Internet Draft Network Working Group R. R. Stewart INTERNET-DRAFT Q. Xie Motorola K. Morneault C. Sharp Cisco H. J. Schwarzbauer Siemens T. Taylor Nortel Networks I. Rytina Ericsson M. Kalla Telcordia L. Zhang UCLA V. Paxson ACIRI expires in six months March 2,2000 Simple Control Transmission Protocol <draft-ietf-sigtran-sctp-07.txt> Status of This Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC 2026. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. Stewart, et al [Page 1] Internet Draft Simple Control Transmission Protocol March 2000 Abstract This document describes the Simple Control Transmission Protocol (SCTP). SCTP is designed to transport PSTN signaling messages over IP networks, but is capable of broader applications. SCTP is a reliable datagram transfer protocol operating on top of an unreliable routed packet network such as IP. It offers the following services to its users: -- acknowledged error-free non-duplicated transfer of user data, -- data segmentation to conform to discovered path MTU size, -- sequenced delivery of user messages within multiple streams, with an option for order-of-arrival delivery of individual user messages, -- optional multiplexing of user messages into SCTP datagrams, and -- network-level fault tolerance through supporting of multi-homing at either or both ends of an association. The design of SCTP includes appropriate congestion avoidance behavior and resistance to flooding and masquerade attacks. Stewart, et al [Page 2] Internet Draft Simple Control Transmission Protocol March 2000 TABLE OF CONTENTS 1. Introduction..................................................5 1.1 Motivation..................................................5 1.2 Architectural View of SCTP..................................5 1.3 Functional View of SCTP.....................................6 1.3.1 Association Startup and Takedown........................7 1.3.2 Sequenced Delivery within Streams.......................7 1.3.3 User Data Segmentation..................................8 1.3.4 Acknowledgment and Congestion Avoidance.................8 1.3.5 Chunk Multiplex.........................................8 1.3.6 Path Management.........................................8 1.3.7 Message Validation......................................9 1.4 Recapitulation of Key Terms.................................9 1.5 Abbreviations...............................................11 2. SCTP Datagram Format..........................................11 2.1 SCTP Common Header Field Descriptions.......................12 2.2 Chunk Field Descriptions....................................13 2.2.1 Optional/Variable-length Parameter Format...............14 2.2.2 Vendor-Specific Extension Parameter Format..............15 2.3 SCTP Chunk Definitions......................................17 2.3.1 Initiation (INIT).......................................17 2.3.1.1 Optional or Variable Length Parameters..............19 2.3.2 Initiation Acknowledgment (INIT ACK)....................20 2.3.2.1 Optional or Variable Length Parameters..............21 2.3.3 Selective Acknowledgment (SACK).........................22 2.3.4 Heartbeat Request (HEARTBEAT)...........................25 2.3.5 Heartbeat Acknowledgment (HEARTBEAT ACK)................26 2.3.6 Abort Association (ABORT)...............................26 2.3.7 Shutdown Association (SHUTDOWN).........................27 2.3.8 Shutdown Acknowledgment (SHUTDOWN ACK)..................28 2.3.9 Operation Error (ERROR).................................28 2.3.10 State Cookie (COOKIE)..................................30 2.3.11 Cookie Acknowledgment (COOKIE ACK).....................31 2.3.12 Payload Data (DATA)....................................31 2.4 Vendor-Specific Chunk Extensions............................33 3. SCTP Association State Diagram.................................34 4. Association Initialization.....................................36 4.1 Normal Establishment of an Association......................37 4.1.1 Handle Stream Parameters................................39 4.1.2 Handle Address Parameters...............................39 4.1.3 Generating State Cookie.................................39 4.1.4 Cookie Processing.......................................40 4.1.5 Cookie Authentication...................................40 4.1.6 An Example of Normal Association Establishment..........41 4.2 Handle Duplicate INIT, INIT ACK, COOKIE, and COOKIE ACK.....42 4.2.1 Handle Duplicate INIT in COOKIE-WAIT or COOKIE-SENT States...................................43 4.2.2 Handle Duplicate INIT in Other States...................43 4.2.3 Handle Duplicate INIT ACK...............................43 4.2.4 Handle Duplicate COOKIE.................................43 4.2.5 Handle Duplicate COOKIE-ACK.............................45 4.2.6 Handle Stale COOKIE Error...............................45 4.3 Other Initialization Issues.................................45 Stewart, et al [Page 3] Internet Draft Simple Control Transmission Protocol March 2000 4.3.1 Selection of Tag Value..................................45 5. User Data Transfer.............................................46 5.1 Transmission of DATA Chunks.................................47 5.2 Acknowledgment of Reception of DATA Chunks..................48 5.2.1 Tracking Peer's Receive Buffer Space....................49 5.3 Management Retransmission Timer.............................50 5.3.1 RTO Calculation.........................................50 5.3.2 Retransmission Timer Rules..............................51 5.3.3 Handle T3-rxt Expiration................................52 5.4 Multi-homed SCTP Endpoints..................................53 5.4.1 Failover from Inactive Destination Address..............54 5.5 Stream Identifier and Stream Sequence Number................54 5.6 Ordered and Un-ordered Delivery.............................54 5.7 Report Gaps in Received DATA TSNs...........................55 5.8 Adler-32 Checksum Calculation...............................56 5.9 Segmentation................................................57 5.10 Bundling and Multiplexing..................................58 6. Congestion Control ..........................................58 6.1 SCTP Differences from TCP Congestion Control................59 6.2 SCTP Slow-Start and Congestion Avoidance....................59 6.2.1 Slow-Start..............................................60 6.2.2 Congestion Avoidance....................................61 6.2.3 Congestion Control......................................61 6.2.4 Fast Retransmit on Gap Reports..........................62 6.3 Path MTU Discovery..........................................63 7. Fault Management..............................................64 7.1 Endpoint Failure Detection..................................64 7.2 Path Failure Detection......................................64 7.3 Path Heartbeat..............................................65 7.4 Handle "Out of the blue" Packets............................66 7.5 Verification Tag............................................67 7.5.1 Exceptions in Verification Tag Rules....................67 8. Termination of Association.....................................68 8.1 Close of an Association.....................................68 8.2 Shutdown of an Association..................................68 9. Interface with Upper Layer.....................................69 9.1 ULP-to-SCTP.................................................70 9.2 SCTP-to-ULP.................................................78 10. Security Considerations.......................................82 10.1 Security Objectives........................................82 10.2 SCTP Responses To Potential Threats........................82 10.2.1 Countering Insider Attacks.............................82 10.2.2 Protecting against Data Corruption in the Network......83 10.2.3 Protecting Confidentiality.............................83 10.2.4 Protecting against Blind Denial of Service Attacks.....83 10.2.4.1 Flooding...........................................84 10.2.4.2 Masquerade.........................................84 10.2.4.3 Improper Monopolization of Services................85 10.3 Protection against Fraud and Repudiation...................85 11. Recommended Transmission Control Block (TCB) Parameters.......86 11.1 Parameters necessary for the SCTP instance.................86 11.2 Parameters necessary per association (i.e. the TCB)........87 11.3 Per Transport Address Data.................................88 11.4 General Parameters Needed..................................89 12. IANA Consideration............................................89 12.1 IETF-defined Chunk Extension...............................89 12.2 IETF-defined Chunk Parameter Extension.....................90 12.3 IETF-defined Additional Error Causes.......................91 12.4 Payload Protocol Identifiers...............................92 Stewart, et al [Page 4] Internet Draft Simple Control Transmission Protocol March 2000 13. Suggested SCTP Protocol Parameter Values......................92 14. Acknowledgments...............................................92 15. Authors' Addresses............................................93 16. References....................................................94 Appendix A .......................................................95 1. Introduction This section explains the reasoning behind the development of the Simple Control Transmission Protocol (SCTP), the services it offers, and the basic concepts needed to understand the detailed description of the protocol. 1.1 Motivation TCP [8] has performed immense service as the primary means of reliable data transfer in IP networks. However, an increasing number of recent applications have found TCP too limiting, and have incorporated their own reliable data transfer protocol on top of UDP [9]. The limitations which users have wished to bypass include the following: -- TCP provides both reliable data transfer and strict order- of-transmission delivery of data. Some applications need reliable transfer without sequence maintenance, while others would be satisfied with partial ordering of the data. In both of these cases the head-of-line blocking offered by TCP causes unnecessary delay. -- The stream-oriented nature of TCP is often an inconvenience. Applications must add their own record marking to delineate their messages, and must make explicit use of the push facility to ensure that a complete message is transferred in a reasonable time. -- The limited scope of TCP sockets complicates the task of providing highly-available data transfer capability using multi-homed hosts. -- TCP is relatively vulnerable to denial of service attacks, such as SYN attacks. Transport of PSTN signaling across the IP network is an application for which all of these limitations of TCP are relevant. While this application directly motivated the development of SCTP, other applications may find SCTP a good match to their requirements. 1.2 Architectural View of SCTP SCTP is viewed as a layer between the SCTP user application ("SCTP user" for short) and an unreliable routed packet network service such as IP. The basic service offered by SCTP is the reliable transfer of user messages between peer SCTP users. It performs this service Stewart, et al [Page 5] Internet Draft Simple Control Transmission Protocol March 2000 within the context of an association between two SCTP nodes. Chapter 9 of this document sketches the API which should exist at the boundary between the SCTP and the SCTP user layers. SCTP is connection-oriented in nature, but the SCTP association is a broader concept than the TCP connection. SCTP provides the means for each SCTP endpoint (Section 1.4) to provide the other during association startup with a list of transport addresses (e.g. multiple IP addresses in combination with an SCTP port) through which that endpoint can be reached and from which it will originate messages. The association spans transfers over all of the possible source/destination combinations which may be generated from the two endpoint lists. _____________ _____________ | SCTP User | | SCTP User | | Application | | Application | |-------------| |-------------| | SCTP | | SCTP | | Transport | | Transport | | Service | | Service | |-------------| |-------------| | |One or more ---- One or more| | | IP Network |IP address \/ IP address| IP Network | | Service |appearances /\ appearances| Service | |_____________| ---- |_____________| SCTP Node A |<-------- Network transport ------->| SCTP Node B Figure 1: An SCTP Association 1.3 Functional View of SCTP The SCTP transport service can be decomposed into a number of functions. These are depicted in Figure 2 and explained in the remainder of this section. SCTP User Application ..----------------------------------------------------- .. _____________ ____________________ | | | Sequenced delivery | | Association | | within streams | | | |____________________| | startup | ..| | ____________________________ | and | | User Data Segmentation | | | |____________________________| | takedown | Stewart, et al [Page 6] Internet Draft Simple Control Transmission Protocol March 2000 ..| | ____________________________ | | | Acknowledgment | | | | and | | | | Congestion Avoidance | ..| | |____________________________| | | | | ____________________________ | | | Chunk Multiplex | | | |____________________________| | | | | ________________________________ | | | Message Validataion | | | |________________________________| | | | | ________________________________ | | | Path Management | |______________ |________________________________| Figure 2: Functional View of the SCTP Transport Service 1.3.1 Association Startup and Takedown An association is initiated by a request from the SCTP user (see the description of the ASSOCIATE primitive in Chapter 9). A cookie mechanism, taken from that devised by Karn and Simpson in RFC 2522 [6], is employed during the initialization to provide protection against security attacks. The cookie mechanism uses a four-way handshaking, but the last two legs of which are allowed to carry user data for fast setup. The startup sequence is described in chapter 4 of this document. SCTP provides for graceful takedown of an active association on request from the SCTP user. See the description of the TERMINATE primitive in chapter 9. SCTP also allows ungraceful takedown, either on request from the user (ABORT primitive) or as a result of an error condition detected within the SCTP layer. Chapter 8 describes both the graceful and the ungraceful takedown procedures. 1.3.2 Sequenced Delivery within Streams The term "stream" is used in SCTP to refer to a sequence of user messages. This is in contrast to its usage in TCP, where it refers to a sequence of bytes. The SCTP user can specify at association startup time the number of streams to be supported by the association. This number is negotiated with the remote end (see section 4.1.1). User messages are associated with stream numbers (SEND, RECEIVE primitives, Chapter 9). Internally, SCTP assigns a stream sequence number to each message passed to it by Stewart, et al [Page 7] Internet Draft Simple Control Transmission Protocol March 2000 the SCTP user. On the receiving side, SCTP ensures that messages are delivered to the SCTP user in sequence within a given stream. However, while one stream may be blocked waiting for the next in-sequence user message, delivery from other streams may proceed. SCTP provides a mechanism for bypassing the sequenced delivery service. User messages sent using this mechanism are delivered to the SCTP user as soon as they are received. 1.3.3 User Data Segmentation SCTP can segment user messages to ensure that the SCTP datagram passed to the lower layer conforms to the path MTU. Segments are reassembled into complete messages before being passed to the SCTP user. 1.3.4 Acknowledgment and Congestion Avoidance SCTP assigns a Transmission Sequence Number (TSN) to each user data segment or unsegmented message. The TSN is independent of any stream sequence number assigned at the stream level. The receiving end acknowledges all TSNs received, even if there are gaps in the sequence. In this way, reliable delivery is kept functionally separate from sequenced delivery. The Acknowledgment and Congestion Avoidance function is responsible for message retransmission when timely acknowledgment has not been received. Message retransmission is conditioned by congestion avoidance procedures similar to those used for TCP. See Chapters 5 and 6 for a detailed description of the protocol procedures associated with this function. 1.3.5 Chunk Multiplex As described in Chapter 2, the SCTP datagram as delivered to the lower layer consists of a common header followed by one or more chunks. Each chunk may contain either user data or SCTP control information. The SCTP user has the option to request "bundling", or multiplexing of more than one user messages into a single SCTP datagram. The chunk multiplex function of SCTP is responsible for assembly of the complete SCTP datagram and its disassembly at the receiving end. 1.3.6 Path Management The sending SCTP user is able to manipulate the set of transport addresses used as destinations for SCTP datagrams, through the primitives described in Chapter 9. The SCTP path management function chooses the destination transport address for each outgoing SCTP datagram based on the SCTP user's instructions and the currently perceived reachability status of the eligible destination set. Stewart, et al [Page 8] Internet Draft Simple Control Transmission Protocol March 2000 The path management function monitors reachability through heartbeat messages when other message traffic is inadequate to provide this information, and advises the SCTP user when reachability of any far- end transport address changes. The path management function is also responsible for reporting the eligible set of local transport addresses to the far end during association startup, and for reporting the transport addresses returned from the far end to the SCTP user. At association start-up, a primary destination transport address is defined for each SCTP endpoint, and is used for normal sending of SCTP datagrams. On the receiving end, the path management is responsible for verifying the existence of a valid SCTP association to which the inbound SCTP datagram belongs before passing it for further processing. 1.3.7 Message Validation A mandatory verification tag and an Adler-32 checksum [2] fields are included in the SCTP common header. The verification tag value is chosen by each end of the association during association startup. Messages received without the verification tag value expected by the receiver are discarded, as a protection against blind masquerade attacks and against stale datagrams from a previous association. The Adler-32 checksum should be set by the sender of each SCTP datagram, to provide additional protection against data corruption in the network beyond that provided by lower layers (e.g. the IP checksum). 1.4 Recapitulation of Key Terms The language used to describe SCTP has been introduced in the previous sections. This section provides a consolidated list of the key terms and their definitions. o SCTP user application (SCTP user): The logical higher-layer application entity which uses the services of SCTP, also called the Upper-layer Protocol (ULP). o User message: the unit of data delivery across the interface between SCTP and its user. o SCTP datagram: the unit of data delivery across the interface between SCTP and the unreliable packet network (e.g. IP) which it is using. An SCTP datagram includes the common SCTP header, possible SCTP control chunks, and user data encapsulated within SCTP DATA chunks. o Transport address: an address which serves as a source or destination for the unreliable packet transport service used by SCTP. In IP networks, a transport address is defined by the combination of an IP address and an SCTP port number. Stewart, et al [Page 9] Internet Draft Simple Control Transmission Protocol March 2000 Note, only one SCTP port may be defined for each endpoint, but each endpoint may have multiple IP addresses. o SCTP endpoint: the logical sender/receiver of SCTP datagrams. On a multi-homed host, an SCTP endpoint is represented to its peers as a combination of a set of eligible destination transport addresses to which SCTP datagrams can be sent and a set of eligible source transport addresses from which SCTP datagrams can be received. Note, a source or destination transport address can only be included in one unique SCTP endpoint, i.e., it is NOT allowed to have the same SCTP source or destination transport address appear in more than one SCTP endpoint. o SCTP association: a protocol relationship between SCTP endpoints, comprising the two SCTP endpoints and protocol state information including verification tags and the currently active set of Transmission Sequence Numbers (TSNs), etc. o Chunk: a unit of information within an SCTP datagram, consisting of a chunk header and chunk-specific content. o Transmission Sequence Number (TSN): a 32-bit sequence number used internally by SCTP. One TSN is attached to each chunk containing user data to permit the receiving SCTP endpoint to acknowledge its receipt and detect duplicate deliveries. o Stream: a uni-directional logical channel established from one to another associated SCTP endpoints, within which all user messages are delivered in sequence except for those submitted to the un-ordered delivery service. Note: The relationship between stream numbers in opposite directions is strictly a matter of how the applications use them. It is the responsibility of the SCTP user to create and manage these correlations if they are so desired. o Stream Sequence Number: a 16-bit sequence number used internally by SCTP to assure sequenced delivery of the user messages within a given stream. One stream sequence number is attached to each user message. o Path: the route taken by the SCTP datagrams sent by one SCTP endpoint to a specific destination transport address of its peer SCTP endpoint. Note, sending to different destination transport addresses does not necessarily guarantee getting separate paths. o Bundling: an optional multiplexing operation, whereby more than one user messages may be carried in the same SCTP datagram. Each user message occupies its own DATA chunk. o Outstanding TSN (at an SCTP endpoint): a TSN (and the associated DATA chunk) which have been sent by the endpoint but for which it has not yet received an acknowledgment. Stewart, et al [Page 10] Internet Draft Simple Control Transmission Protocol March 2000 o Unacknowledged TSN (at an SCTP endpoint): a TSN (and the associated DATA chunk) which have been received by the endpoint but for which an acknowledgment has not yet been sent. o Receiver Window (rwnd): The most recently calculated receiver window, in number of octets. This gives an indication of the space available in the receiver's inbound buffer. o Congestion Window (cwnd): An SCTP variable that limits the data, in number of octets, a sender can send into the network before receiving an acknowledgment on a particular destination Transport address. o Slow Start Threshold (ssthresh): An SCTP variable. This is the threshold which the endpoint will use to determine whether to perform slow start or congestion avoidance on a particular destination transport address. Ssthresh is in number of octets. o Transmission Control Block (TCB): an internal data structure created by an SCTP endpoint for each of its existing SCTP associations to other SCTP endpoints. TCB contains all the status and operational information for the endpoint to maintain and manage the corresponding association. o Network Byte Order: Most significant byte first, a.k.a Big Endian. 1.5. Abbreviations ICV - Integrity Check Value [4] RTO - Retransmission Time-out RTT - Round-trip Time RTTVAR - Round-trip Time Variation SCTP - Simple Control Transmission Protocol SRTT - Smoothed RTT TCB - Transmission Control Block TLV - Type-Length-Value Coding Format TSN - Transmission Sequence Number ULP - Upper-layer Protocol 2. SCTP Datagram Format An SCTP datagram is composed of a common header and chunks. A chunk contains either control information or user data. Stewart, et al [Page 11] Internet Draft Simple Control Transmission Protocol March 2000 The SCTP datagram format is shown below: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Common Header | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Chunk #1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | ... | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Chunk #n | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Multiple chunks can be multiplexed into one SCTP datagram up to the MTU size, except for the INIT, INIT ACK, and SHUTDOWN ACK chunks. These chunks MUST NOT be multiplexed with any other chunk in a datagram. See Section 5.10 for more details on chunk multiplexing. If an user data message doesn't fit into one SCTP datagram it can be segmented into multiple chunks using the procedure defined in Section 5.9. All integer fields in an SCTP datagram MUST be transmitted in the network byte order, unless otherwise stated. 2.1 SCTP Common Header Field Descriptions SCTP Common Header Format 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Source Port Number | Destination Port Number | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Verification Tag | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Adler-32 Checksum | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Source Port Number: 16 bit u_int This is the SCTP sender's port number. It can be used by the receiver, in combination with the source IP address, to identify the association to which this datagram belongs. Destination Port Number: 16 bit u_int This is the SCTP port number to which this datagram is destined. The receiving host will use this port number to de-multiplex the SCTP datagram to the correct receiving endpoint/application. Stewart, et al [Page 12] Internet Draft Simple Control Transmission Protocol March 2000 Verification Tag: 32 bit u_int The receiver of this datagram uses the Verification Tag to validate the sender of this SCTP datagram. On transmit, the value of this Verification Tag MUST be set to the value of the Initiate Tag received from the peer endpoint during the association initialization. For datagrams carrying the INIT chunk, the transmitter MUST set the Verification Tag to all 0's. If the receiver receives a datagram with an all-zeros Verification Tag field, it checks the Chunk ID immediately following the common header. If the Chunk Type is neither INIT nor SHUTDOWN ACK, the receiver MUST drop the datagram. For datagrams carrying the SHUTDOWN ACK chunk, the transmitter SHOULD set the Verification Tag to the Initiate Tag received from the peer endpoint during the association initialization, if known. Otherwise, the Verification Tag MUST be set to all 0's. Adler-32 Checksum: 32 bit u_int This field MUST contain an Adler-32 checksum of this SCTP datagram. Its calculation is discussed in Section 5.8. 2.2 Chunk Field Descriptions The figure below illustrates the field format for the chunks to be transmitted in the SCTP datagram. Each chunk is formatted with a Chunk ID field, a chunk-specific Flag field, a Length field, and a Value field. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Chunk ID | Chunk Flags | Chunk Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Chunk Value / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk ID: 8 bits, u_int This field identifies the type of information contained in the Chunk Value field. It takes a value from 0x00 to 0xFF. The value of 0xFE is reserved for vendor-specific extensions. The value of 0xFF is reserved for future use as an extension field. Procedures for extending this field by vendors are defined in Section 2.4. The values of Chunk ID are defined as follows: Stewart, et al [Page 13] Internet Draft Simple Control Transmission Protocol March 2000 ID Value Chunk Type ----- ---------- 00000000 - Payload Data (DATA) 00000001 - Initiation (INIT) 00000010 - Initiation Acknowledgment (INIT ACK) 00000011 - Selective Acknowledgment (SACK) 00000100 - Heartbeat Request (HEARTBEAT) 00000101 - Heartbeat Acknowledgment (HEARTBEAT ACK) 00000110 - Abort (ABORT) 00000111 - Shutdown (SHUTDOWN) 00001000 - Shutdown Acknowledgment (SHUTDOWN ACK) 00001001 - Operation Error (ERROR) 00001010 - State Cookie (COOKIE) 00001011 - Cookie Acknowledgment (COOKIE ACK) 00001100 - Reserved for Explict Congestion Notification Echo (ECNE) 00001101 - Reserved for Congestion Window Reduced (CWR) 00001110 to 11111101 - reserved by IETF 11111110 - Vendor-specific Chunk Extensions 11111111 - IETF-defined Chunk Extensions Note: The ENCE and CWR chunk types are reserved for future use of Explicit Congestion Notification (ECN). Chunk Flags: 8 bits The usage of these bits depends on the chunk type as given by the Chunk ID. Unless otherwise specified, they are set to zero on transmit and are ignored on receipt. Chunk Length: 16 bits (u_int) This value represents the size of the chunk in octets including the Chunk ID, Flags, Length, and Value fields. Therefore, if the Value field is zero-length, the Length field will be set to 0x0004. The Length field does not count any padding. Chunk Value: variable length The Chunk Value field contains the actual information to be transferred in the chunk. The usage and format of this field is dependent on the Chunk ID. The Chunk Value field MUST be aligned on 32-bit boundaries. If the length of the chunk does not align on 32-bit boundaries, it is padded at the end with all zero octets. SCTP defined chunks are described in detail in Section 2.3. The guideline for vendor-specific chunk extensions is discussed in Section 2.4. And the guidelines for IETF-defined chunk extensions can be found in Section 12.1 of this document. 2.2.1 Optional/Variable-length Parameter Format The optional and variable-length parameters contained in a chunk are defined in a Type-Length-Value format as shown below. Stewart, et al [Page 14] Internet Draft Simple Control Transmission Protocol March 2000 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Parameter Type | Parameter Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Parameter Value / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Parameter Type: 16 bit u_int The Type field is a 16 bit identifier of the type of parameter. It takes a value of 0x0000 to 0xFFFF. The value of 0xFFFE is reserved for vendor-specific extensions if the specific chunk allows such extensions. The value of 0xFFFF is reserved for IETF-defined extensions. Values other than those defined in specific SCTP chunk description are reserved for use by IETF. Parameter Length: 16 bit u_int The Length field contains the size of the parameter in octets, including the Type, Length, and Value fields. Thus, a parameter with a zero-length Value field would have a Length field of 0x0004. The Length does not include any padding octets. Parameter Value: variable-length. The Value is dependent on the value of the Type field. The value field MUST be aligned on 32-bit boundaries. If the value field is not aligned on 32-bit boundaries it is padded at the end with all zero octets. The value field must be an integer number of octets. The actual SCTP parameters are defined in the specific SCTP chunk section. The guidelines for vendor-specific parameter extensions are discussed in Section 2.2.2. And the rules for IETF-defined parameter extensions are defined in Section 12.2. 2.2.2 Vendor-Specific Extension Parameter Format This is to allow vendors to support their own extended parameters not defined by the IETF. It MUST not affect the operation of SCTP. Endpoints not equipped to interpret the vendor-specific information sent by a remote endpoint MUST ignore it (although it may be reported). Endpoints that do not receive desired vendor-specific information SHOULD make an attempt to operate without it, although they may do so (and report they are doing so) in a degraded mode. A summary of the Vendor-specific extension format is shown below. The fields are transmitted from left to right. Stewart, et al [Page 15] Internet Draft Simple Control Transmission Protocol March 2000 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Parameter Type = 0xFFFE | Parameter Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Vendor-Id | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Parameter Value / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Type: 16 bit u_int 0xFFFE for all Vendor-Specific parameters. Length: 16 bit u_int Indicate the size of the parameter in octets, including the Type, Length, Vendor-Id, and Value fields. Vendor-Id: 32 bit u_int The high-order octet is 0 and the low-order 3 octets are the SMI Network Management Private Enterprise Code of the Vendor in network byte order, as defined in the Assigned Numbers (RFC 1700). Value: variable length The Value field is one or more octets. The actual format of the information is site or application specific, and a robust implementation SHOULD support the field as undistinguished octets. The codification of the range of allowed usage of this field is outside the scope of this specification. It SHOULD be encoded as a sequence of vendor type / vendor length / value fields, as follows. The parameter field is dependent on the vendor's definition of that attribute. An example encoding of the Vendor-Specific attribute using this method follows: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Parameter Type = 0xFFFE | Parameter Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Vendor-Id | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | VS-Type | VS-Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / VS-Value / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Stewart, et al [Page 16] Internet Draft Simple Control Transmission Protocol March 2000 VS-Type: 16 bit u_int This field identifies the parameter included in the VS-Value field. It is assigned by the vendor. VS-Length: 16 bit u_int This field is the length of the vendor-specific parameter and Includes the VS-Type, VS-Length and VS-Value (if included) fields. VS-Value: Variable Length This field contains the parameter identified by the VS-Type field. It's meaning is identified by the vendor. 2.3 SCTP Chunk Definitions This section defines the format of the different SCTP chunk types. 2.3.1 Initiation (INIT) (00000001) This chunk is used to initiate a SCTP association between two endpoints. The format of the INIT message is shown below: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 0 0 1| Chunk Flags | Chunk Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Initiate Tag | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Advertised Receiver Window Credit (a_rwnd) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Number of Outbound Streams | Number of Inbound Streams | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Initial TSN | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Optional/Variable-Length Parameters / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The INIT chunk contains the following parameters. Unless otherwise noted, each parameter MUST only be included once in the INIT chunk. Fixed Parameters Status ---------------------------------------------- Initiate Tag Mandatory Advertised Receiver Window Credit Mandatory Number of Outbound Streams Mandatory Number of Inbound Streams Mandatory Initial TSN Mandatory Stewart, et al [Page 17] Internet Draft Simple Control Transmission Protocol March 2000 Variable Parameters Status Type Value ------------------------------------------------------------- IPv4 Address (Note 1) Optional 0x0005 IPv6 Address (Note 1) Optional 0x0006 Cookie Preservative Optional 0x0009 Reserved for ECN Capable Optional 0x000a Note 1: The INIT chunks may contain multiple addresses that may be IPv4 and/or IPv6 in any combination. Note 2: The ECN capable field is reserved for future use of Explicit Congestion Notification. Chunk Flags field in INIT is reserved, and all bits in it should be set to 0 by the sender and ignored by the receiver. The sequence of parameters within an INIT may be processed in any order. Initiate Tag: 32 bit u_int The receiver of the INIT (the responding end) records the value of the Initiate Tag parameter. This value MUST be placed into the Verification Tag field of every SCTP datagram that the responding end transmits within this association. The valid range for Initiate Tag is from 0x1 to 0xffffffff. See Section 4.3.1 for more on the selection of the tag value. If the value of the Initiate Tag in a received INIT chunk is found to be 0x0, the receiver MUST treat it as an error and silently discard the datagram. Advertised Receiver Window Credit (a_rwnd): 32 bit u_int This value represents the dedicated buffer space, in number of octets, the sender of the INIT has placed in association with this window. During the life of the association this buffer space SHOULD not be lessened (i.e. dedicated buffers taken away from this association). Number of Outbound Streams (OS): 16 bit u_int Defines the number of outbound streams the sender of this INIT chunk wishes to create in this association. The value of 0 MUST NOT be used. Number of Inbound Streams (MIS) : 16 bit u_int Defines the maximum number of streams the sender of this INIT chunk allows the peer end to create in this association. The value 0 MUST NOT be used. Initial TSN (I-TSN) : 32 bit u_int Defines the initial TSN that the sender will use. The valid range is from 0x0 to 0xffffffff. This field MAY be set to the value of the Initiate Tag field. Stewart, et al [Page 18] Internet Draft Simple Control Transmission Protocol March 2000 Vendor-specific parameters are allowed in INIT. However, they MUST be appended to the end of the above INIT chunks. The format of the vendor-specific parameters MUST follow the Type-Length-value format as defined in Section 2.2.2. In case an endpoint does not support the vendor-specific chunks received, it MUST ignore them. 2.3.1.1 Optional/Variable Length Parameters in INIT The following parameters follow the Type-Length-Value format as defined in Section 2.2.1. The IP address fields MUST come after the fixed-length fields defined in the previous Section. Any extensions MUST come after the IP address fields. IPv4 Address Parameter 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1|0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | IPv4 Address | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ IPv4 Address: 32 bit Contains an IPv4 address of the sending endpoint. It is binary encoded. IPv6 Address: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0|0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | | IPv6 Address | | | | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ IPv6 Address: 128 bit Contains an IPv6 address of the sending endpoint. It is binary encoded. Combining with the Source Port Number in the SCTP common header, the value passed in an IPv4 or IPv6 Address parameter indicates a transport address the sender of the INIT will support for the association being initiated. That is, during the lifetime of this association, this IP address may appear in the source address field Stewart, et al [Page 19] Internet Draft Simple Control Transmission Protocol March 2000 of a datagram sent from the sender of the INIT, and may be used as a destination address of a datagram sent from the receiver of the INIT. More than one IP Address parameters can be included in an INIT chunk when the INIT sender is multi-homed. Moreover, a multi-homed endpoint may have access to different types of network, thus more than one address type may be present in one INIT chunk, i.e., IPv4 and IPv6 transport addresses are allowed in the same INIT message. If the INIT contains at least one IP Address parameter, then only the transport address(es) provided within the INIT may be used as destinations by the responding end. If the INIT does not contain any IP Address parameters, the responding end MUST use the source address associated with the received SCTP datagram as its sole destination address for the association. Cookie Preservative The sender of the INIT shall use this parameter to suggest to the receiver of the INIT for a longer life-span of the State Cookie. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1|0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Suggested Cookie Life-span Increment (msec.) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Suggested Cookie Life-span Increment: 32bit u_int This parameter indicates to the receiver how much increment the sender wishes the receiver to add to its default cookie life-span. This optional parameter should be added to the INIT message by the sender when it re-attempts establishing an association with a peer to which its previous attempt of establishing the association failed due to a Stale COOKIE error. Note, the receiver MAY choose to ignore the suggested cookie life-span increase for its own security reasons. 2.3.2 Initiation Acknowledgment (INIT ACK) (00000010): The INIT ACK chunk is used to acknowledge the initiation of an SCTP association. The parameter part of INIT ACK is formatted similarly to the INIT chunk. It uses two extra variable parameters: The State Cookie and the Unrecognized Parameter: The format of the INIT ACK message is shown below: Stewart, et al [Page 20] Internet Draft Simple Control Transmission Protocol March 2000 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 0 1 0| Chunk Flags | Chunk Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Initiate Tag | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Advertised Receiver Window Credit | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Number of Outbound Streams | Number of Inbound Streams | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Initial TSN | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Optional/Variable-Length Parameters / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The INIT ACK contains the following parameters. Unless otherwise noted, each parameter MUST only be included once in the INIT ACK chunk. Fixed Parameters Status ---------------------------------------------- Initiate Tag Mandatory Advertised Receiver Window Credit Mandatory Number of Outbound Streams Mandatory Number of Inbound Streams Mandatory Initial TSN Mandatory Variable Parameters Status Type Value ------------------------------------------------------------- State Cookie Mandatory 0x0007 IPv4 Address (Note 1) Optional 0x0005 IPv6 Address (Note 1) Optional 0x0006 Unrecognized Parameters Optional 0x0008 Reserved for ECN Capable Optional 0x000a Note 1: The INIT ACK chunks may contain any number of IP address parameters that may be IPv4 and/or IPv6 in any combination. Note 2: The ECN capable field is reserved for future use of Explicit Congestion Notification. Same as with INIT, in combination with the Source Port carried in the SCTP common header, each IP Address parameter in the INIT ACK indicates to the receiver of the INIT ACK a valid transport address supported by the sender of the INIT ACK for the lifetime of the association being initiated. If the INIT ACK contains at least one IP Address parameter, then only the transport address(es) explicitly indicated in the INIT ACK may be used as the destination(s) by the receiver of the INIT ACK. However, if the INIT ACK contains no IP Address parameter, the receiver of the INIT ACK MUST take the source IP address associated with this INIT ACK as its sole destination address for this association. Stewart, et al [Page 21] Internet Draft Simple Control Transmission Protocol March 2000 The State Cookie and Unrecognized Parameters use the Type-Length- Value format as defined in Section 2.2.1 and are described below. The other fields are defined the same as their counterparts in the INIT message. 2.3.2.1 Optional or Variable Length Parameters State Cookie: variable size, depending on Size of Cookie This field MUST contain all the necessary state and parameter information required for the sender of this INIT ACK to create the association, along with an Integrity Check Value (ICV). See Section 4.1.3 for details on Cookie definition. The Cookie MUST be padded with '0' to the next 32-bit word boundary. The internal format of the Cookie is implementation-specific. Unrecognized Parameters: Variable Size. This parameter is returned to the originator of the INIT message if the receiver does not recognize one or more Optional TLV parameters in the INIT chunk. This parameter field will contain the unrecognized parameters copied from the INIT message complete with TLV. Vendor-Specific parameters are allowed in INIT ACK. However, they MUST be defined using the format described in Section 2.2.2, and be appended to the end of the above INIT ACK chunk. In case the receiver of the INIT ACK does not support the vendor-specific parameters received, it MUST ignore those fields. 2.3.3 Selective Acknowledgment (SACK) (00000011): This chunk is sent to the remote endpoint to acknowledge received DATA chunks and to inform the remote endpoint of gaps in the received subsequences of DATA chunks as represented by their TSNs. The SACK MUST contain the Cumulative TSN ACK and Advertised Receiver Window Credit (a_rwnd) parameters. By definition, the value of the Cumulative TSN ACK parameter is the last TSN received at the time the Selective ACK is sent, before a break in the sequence of received TSNs occurs; the next TSN value following this one has not yet been received at the reporting end. This parameter therefore acknowledges receipt of all TSNs up to and including the value given. The handling of the a_rwnd by the receiver of the SACK is discussed in detail in Section 5.2.1. The Selective ACK also contains zero or more fragment reports. Each fragment report acknowledges a subsequence of TSNs received following a break in the sequence of received TSNs. By definition, all TSNs acknowledged by fragment reports are higher than the value of the Cumulative TSN ACK. Stewart, et al [Page 22] Internet Draft Simple Control Transmission Protocol March 2000 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 0 1 1|Chunk Flags | Chunk Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cumulative TSN ACK | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Advertised Receiver Window Credit (a_rwnd) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Number of Fragments = N | Number of Duplicate TSNs = X | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Fragment #1 Start | Fragment #1 End | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / / \ ... \ / / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Fragment #N Start | Fragment #N End | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Duplicate TSN 1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / / \ ... \ / / +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Duplicate TSN X | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: Set to all zeros on transmit and ignored on receipt. Cumulative TSN ACK: 32 bit u_int This parameter contains the TSN of the last DATA chunk received in sequence before a gap. Advertised Receiver Window Credit (a_rwnd): 32 bit u_int This field indicates the updated receive buffer space in octets of the sender of this SACK, see Section 5.2.1 for details. Number of Fragments: 16 bit u_int Indicates the number of TSN fragments included in this Selective ACK. Number of Duplicate TSNs: 16 bit This field contains the number of duplicate TSNs the endpoint has received. Each duplicate TSN is listed following the fragment list. Fragments: These fields contain the ack fragments. They are repeated for each fragment up to the number of fragments defined in the Number of Fragments field. All DATA chunks with TSNs between the (Cumulative TSN ACK + Fragment Start) and (Cumulative TSN ACK + Fragment End) of each fragment are assumed to have been received correctly. Stewart, et al [Page 23] Internet Draft Simple Control Transmission Protocol March 2000 Fragment Start: 16 bit u_int Indicates the Start offset TSN for this fragment. To calculate the actual TSN number the Cumulative TSN ACK is added to this offset number to yield the TSN. This calculated TSN identifies the first TSN in this fragment that has been received. Fragment End: 16 bit u_int Indicates the End offset TSN for this fragment. To calculate the actual TSN number the Cumulative TSN ACK is added to this offset number to yield the TSN. This calculated TSN identifies the TSN of the last DATA chunk received in this fragment. Duplicate TSN: 32 bit u_int Indicates a TSN that was received in duplicate. For example, assume the receiver has the following datagrams newly arrived at the time when it decides to send a Selective ACK, ---------- | TSN=17 | ---------- | | <- still missing ---------- | TSN=15 | ---------- | TSN=14 | ---------- | | <- still missing ---------- | TSN=12 | ---------- | TSN=11 | ---------- | TSN=10 | ---------- then, the parameter part of the Selective ACK MUST be constructed as follows (assuming the new a_rwnd is set to 0x1234 by the sender): +---------------+--------------+ | Cumulative TSN ACK = 12 | ----------------+--------------- | a_rwnd = 0x1234 | ----------------+--------------- | num of frag=2 | (rev = 0) | ----------------+--------------- |frag #1 strt=2 |frag #1 end=3 | ----------------+--------------- |frag #2 strt=5 |frag #2 end=5 | -------------------------------- Stewart, et al [Page 24] Internet Draft Simple Control Transmission Protocol March 2000 2.3.4 Heartbeat Request (HEARTBEAT) (00000100): An endpoint should send this chunk to its peer endpoint of the current association to probe the reachability of a particular destination transport address defined in the present association. The parameter field contains the Heartbeat Information which is a variable length opaque data structure understood only by the sender. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 1 0 0| Chunk Flags | Heartbeat Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Heartbeat Information (Variable-Length) / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: Set to zero on transmit and ignored on receipt. Heartbeat Length: Set to the size of the chunk in octets, including the chunk header and the Heartbeat Information field. Heartbeat Information: defined as a variable-length parameter using the format described in Section 2.2.1, i.e.: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Heartbeat Info Type=1 | HB Info Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / Sender-specific Heartbeat Info / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The Sender-specific Heartbeat Info field should normally include information about the sender's current time when this HEARTBEAT message is sent and the destination transport address to which this HEARTBEAT is sent (see Section 7.3). Stewart, et al [Page 25] Internet Draft Simple Control Transmission Protocol March 2000 2.3.5 Heartbeat Acknowledgment (HEARTBEAT ACK) (00000101): An endpoint should send this chunk to its peer endpoint as a response to a Heartbeat Request (see Section 7.3). The parameter field contains a variable length opaque data structure. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 1 0 1| Chunk Flags | Heartbeat Ack Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Heartbeat Information (Variable-Length) / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: Set to zero on transmit and ignored on receipt. Heartbeat Ack Length: Set to the size of the chunk in octets, including the chunk header and the Heartbeat Information field. Heartbeat Information: The values of this field SHALL be copied from the Heartbeat Information field found in the Heartbeat Request to which this Heartbeat Acknowledgment is responding. 2.3.6 Abort Association (ABORT) (00000110): The ABORT chunk is sent to the peer of an association to terminate the association. The ABORT chunk may contain cause parameters to inform the receiver the reason of the abort. DATA chunks MUST not be bundled with ABORT. Control chunks MAY be bundled with an ABORT but they MUST be placed before the ABORT in the SCTP datagram, or they will be ignored by the receiver. If an endpoint receives an ABORT with a format error or for an association that doesn't exist, it MUST silently discard it. Moreover, under any circumstances, an endpoint that receives an ABORT MUST never respond to that ABORT by sending an ABORT of its own. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 1 1 0| Chunk Flags | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / zero or more Error Causes / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Stewart, et al [Page 26] Internet Draft Simple Control Transmission Protocol March 2000 Chunk Flags: Set to zero on transmit and ignored on receipt. Length: Set to the size of the chunk in octets, including the chunk header and all the Error Cause fields present. See Section 2.3.9 for Error Cause definitions. Note: Special rules apply to the Verification Tag field of SCTP datagrams which carry an ABORT, see Section 7.5 for details. 2.3.7 SHUTDOWN (00000111): An endpoint in an association MUST use this chunk to initiate a graceful termination of the association with its peer. This chunk has the following format. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 1 1 1|Chunk Flags |0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cumulative TSN ACK | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: Set to zero on transmit and ignored on receipt. Cumulative TSN ACK: 32 bit u_int This parameter contains the TSN of the last chunk received in sequence before any gaps. Stewart, et al [Page 27] Internet Draft Simple Control Transmission Protocol March 2000 2.3.8 Shutdown Acknowledgment (SHUTDOWN ACK) (00001000): This chunk MUST be used to acknowledge the receipt of the SHUTDOWN chunk at the completion of the shutdown process, see Section 8.2 for details. The SHUTDOWN ACK chunk has no parameters. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 1 0 0 0|Chunk Flags |0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: Set to zero on transmit and ignored on receipt. Note: if the endpoint that receives the SHUTDOWN message does not have a TCB or tag for the sender of the SHUTDOWN, the receiver MUST still respond. In such cases, the receiver MUST send back a stand-alone SHUTDOWN ACK chunk in an SCTP datagram with the Verification Tag field of the common header filled with all '0's. 2.3.9 Operation Error (ERROR) (00001001): This chunk is sent to the other endpoint in the association to notify certain error conditions. It contains one or more error causes. It has the following parameters: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 1 0 0 1| Chunk Flags | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / one or more Error Causes / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: Set to zero on transmit and ignored on receipt. Length: Set to the size of the chunk in octets, including the chunk header and all the Error Cause fields present. Error causes are defined as variable-length parameters using the format described in 2.2.1, i.e.: Stewart, et al [Page 28] Internet Draft Simple Control Transmission Protocol March 2000 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cause Code | Cause Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ / Cause-specific Information / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Cause Code: 16 bit u_int Defines the type of error conditions being reported. Cause Length: 16 bit u_int Set to the size of the parameter in octets, including the Cause Code, Cause Length, and Cause-Specific Information fields Cause-specific Information: variable length This field carries the details of the error condition. Currently SCTP defines the following error causes: Cause of error --------------- Invalid Stream Identifier: indicating receiving a DATA sent to a nonexistent stream. +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cause Code=1 | Cause Length=8 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Stream Identifier | (Reserved) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Cause of error --------------- Missing Mandatory Parameter: indicating that mandatory one or more TLV parameters are missing in a received INIT or INIT ACK. +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cause Code=2 | Cause Length=8+N*2 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Number of missing params=N | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Missing Param Type #1 | Missing Param Type #2 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Missing Param Type #N-1 | Missing Param Type #N | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Each missing mandatory parameter type should be specified. Stewart, et al [Page 29] Internet Draft Simple Control Transmission Protocol March 2000 Cause of error -------------- Stale Cookie Error: indicating the receiving of a valid cookie which is however expired. +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cause Code=3 | Cause Length=8 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Measure of Staleness (usec.) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ The sender of this error cause MAY choose to report how long past expiration the cookie is, by putting in the Measure of Staleness field the difference, in microseconds, between the current time and the time the cookie expired. If the sender does not wish to provide this information it should set Measure of staleness to 0. Cause of error --------------- Out of Resource: indicating that the sender is out of resource. This is usually sent in combination with or within an ABORT. +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cause Code=4 | Cause Length=4 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Guidelines for IETF-defined Error Cause extensions are discussed in Section 12.3 of this document. 2.3.10 State Cookie (COOKIE) (00001010): This chunk is used only during the initialization of an association. It is sent by the initiator of an association to its peer to complete the initialization process. This chunk MUST precede any chunk sent within the association, but MAY be bundled with one or more DATA chunks in the same datagram. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 1 0 1 0|Chunk Flags | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Cookie | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: 8 bit Set to zero on transmit and ignored on receipt. Length: 16 bit u_int Set to the size of the chunk in octets, including the 4 octets of the chunk header and the size of the Cookie. Stewart, et al [Page 30] Internet Draft Simple Control Transmission Protocol March 2000 Cookie: variable size This field must contain the exact cookie received in a previous INIT ACK. 2.3.11 Cookie Acknowledgment (COOKIE ACK) (00001011): This chunk is used only during the initialization of an association. It is used to acknowledge the receipt of a COOKIE chunk. This chunk MUST precede any chunk sent within the association, but MAY be bundled with one or more DATA chunks in the same SCTP datagram. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 1 0 1 1|Chunk Flags |0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Chunk Flags: Set to zero on transmit and ignored on receipt. 2.3.12 Payload Data (DATA) (00000000): The following format MUST be used for the DATA chunk: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0 0 0 0 0 0 0 0| Reserved|U|B|E| Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | TSN | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Stream Identifier S | Stream Sequence Number n | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Payload Protocol Identifier | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / User Data (seq n of Stream S) / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Reserved: 5 bits should be set to all '0's and ignored by the receiver. U bit: 1 bit The (U)nordered bit, if set, indicates that this is an unordered data chunk, and there is NO Stream Sequence Number assigned to this DATA chunk. Therefore, the receiver MUST ignore the Stream Sequence Number field. Stewart, et al [Page 31] Internet Draft Simple Control Transmission Protocol March 2000 After re-assembly (if necessary), unordered data chunks MUST be dispatched to the upper layer by the receiver without any attempt of re-ordering. Note, if an unordered user message is segmented, each segment of the message MUST have its U bit set to 1. B bit: 1 bit The (B)eginning segment bit, if set, indicates the first segment of a user message. E bit: 1 bit The (E)nding segment bit, if set, indicates the last segment of a user message. A non-segmented user message shall have both the B and E bits set to 1. Setting both B and E bits to 0 indicates a middle segment of a multi-segment user message, as summarized in the following table: B E Description ============================================================ | 1 0 | First piece of a segmented user message | +----------------------------------------------------------+ | 0 0 | Middle piece of a segmented user message | +----------------------------------------------------------+ | 0 1 | Last piece of a segmented user message | +----------------------------------------------------------+ | 1 1 | Un-segmented Message | ============================================================ Length: 16 bits (16 bit u_int) This field indicates the length of the DATA chunk in octets. It includes the Type field, the Reserved field, the U and B/E bits, the Length field, TSN, the Stream Identifier, the Stream Sequence Number, and the User Data fields. It does not include any padding. TSN : 32 bits (32 bit u_int) This value represents the TSN for this DATA chunk. The valid range of TSN is from 0x0 to 0xffffffff. Stream Identifier S: 16 bit u_int Identifies the stream to which the following user data belongs. Stream Sequence Number n: 16 bit u_int This value presents the stream sequence number of the following user data within the stream S. Valid range is 0x0 to 0xFFFF. Note, when a user message is segmented by SCTP for transport, the same stream sequence number MUST be carried in each of the segments of the message. Stewart, et al [Page 32] Internet Draft Simple Control Transmission Protocol March 2000 Payload Protocol Identifier: 32 bits (32 bit u_int) This value represents an application (or upper layer) specified protocol identifier. This value is passed to SCTP by its upper layer and sent to its peer. This identifier is not used by SCTP but may be used by certain network entities as well as the peer application to identify the type of information being carried in this DATA chunk. The value 0x0 indicates no application identifier is specified by the upper layer for this payload data. User Data: variable length This is the payload user data. The implementation MUST pad the end of the data to a 32 bit boundary with 0 octets. Any padding MUST NOT be included in the length field. 2.4 Vendor-Specific Chunk Extensions This Chunk type is available to allow vendors to support their own extended data formats not defined by the IETF. It MUST not affect the operation of SCTP. In particular, when adding a Vendor Specific chunk type, the vendor defined chunks MUST obey the congestion avoidance rules defined in this document if they carry user data. User data is defined as any data transported over the association that is delivered to the upper layer of the receiver. Endpoints not equipped to interpret the vendor-specific chunk sent by a remote endpoint MUST ignore it. Endpoints that do not receive desired vendor specific information SHOULD make an attempt to operate without it, although they may do so (and report they are doing so) in a degraded mode. A summary of the Vendor-Specific Chunk format is shown below. The fields are transmitted from left to right. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Flags | Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Vendor-Id | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Value / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Type: 8 bit u_int 0xFE for all Vendor-Specific chunks. Stewart, et al [Page 33] Internet Draft Simple Control Transmission Protocol March 2000 Flags: 8 bit u_int Vendor specific flags. Length: 16 bit u_int Size of this Vendor-Specific chunks in octets, including the Type, Flags, Length, Vendor-Id, and Value fields. Vendor-Id: 32 bit u_int The high-order octet is 0 and the low-order 3 octets are the SMI Network Management Private Enterprise Code of the Vendor in network byte order, as defined in the Assigned Numbers (RFC 1700). Value: Variable length The Value field is one or more octets. The actual format of the information is site or application specific, and a robust implementation SHOULD support the field as undistinguished octets. The codification of the range of allowed usage of this field is outside the scope of this specification. 3. SCTP Association State Diagram During the lifetime of an SCTP association, the SCTP endpoints progress from one state to another in response to various events. The events that may potentially advance an endpoint's state include: o SCTP user primitive calls, e.g., [ASSOCIATE], [TERMINATE], [ABORT], o reception of INIT, COOKIE, ABORT, SHUTDOWN, etc. control chunks, or o some timeout events. The state diagram in the figures below illustrates state changes, together with the causing events and resulting actions. Note that some of the error conditions are not shown in the state diagram. Full description of all special cases should be found in the text. Note, chunk names are given in all capital letters, while parameter names have the first letter capitalized, e.g., COOKIE chunk type vs. Cookie parameter. Stewart, et al [Page 34] Internet Draft Simple Control Transmission Protocol March 2000 ----- -------- (frm any state) / \ / rcv ABORT [ABORT] rcv INIT | | | ---------- or ---------- --------------- | v v delete TCB snd ABORT generate Cookie \ +---------+ delete TCB snd INIT.ACK ---| CLOSED | +---------+ / \ [ASSOCIATE] / \ --------------- | | create TCB | | snd INIT | | strt init timer rcv valid COOKIE | v (1) ---------------- | +------------+ create TCB | | COOKIE_WAIT| (2) snd COOKIE.ACK | +------------+ | | | | rcv INIT.ACK | | ----------------- | | snd COOKIE | | stop init timer | | strt cookie timer | v | +------------+ | | COOKIE_SENT| (3) | +------------+ | | | | rcv COOKIE.ACK | | ----------------- | | stop cookie timer v v +---------------+ | ESTABLISHED | +---------------+ (from the ESTABLISHED state only) | | /--------+--------\ [TERMINATE] / \ ----------------- | | check outstanding | | data chunks | | v | +---------+ | |SHUTDOWN | | rcv SHUTDOWN |PENDING | | ---------------- +---------+ | x | | No more outstanding | | ------------------- | | snd SHUTDOWN | | strt shutdown timer | | v v Stewart, et al [Page 35] Internet Draft Simple Control Transmission Protocol March 2000 +---------+ +-----------+ (4) |SHUTDOWN | | SHUTDOWN | (5) |SENT | | RECEIVED | +---------+ +-----------+ | | rcv SHUTDOWN.ACK | | x ------------------- | |----------------- stop shutdown timer | | retransmit missing DATA delete TCB | | send SHUTDOWN.ACK | | delete TCB | | \ +---------+ / \-->| CLOSED |<--/ +---------+ Note: (1) If the received COOKIE is invalid (i.e., failed to pass the authentication check), the receiver MUST silently discard the datagram. Or, if the received COOKIE is expired (see Section 4.1.5), the receiver SHALL send back an ERROR chunk. In either case, the receiver stays in the CLOSED state. (2) If the init timer expires, the endpoint SHALL retransmit INIT and re-start the init timer without changing state. This SHALL be repeated up to 'Max.Init.Retransmits' times. After that, the endpoint SHALL abort the initialization process and report the error to SCTP user. (3) If the cookie timer expires, the endpoint SHALL retransmit COOKIE and re-start the cookie timer without changing state. This SHALL be repeated up to 'Max.Init.Retransmits' times. After that, the endpoint SHALL abort the initialization process and report the error to SCTP user. (4) In SHUTDOWN-SENT state the endpoint SHALL acknowledge any received DATA chunks without delay (5) In SHUTDOWN-RECEIVED state, the endpoint MUST NOT accept any new send request from its SCTP user. 4. Association Initialization Before the first data transmission can take place from one SCTP endpoint ("A") to another SCTP endpoint ("Z"), the two endpoints must complete an initialization process in order to set up an SCTP association between them. The SCTP user at an endpoint should use the ASSOCIATE primitive to initialize an SCTP association to another SCTP endpoint. Stewart, et al [Page 36] Internet Draft Simple Control Transmission Protocol March 2000 IMPLEMENTATION NOTE: From an SCTP-user's point of view, an association may be implicitly opened, without an ASSOCIATE primitive (see 9.1 B) being invoked, by the initiating endpoint's sending of the first user data to the destination endpoint. The initiating SCTP will assume default values for all mandatory and optional parameters for the INIT/INIT ACK. Once the association is established, unidirectional streams will be open for data transfer on both ends (see Section 4.1.1). 4.1 Normal Establishment of an Association The initialization process consists of the following steps (assuming that SCTP endpoint "A" tries to set up an association with SCTP endpoint "Z" and "Z" accepts the new association): A) "A" shall first send an INIT message to "Z". In the INIT, "A" must provide its security tag "Tag_A" in the Initiate Tag field. Tag_A SHOULD be a random number in the range of 0x1 to 0xffffffff (see 4.3.1 for Tag value selection). After sending the INIT, "A" starts the T1-init timer and enters the COOKIE-WAIT state. B) "Z" shall respond immediately with an INIT ACK message. In the message, besides filling in other parameters, "Z" must set the Verification Tag field to Tag_A, and also provide its own security tag "Tag_Z" in the Initiate Tag field. Moreover, "Z" MUST generate and send along with the INIT ACK an State Cookie. See Section 4.1.3 for State Cookie generation. Note: after sending out INIT ACK with the cookie, "Z" MUST not allocate any resources, nor keep any states for the new association. Otherwise, "Z" will be vulnerable to resource attacks. C) Upon reception of the INIT ACK from "Z", "A" shall stop the T1-init timer and leave COOKIE-WAIT state. "A" shall then send the cookie received in the INIT ACK message in a cookie chunk, restart the T1-init timer, and enter the COOKIE-SENT state. Note, the cookie chunk can be bundled with any pending outbound DATA chunks, but it MUST be the first chunk in the datagram AND until the COOKIE ACK is returned the sender MUST NOT send any other datagrams to the peer. D) Upon reception of the COOKIE chunk, Endpoint "Z" will reply with a COOKIE ACK chunk after building a TCB and marking itself to the ESTABLISHED state. A COOKIE ACK chunk may be combined with any pending DATA chunks (and/or SACK chunks), but the COOKIE ACK chunk MUST be the first chunk in the datagram. IMPLEMENTATION NOTE: an implementation may choose to send the Communication Up notification to the SCTP user upon reception of a valid COOKIE. Stewart, et al [Page 37] Internet Draft Simple Control Transmission Protocol March 2000 E) Upon reception of the COOKIE ACK, endpoint "A" will move from the COOKIE-SENT state to the ESTABLISHED state, stopping the T1-init timer, and it may also notify its ULP about the successful establishment of the associate with a Communication Up notification (see Section 9). Note: A DATA chunk MUST NOT be carried in the INIT or INIT ACK message. Note: T1-init timer shall follow the same rules given in Section 5.3. Note: if an endpoint receives an INIT, INIT ACK, or COOKIE chunk but decides not to establish the new association due to missing mandatory parameters in the received INIT or INIT ACK, invalid parameter values, or, lack of local resources, it SHALL respond with an ABORT chunk. It SHOULD also specify the cause of abort, such as the type of the missing mandatory parameters, etc., by either including cause parameters or bundling with the ABORT one or more Operational ERROR chunks. The Verification Tag field in the common header of the outbound abort datagram MUST be set to equal the Initiate Tag value of the peer. Note: After the reception of the first data chunk in an association the receiver MUST immediately respond with a SACK to acknowledge the data chunk, subsequent acknowledgments should be done as described in section 5.2. Note: When an SCTP endpoint sends an INIT or INIT ACK it SHOULD include all of its transport addresses in the parameter section. This is because it may NOT be possible to control the "sending" address that a receiver of an SCTP datagram sees. A receiver thus MUST know every address that may be a source address for a peer SCTP endpoint, this assures that the inbound SCTP datagram can be matched to the proper association. Note: At the time when the TCB is created, either end MUST set its internal cumulative TSN acknowledgment point to its peer's Initial TSN minus one. IMPLEMENTATION Note: The IP address and SCTP port(s) are generally used as the key to find the TCB within an SCTP instance. 4.1.1 Handle Stream Parameters In the INIT and INIT ACK messages, the sender of the message shall indicate the number of outbound streams (OS) it wishes to have in the association, as well as the maximal inbound streams (MIS) it will accept from the other endpoint. After receiving these stream configuration information from the other side, each endpoint shall perform the following check: if the peer's MIS is less than the endpoint's OS, meaning that the peer is incapable of supporting all the outbound streams the endpoint wants to configure, the endpoint MUST either settle with MIS outbound streams, or abort the association and report to its upper layer the resources shortage at its peer. Stewart, et al [Page 38] Internet Draft Simple Control Transmission Protocol March 2000 After the association is initialized, the valid outbound stream identifier range for either endpoint shall be 0 to min(local OS, remote MIS)-1. 4.1.2 Handle Address Parameters During the association initialization, an endpoint shall use the following rules to discover and collect the destination transport address(es) of its peer. On reception of an INIT or INIT ACK message, the receiver shall record any transport addresses. The transport address(es) are derived by the combination of SCTP source port (from the common header) and the IP address parameter(s) carried in the INIT or INIT ACK message. The receiver should use only these transport addresses as destination transport addresses when sending subsequent datagrams to its peer. If no IP address parameters are specified in the INIT or INIT ACK message, then the source IP address from which the message arrives should be combined with SCTP source port number and be considered as the only destination transport address to use. An initial primary destination transport address shall be selected for either endpoint, using the following rules: For the initiator: any valid transport address obtained from the INIT ACK message. If no transport address is specified in the INIT ACK message, use the source transport address from which the INIT ACK message arrived. For the responder: any valid transport address obtained from the INIT message. If no transport address is specified in the INIT message, use the source transport address from which the INIT message arrived. 4.1.3 Generating State Cookie When sending an INIT ACK as a response to an INIT message, the sender of INIT ACK should create an State Cookie and send it as part of the INIT ACK. Inside this State Cookie, the sender should include a ICV security signature or MAC (message Authentication code) [4], a time stamp on when the cookie is created, and the lifespan of the cookie, along with all the information necessary for it to establish the association. The following steps SHOULD be taken to generate the cookie: 1) create an association TCB using information from both the received INIT and the outgoing INIT ACK messages, 2) in the TCB, set the creation time to the current time of day, and the lifespan to the protocol parameter 'Valid.Cookie.Life', 3) Generate a MAC signature using the TCB and a Private Key (see [4] for details on generating the MAC), and Stewart, et al [Page 39] Internet Draft Simple Control Transmission Protocol March 2000 4) generate the State Cookie by combining the TCB and the resultant ICV signature. After sending the INIT ACK with the cookie, the sender SHOULD delete the TCB and any other local resource related to the new association, so as to prevent resource attacks. The ICV and hashing method used to generate the MAC is strictly a private matter for the receiver of the INIT message. The use of a MAC is mandatory to prevent denial of service attacks. The Private Key MUST be random per RFC1750 [1]; it SHOULD be changed reasonably frequently, and the timestamp in the cookie MAY be used to determine which key should be used to verify the MAC. 4.1.4 Cookie Processing When an endpoint receives an INIT ACK chunk in response to its INIT chunk, and the INIT ACK contains an State Cookie parameter, it MUST immediately send a COOKIE chunk to its peer with the received cookie. The sender MAY also add any pending DATA chunks to the message. The sender shall also start the cookie timer after sending out the COOKIE chunk. If the timer expires, the sender shall retransmit the COOKIE chunk and restart the cookie timer. This is repeated until either a COOKIE ACK is received or the endpoint is marked unreachable (and thus the association enters the CLOSED state). 4.1.5 Cookie Authentication When an endpoint receives a COOKIE chunk from another endpoint with which it has no association, it shall take the following actions: 1) compute a MAC signature using the TCB data carried in the cookie and the Private Key (note the timestamp in the cookie MAY be used to determine which Private Key to use) reference [4] SHOULD be used has a guideline for generating the MAC, 2) authenticate the cookie as one that it previously generated by comparing the computed MAC signature against the one carried in the cookie. If this comparison fails, the datagram, including the COOKIE and the attached user data, should be silently discarded, 3) compare the creation time stamp in the cookie to the current local time, if the elapsed time is longer than the lifespan carried in the cookie, then the datagram, including the COOKIE and the attached user data, SHOULD be discarded and the endpoint MUST transmit a stale cookie operational error to the sending endpoint, 4) if the cookie is valid, create an association to the sender of the COOKIE message with the information in the TCB data carried in the COOKIE, and enter the ESTABLISHED state, 5) immediately acknowledge any DATA chunk in the datagram with a SACK (subsequent datagram acknowledgement should following the rules defined in Section 5.2), and, Stewart, et al [Page 40] Internet Draft Simple Control Transmission Protocol March 2000 6) send a COOKIE ACK chunk to the sender acknowledging reception of the cookie. The COOKIE ACK MAY be piggybacked with any outbound DATA chunk or SACK chunk. Note that if a COOKIE is received from an endpoint with which the receiver of the COOKIE has an existing association, the procedures in section 4.2 should be followed. 4.1.6 An Example of Normal Association Establishment In the following example, "A" initiates the association and then sends a user message to "Z", then "Z" sends two user messages to "A" later (assuming no bundling or segmentation occurs): Endpoint A Endpoint Z {app sets association with Z} (build TCB) INIT [INIT Tag=Tag_A & other info] --------\ (Start T1-init timer) \ (Enter COOKIE-WAIT state) \---> (compose temp TCB and Cookie_Z) /--- INIT ACK [Veri Tag=Tag_A, / INIT Tag=Tag_Z, (Cancel T1-init timer) <------/ Cookie_Z, & other info] (destroy temp TCB) COOKIE [Cookie_Z] -----------\ (Start T1-init timer) \ (Enter COOKIE-SENT state) \---> (build TCB enter ESTABLISHED state) /---- COOKIE-ACK / (Cancel T1-init timer, <-----/ Enter established state) ... {app sends 1st user data; strm 0} DATA [TSN=initial TSN_A Strm=0,Seq=1 & user data]--\ (Start T3-rxt timer) \ \-> /----- SACK [TSN ACK=init TSN_A,Frag=0] (Cancel T3-rxt timer) <------/ ... Stewart, et al [Page 41] Internet Draft Simple Control Transmission Protocol March 2000 ... {app sends 2 datagrams;strm 0} /---- DATA / [TSN=init TSN_Z <--/ Strm=0,Seq=1 & user data 1] SACK [TSN ACK=init TSN_Z, /---- DATA Frag=0] --------\ / [TSN=init TSN_Z +1, \/ Strm=0,Seq=2 & user data 2] <------/\ \ \------> Note that If T1-init timer expires at "A" after the INIT or COOKIE chunks are sent, the same INIT or cookie chunk with the same Initiate Tag (i.e., Tag_A) or cookie shall be retransmitted and the timer restarted. This shall be repeated Max.Init.Retransmits times before "A" considers "Z" unreachable and reports the failure to its upper layer (and thus the association enters the CLOSED state). When retransmitting the INIT, the endpoint SHALL following the rules defined in 5.3 to determine the proper timer value. 4.2 Handle Duplicate INIT, INIT ACK, COOKIE, and COOKIE ACK During the life time of an association (in one of the possible states), an endpoint may receive from its peer endpoint one of the setup chunks (INIT, INIT ACK, COOKIE, and COOKIE ACK). The receiver shall treat such a setup chuck as a duplicate and process it as described in this section. The following scenarios can cause duplicated chunks: A) The peer has crashed without being detected, and re-started itself and sent out a new INIT Chunk trying to restore the association, B) Both sides are trying to initialize the association at about the same time, C) The chunk is from a staled datagram that was used to establish the present association or a past association which is no longer in existence, D) The chunk is a false message generated by an attacker, or E) The peer never received the COOKIE ACK and is retransmitting its COOKIE. In case A), the endpoint shall reset the present association and set a new association with its peer. Case B) is unique and is discussed in Section 4.2.1. However, in cases C), D) and E), the endpoint must retain the present association. The rules in the following sections shall be applied in order to identify and correctly handle these cases. Stewart, et al [Page 42] Internet Draft Simple Control Transmission Protocol March 2000 4.2.1 Handle Duplicate INIT in COOKIE-WAIT or COOKIE-SENT State This usually indicates an initialization collision, i.e., both endpoints are attempting at about the same time to establish an association with the other endpoint. In such a case, each of the two side shall respond to the other side with an INIT ACK, with the Verification Tag field of the common header set to the tag value received from the INIT message, and the Initiate Tag field set to its own tag value (the same tag used in the INIT message sent out by itself). Each responder shall also generate a cookie with the INIT ACK. After that, no other actions shall be taken by either side, i.e., the endpoint shall not change its state, and the T1-init timer shall be left running. The normal procedures for handling cookies will resolve the duplicate INITs to a single association. 4.2.2 Handle Duplicate INIT in Other States Upon reception of the duplicated INIT, the receiver shall generate an INIT ACK with an State Cookie. In the outbound INIT ACK, the endpoint shall set the Verification Tag field in the common header to the peer's new tag value (from the duplicated INIT message), and the Initiate Tag field to its own tag value (unchanged from the existing association). The included State Cookie shall be generated using the current time and a temporary TCB constructed with the information provided in the duplicated INIT message (see Section 4.1.3). This temporary TCB MUST be destroyed after the outbound INIT ACK is built. After sending out the INIT ACK, the endpoint shall take no further actions, i.e., the existing association, including its current state, and the corresponding TCB MUST not be changed. 4.2.3 Handle Duplicate INIT ACK If an INIT ACK is received by an endpoint in any state other than the COOKIE-WAIT state, the endpoint should discard the INIT ACK message. A duplicate INIT ACK usually indicates the processing of an old INIT or duplicated INIT message. 4.2.4 Handle Duplicate Cookie When a duplicated COOKIE chunk is received in any state for an existing association the following rules shall be applied: 1) compute an MD5 signature using the TCB data carried in the cookie along with the receiver's private security key, Stewart, et al [Page 43] Internet Draft Simple Control Transmission Protocol March 2000 2) authenticate the cookie by comparing the computed MD5 signature against the one carried in the cookie. If this comparison fails, the datagram, including the COOKIE and the attached user data, should be silently discarded (this is case C or D above). 3) compare the timestamp in the cookie to the current time, if the cookie is older than the lifespan carried in the cookie, the datagram, including the COOKIE and the attached user data, should be discarded and the endpoint MUST transmit a stale cookie error to the sending endpoint only if the Verification tags of the cookie's TCB does NOT match the current tag values in the association (this is case C or D above). If both Verification tags do match consider the cookie valid (this is case E). 4) If the cookie proves to be valid, unpack the TCB into a temporary TCB. 5) If the Verification Tags in the Temporary TCB matches the Verification Tags in the existing TCB, the cookie is a duplicate cookie. A cookie ack should be sent to the peer endpoint but NO update should be made to the existing TCB. 6) If the the local Verification Tag in the temporary TCB does not match the local Verification Tag in the existing TCB, then the cookie is an old stale cookie and does not correspond to the existing association (case C above). The datagram should be silently discarded. 7) If the peer's Verification Tag in the temporary TCB does not match the peer's Verification Tag in the existing TCB, then a restart of the peer has occurred (case A above). In such a case, the endpoint should report the restart to its ULP and respond the peer with a COOKIE ACK message. It shall also update the Verification Tag, initial TSN, and the destination address list of the existing TCB with the information from the temporary TCB. After that the temporary TCB can be discarded. Furthermore, all the congestion control parameters (e.g., cwnd, ssthresh) related to this peer shall be reset to their initial values (see Section 6.2.1). IMPLEMENTATION NOTE: It is an implementation decision on how to handle any pending datagrams. The implementation may elect to either A) send all messages back to its upper layer with the restart report, or B) automatically re-queue any datagrams pending by marking all of them as never-sent and assigning new TSN's at the time of their initial transmissions based upon the updated starting TSN (as defined in section 5). Note: The "peer's Verification Tag" is the tag received in the INIT or INIT ACK chunk. Stewart, et al [Page 44] Internet Draft Simple Control Transmission Protocol March 2000 4.2.5 Handle Duplicate COOKIE-ACK. At any state other than COOKIE-SENT, an endpoint may receive a duplicated COOKIE ACK chunk. If so, the chunk should be silently discarded. 4.2.6 Handle Stale COOKIE Error A stale cookie error indicates one of a number of possible events: A) that the association failed to completely setup before the cookie issued by the sender was processed. B) an old cookie was processed after setup completed. C) an old cookie is received from someone that the receiver is not interested in having an association with and the ABORT message was lost. When processing a stale cookie an endpoint should first examine if an association is in the process of being setup, i.e. the association is in the COOKIE-SENT state. In all cases if the association is NOT in the COOKIE-SENT state, the stale cookie message should be silently discarded. If the association is in the COOKIE-SENT state, the endpoint may elect one of the following three alternatives. 1) Send a new INIT message to the endpoint, to generate a new cookie and re-attempt the setup procedure. 2) Discard the TCB and report to the upper layer the inability of setting-up the association. 3) Send a new INIT message to the endpoint, adding a cookie preservative parameter requesting an extension on the life time of the cookie. When calculating the time extension, an implementation SHOULD use the RTT information measured based on the previous COOKIE / Stale COOKIE message exchange, and should add no more than 1 second beyond the measured RTT, due to a long cookie life time makes the endpoint more subject to a replay attack. 4.3 Other Initialization Issues 4.3.1 Selection of Tag Value Initiate Tag values should be selected from the range of 0x1 to 0xffffffff. It is very important that the Tag value be randomized to help protect against "man in the middle" and "sequence number" attacks. It is suggested that RFC 1750 [1] be used for the Tag randomization. Stewart, et al [Page 45] Internet Draft Simple Control Transmission Protocol March 2000 Moreover, the tag value used by either endpoint in a given association MUST never be changed during the lifetime of the association. However, a new tag value MUST be used each time the endpoint tears-down and then re-establishes the association to the same peer. 5. User Data Transfer For transmission efficiency, SCTP defines mechanisms for bundling of small user messages and segmentation of large user messages. The following diagram depicts the flow of user messages through SCTP. +--------------------------+ | User Messages | +--------------------------+ SCTP user ^ | ==================|==|======================================= | v (1) +------------------+ +--------------------+ | SCTP DATA Chunks | |SCTP Control Chunks | +------------------+ +--------------------+ ^ | ^ | | v (2) | v (2) +--------------------------+ | SCTP datagrams | +--------------------------+ SCTP ^ | ===========================|==|=========================== | v Unreliable Packet Transfer Service (e.g., IP) Note: (1) When converting user messages into Data chunks, SCTP sender will segment user messages larger than the current path MTU into multiple data chunks. The segmented message will normally be reassembled from data chunks before delivery to the user by the SCTP receiver (see Section 5.9 for details). (2) Multiple data and control chunks may be multiplexed by the sender into a single SCTP datagram for transmission, as long as the final size of the datagram does not exceed the current path MTU. The receiver will de-multiplex the datagram back into the original chunks. The segmentation and bundling mechanisms, as detailed in Sections 5.9 and 5.10, are optional to implement by the data sender, but they MUST be implemented by the data receiver, i.e., an SCTP receiver MUST be prepared to receive and process bundled or segmented data. Stewart, et al [Page 46] Internet Draft Simple Control Transmission Protocol March 2000 5.1 Transmission of DATA Chunks The following general rules SHALL be applied by the sender for transmission and/or retransmission of outbound DATA chunks: A) At any given time, the sender MUST NOT transmit new data onto any destination transport address if its peer's rwnd indicates that the peer has no buffer space (i.e. rwnd is 0, see Section 5.2.1). However, regardless of the value of rwnd (including if it is 0), the sender can always have ONE data packet in flight to the receiver if allowed by cwnd (see rule B below). This rule allows the sender to probe for a change in rwnd that the sender missed due to the update having been lost in transmission from the receiver to the sender. B) At any given time, the sender MUST NOT transmit new data onto a given transport address if it has cwnd or more octets of data outstanding on that transport address. C) When the time comes for the sender to transmit, before sending new DATA chunks, the sender MUST first transmit any outstanding DATA chunks which are marked for retransmission (limited by the current cwnd). D) Then, the sender can send out as many new DATA chunks as Rule A and Rule B above allow. Note: multiple DATA chunks committed for transmission MAY be bundled in a single packet, unless bundling is explicitly disallowed by ULP of the data sender. Furthermore, DATA chunks being retransmitted MAY be bundled with new DATA chunks, as long as the resulting packet size does not exceed the path MTU. Note: before a sender transmits a data packet, if any received DATA chunks have not been acknowledged (e.g., due to delayed ack), the sender should create a SACK and bundle it with the outbound DATA chunk, as long as the size of the final SCTP datagram does not exceed the current MTU. See Section 5.2. IMPLEMENTATION Note: when the window is full (i.e., transmission is disallowed by Rule A and/or Rule B), the sender MAY still accept send requests from its upper layer, but SHALL transmit no more DATA chunks until some or all of the outstanding DATA chunks are acknowledged and transmission is allowed by Rule A and Rule B again. Whenever a transmission or retransmission is made to any address, if the T3-rxt timer of that address is not currently running, the sender MUST start that timer. However, if the timer of that address is already running, the sender SHALL restart the timer ONLY IF the earliest (i.e., lowest TSN) outstanding DATA chunk sent to that address is being retransmitted. Stewart, et al [Page 47] Internet Draft Simple Control Transmission Protocol March 2000 When starting or restarting the T3-rxt timer, the timer value must be adjusted according to the timer rules defined in Sections 5.3.2, and 5.3.3. 5.2 Acknowledgment on Reception of DATA Chunks The SCTP receiver MUST always acknowledge the SCTP sender about the reception of each DATA chunk. The guidelines on delayed acknowledgment algorithm specified in Section 4.2 of RFC 2581 [3] SHOULD be followed. Specifically, an acknowledgment SHOULD be generated for at least every second datagram received, and SHOULD be generated within 200 ms of the arrival of any unacknowledged datagram. IMPLEMENTATION NOTE: the maximal delay for generating an acknowledgment may be configured by the SCTP user, either statically or dynamically, in order to meet the specific timing requirement of the signaling protocol being carried. Acknowledgments MUST be sent in SACK control chunks. A SACK chunk can acknowledge the reception of multiple DATA chunks. See Section 2.3.3 for SACK chunk format. In particular, the SCTP receiver MUST fill in the Cumulative TSN ACK field to indicate the latest cumulative TSN number it has received, and any received segments beyond the Cumulative TSN SHALL also be reported. Upon reception of the SACK, the data sender MUST adjust its total outstanding data count and the outstanding data count on those destination addresses for which one or more data chunks is acknowledged by the SACK. Note: When a datagram arrives with duplicate DATA chunk(s) and no new DATA chunk(s), the receiver MUST immediately send a SACK with no delay. Normally this will occur when the original SACK was lost, and the peers RTO has expired. The duplicate TSN number(s) SHOULD be reported in the SACK as duplicate. When a receiver prepares a SACK, any duplicate DATA chunks received SHOULD be reported in the SACK. When a SACK is received the receiver MAY use the Duplicate TSN information to determine if SACK loss is occuring. Further use of this data is for future study. The following example illustrates the use of delayed acknowledgments: Endpoint A Endpoint Z {App sends 3 messages; strm 0} DATA [TSN=7,Strm=0,Seq=3] ------------> (ack delayed) (Start T3-rxt timer) Stewart, et al [Page 48] Internet Draft Simple Control Transmission Protocol March 2000 DATA [TSN=8,Strm=0,Seq=4] ------------> (send ack) /------- SACK [TSN ACK=8,Frag=0] (cancel T3-rxt timer) <-----/ ... ... DATA [TSN=9,Strm=0,Seq=5] ------------> (ack delayed) (Start T3-rxt timer) ... {App sends 1 message; strm 1} (bundle SACK with DATA) /----- SACK [TSN Ack=9,Frag=0] \ / DATA [TSN=6,Strm=1,Seq=2] (cancel T3-rxt timer) <------/ (Start T3-rxt timer) (ack delayed) ... (send ack) SACK [TSN ACK=6,Frag=0] -------------> (cancel T3-rxt timer) 5.2.1 Tracking Peer's Receive Buffer Space Whenever a SACK arrives, a new updated a_rwnd arrives with it. This value represents the amount of buffer space the sender of the SACK, at the time of transmitting the SACK, has left of its total receive buffer space (as specified in the INIT/INIT-ACK). After processing the SACK, the receiver of the SACK must use the following rules to re-calculate the rwnd, using the received a_rwnd value. A) At the establishment of the association, the endpoint initializes the rwnd to the Advertised Receiver Window Credit (a_rwnd) the peer specified in the INIT or INIT ACK. B) Any time a DATA chunk is transmitted to a peer, the endpoint subtracts the data size of the chunk from the rwnd of that peer. C) Any time a SACK arrives, the endpoint performs the following: If all outstanding TSNs are acknowledged by the SACK, adopt the a_rwnd value in the SACK as the new rwnd. Otherwise, take the value of the current rwnd, and add to it the data size of any newly acknowledged TSNs that has its BE bits set to 11, OR that moved the cumulative TSN point forward. Then, set the rwnd to the lesser of the calculated value and the a_rwnd carried in the SACK. D) Any time the T3-rxt timer expires on any address, causing all outstanding chunks sent to that address to be marked for retransmission, add all of the data sizes of those chunks to the rwnd. Stewart, et al [Page 49] Internet Draft Simple Control Transmission Protocol March 2000 E) Any time a DATA chunk is marked for retransmission via the fast retransmit algorithm (section 6.2.4), add the DATA chunks size to the rwnd. 5.3 Management of Retransmission Timer SCTP uses a retransmission timer T3-rxt to ensure data delivery in the absence of any feedback from the remote data receiver. The duration of this timer is referred to as RTO (retransmission timeout). When the receiver endpoint is multi-homed, the data sender endpoint will calculate a separate RTO for each different destination transport addresses of the receiver endpoint. The computation and management of RTO in SCTP follows closely with how TCP manages its retransmission timer. To compute the current RTO, an SCTP sender maintains two state variables per destination transport address: SRTT (smoothed round-trip time) and RTTVAR (round-trip time variation). 5.3.1 RTO Calculation The rules governing the computation of SRTT, RTTVAR, and RTO are as follows: C1) Until an RTT measurement has been made for a packet sent to the given destination transport address, set RTO to the protocol parameter 'RTO.Initial'. C2) When the first RTT measurement R is made, set SRTT <- R, RTTVAR <- R/2, and RTO <- SRTT + 4 * RTTVAR. C3) When a new RTT measurement R' is made, set RTTVAR <- (1 - RTO.Beta) * RTTVAR + RTO.Beta * |SRTT - R'| SRTT <- (1 - RTO.Alpha) * SRTT + RTO.Alpha * R' Note, the value of SRTT used in the update to RTTVAR is its value *before* updating SRTT itself using the second assignment. After the computation, update RTO <- SRTT + 4 * RTTVAR. Stewart, et al [Page 50] Internet Draft Simple Control Transmission Protocol March 2000 C4) When data is in flight and when allowed by rule C5 below, a new RTT measurement MUST be made each round trip. Furthermore, it is RECOMMENDED that new RTT measurements should be made no more than once per round-trip for a given destination transport address. There are two reasons for this recommendation: first, it appears that measuring more frequently often does not in practice yield any significant benefit [5]; second, if measurements are made more often, then the values of RTO.Alpha and RTO.Beta in rule C3 above should be adjusted so that SRTT and RTTVAR still adjust to changes at roughly the same rate (in terms of how many round trips it takes them to reflect new value) as they would if making only one measurement per round-trip and using RTO.Alpha and RTO.Beta as given in rule C3. However, the exact nature of these adjustments remains a research issue. C5) Karn's algorithm: RTT measurements MUST NOT be made using packets that were retransmitted (and thus for which it is ambiguous whether the reply was for the first instance of the packet or a later instance). C6) Whenever RTO is computed, if it is less than RTO.Min seconds then it is rounded up to RTO.Min seconds. The reason for this rule is that RTOs that do not have a high minimum value are susceptible to unnecessary timeouts [5]. C7) A maximum value may be placed on RTO provided it is at least RTO.max seconds. There is no requirement for the clock granularity G used for computing RTT measurements and the different state variables, other than G1) Whenever RTTVAR is computed, if RTTVAR = 0, then adjust RTTVAR <- G. Experience [5] has shown that finer clock granularities (<= 100 msec) perform somewhat better than more coarse granularities. 5.3.2 Retransmission Timer Rules The rules for managing the retransmission timer are as follows: R1) Every time a packet containing data is sent to any address (including a retransmission), if the T3-rxt timer of that address is not running, start it running so that it will expire after the RTO of that address. The RTO used here is that obtained after any doubling due to previous T3-rxt timer expirations on the corresponding destination address as discussed in rule E2 below. R2) Whenever all outstanding data on an address has been acknowledged, turn off the T3-rxt timer of that address. Stewart, et al [Page 51] Internet Draft Simple Control Transmission Protocol March 2000 R3) Whenever a SACK is received that acknowledges new data chunks including the one with the earliest outstanding TSN on that address, restart T3-rxt timer of that address with its current RTO. The following example shows the use of various timer rules (assuming the receiver uses delayed acks). Endpoint A Endpoint Z {App begins to send} Data [TSN=7,Strm=0,Seq=3] ------------> (ack delayed) (Start T3-rxt timer) {App sends 1 message; strm 1} (bundle ack with data) DATA [TSN=8,Strm=0,Seq=4] ----\ /-- SACK [TSN ACK=7,Frag=0] \ \ / DATA [TSN=6,Strm=1,Seq=2] \ / (Start T3-rxt timer) \ / \ (Re-start T3-rxt timer) <------/ \--> (ack delayed) (ack delayed) ... {send ack} SACK [TSN ACK=6,Frag=0] --------------> (Cancel T3-rxt timer) .. (send ack) (Cancel T3-rxt timer) <-------------- SACK [TSN ACK=8,Frag=0] 5.3.3 Handle T3-rxt Expiration Whenever the retransmission timer T3-rxt expires on a destination address, do the following: E1) On the destination address where the timer expires, adjust its ssthresh with rules defined in Section 6.2.3 and set the cwnd <- MTU. E2) On the destination address where the timer expires, set RTO <- RTO * 2 ("back off the timer"). The maximum value discussed in rule C7 above (RTO.max) may be used to provide an upper bound to this doubling operation. E3) Determine how many of the earliest (i.e., lowest TSN) outstanding Data chunks on the address where the T3-rxt has expired that will fit into a single packet, subject to the MTU constraint for the path corresponding to the destination transport address where the retransmission is being sent to (this may be different from the address where the timer expires [see Section 5.4]). Call this value K. Bundle and retransmit those K data chunks in a single packet to the address. Stewart, et al [Page 52] Internet Draft Simple Control Transmission Protocol March 2000 E4) Start the retransmission timer T3-rxt on the destination address to where the retransmission is sent, if rule R1 above indicates to do so. Note, the RTO to be used for starting T3-rxt should be the one of the destination address to where the retransmission is sent, which, when the receiver is multi-homed, may be different from the destination address where the timer expired (see Section 5.4 below). Note that after retransmitting, once a new RTT measurement is obtained (which can happen only when new data has been sent and acknowledged, per rule C5, or for a measurement made from a Heartbeat [see Section 7.3]), the computation in rule C3 is performed, including the computation of RTO, which may result in "collapsing" RTO back down after it has been subject to doubling (rule E2). The final rule for managing the retransmission timer concerns failover (see Section 5.4.1): F1) Whenever SCTP switches from the current destination transport address to a different one, the current retransmission timers are left running. As soon as SCTP transmits a packet containing data to the new transport address, start the timer on that transport address, using the RTO value of the destination address where the data is being sent, if rule R1 indicates to do so. 5.4 Multi-homed SCTP Endpoints An SCTP endpoint is considered multi-homed if there are more than one transport addresses that can be used as a destination address to reach that endpoint. Moreover, at the sender side, one of the multiple destination addresses of the multi-homed receiver endpoint shall be selected as the primary destination transport address by the UPL (see Sections 4.1.2 and 9.1 for details). When the SCTP sender is transmitting to the multi-homed receiver, by default the transmission SHOULD always take place on the primary transport address, unless the SCTP user explicitly specifies the destination transport address to use. The acknowledgment SHOULD be transmitted to the same destination transport address from which the DATA or control chunk being acknowledged were received. However, when acknowledging multiple DATA chunks in a single SACK, the SACK message may be transmitted to one of the destination transport addresses from which the DATA or control chunks being acknowledged were received. Stewart, et al [Page 53] Internet Draft Simple Control Transmission Protocol March 2000 Furthermore, when the receiver is multi-homed, the SCTP data sender SHOULD try to retransmit a chunk to an active destination transport address that is different from the last destination address where the data chunk was sent to. Note, retransmissions do not affect the total outstanding data count. However, if the data chunk is retransmitted onto a different destination address, both the outstanding data counts on the new destination address and the old destination address where the data chunk was last sent to shall be adjusted accordingly. 5.4.1 Failover from Inactive Destination Address Some of the destination transport addresses of a multi-homed SCTP data receiver may become inactive due to either the occurrence of certain error conditions (see Section 7.2) or adjustments from SCTP user. When there is outbound data to send and the primary destination transport address becomes inactive (e.g., due to failures), or where the SCTP user explicitly requests to send data to an inactive destination transport address, before reporting an error to its ULP, the SCTP sender should try to send the data to an alternate active destination transport address if one exists. 5.5 Stream Identifier and Stream Sequence Number Every DATA chunk MUST carry a valid stream identifier. If a DATA chunk with an invalid stream identifier is received, the receiver shall, after acknowledging the reception of the DATA chunk following the normal procedure, respond immediately with an ERROR message with cause set to Invalid Stream Identifier (see Section 2.3.9) and discard the DATA chunk. The stream sequence number in all the streams shall start from 0x0 when the association is established. Also, when the stream sequence number reaches the value 0xffff the next stream sequence number shall be set to 0x0. 5.6 Ordered and Un-ordered Delivery By default the SCTP receiver shall ensure the DATA chunks within any given stream be delivered to the upper layer according to the order of their stream sequence number. If there are DATA chunks arriving out of order of their stream sequence number, the receiver MUST hold the received DATA chunks from delivery until they are re-ordered. However, an SCTP sender can indicate that no ordered delivery is required on a particular DATA chunk within the stream by setting the U flag of the DATA chunk to 1. Stewart, et al [Page 54] Internet Draft Simple Control Transmission Protocol March 2000 In this case, the receiver must bypass the ordering mechanism and immediately delivery the data to the upper layer (after re-assembly if the user data is segmented by the sender). This provides an effective way of transmitting "out-of-band" data in a given stream. Also, a stream can be used as an "unordered" stream by simply setting the U flag to 1 in all outbound DATA chunks sent through that stream. IMPLEMENTATION NOTE: when sending an unordered DATA chunk, an implementation may choose to place the DATA chunk in an outbound datagram that is at the head of the outbound transmission queue if possible. Note that the 'Stream Sequence Number' field in an un-ordered data chunk has no significance; the sender can fill it with arbitrary value, but the receiver MUST ignore the field. 5.7 Report Gaps in Received DATA TSNs Upon the reception of a new DATA chunk, an SCTP receiver shall examine the continuity of the TSNs received. If the receiver detects that gaps exist in the received DATA chunk sequence, an SACK with fragment reports shall be sent back immediately. Based on the segment reports from the SACK, the data sender can calculate the missing DATA chunks and make decisions on whether to retransmit them (see Section 5.3 for details). Multiple gaps can be reported in one single SACK (see Section 2.3.3). Note that when the data sender is multi-homed, the SCTP receiver SHOULD always try to send the SACK to the same network from where the last DATA chunk was received. Upon the reception of the SACK, the data sender SHALL remove all DATA chunks which have been acknowledged by the SACK. The data sender MUST also treat all the DATA chunks which fall into the gaps between the fragments reported by the SACK as "missing". The number of "missing" reports for each outstanding DATA chunk MUST be recorded by the data sender in order to make retransmission decision, see Section 6.2.4 for details. The following example shows the use of SACK to report a gap. Endpoint A Endpoint Z {App sends 3 messages; strm 0} DATA [TSN=6,Strm=0,Seq=2] ---------------> (ack delayed) (Start T3-rxt timer) Stewart, et al [Page 55] Internet Draft Simple Control Transmission Protocol March 2000 DATA [TSN=7,Strm=0,Seq=3] --------> X (lost) DATA [TSN=8,Strm=0,Seq=4] ---------------> (gap detected, immediately send ack) /----- SACK [TSN ACK=6,Frag=1, / Strt=2,End=2] <-----/ (remove 6 and 8 from out-queue, and strike 7 as "1" missing report) Note: in order to keep the size of the outbound SCTP datagram not to exceed the current path MTU, the maximal number of fragments that can be reported within a single SACK chunk is limited. When a single SACK can not cover all the fragments needed to be reported due to the MTU limitation, the endpoint SHALL send only one SACK, reporting the fragments from the lowest to highest TSNs, within the size limit set by the MTU, and leave the remaining highest TSN fragment numbers unacknowledged. 5.8 Adler-32 Checksum Calculation When sending an SCTP datagram, the sender MUST strengthen the data integrity of the transmission by including the Adler-32 checksum value calculated on the datagram, as described below. After the datagram is constructed (containing the SCTP common header and one or more control or DATA chunks), the sender shall: 1) fill in the proper Verification Tag in the SCTP common header and initialize the Adler-32 checksum filed to 0's. 2) calculate the Adler-32 checksum of the whole datagram, including the SCTP common header and all the chunks. Refer to Sections 8.2 and 9 in [2] for details of the Adler-32 algorithm. And, 3) put the resultant value into the Adler-32 checksum field in the common header, and leave the rest of the bits unchanged. When an SCTP datagram is received, the receiver MUST first check the Adler-32 checksum: 1) store the received Adler-32 checksum value aside, 2) replace the 32 bits of the Adler-32 checksum field in the received SCTP datagram with all '0's and calculate an Adler-32 checksum value of the whole received datagram. And, 3) verify that the calculated Adler-32 checksum is the same as the received Adler-32 checksum, If not, the receiver MUST treat the datagram as an invalid SCTP datagram. The default procedure of handling invalid SCTP datagrams is to silently discard them. Stewart, et al [Page 56] Internet Draft Simple Control Transmission Protocol March 2000 5.9 Segmentation Segmentation MUST be performed by the data sender if the user message to be sent has a large size that causes the outbound SCTP datagram size exceeding the current MTU. Note, if the data receiver is multi-homed, the sender shall choose a size no larger than the latest MTU of the current primary destination address. When determining when to segment, the SCTP implementation MUST take into account the SCTP datagram header as well as the DATA chunk header. The implementation MAY also take account of the space required for a SACK chunk. IMPLEMENTATION NOTE: if segmentation is not support by the sender, an error should be reported to the sender's SCTP user if the data to be sent has a size exceeding the current MTU. In such cases the Send primitive discussed in Section 9.1 would need to return an error to the upper layer. Segmentation takes the following steps: 1) the data sender SHALL break the large user message into a series of DATA chunks, such that each of the chunks can be fit into an IP datagram smaller than or equal to the current MTU, 2) the data sender MUST then assign, in sequence, a separate TSN to each of the DATA chunks in the series, 3) the data sender MUST also set the B/E bits of the first DATA chunk in the series to '10', the B/E bits of the last DATA chunk in the series to '01', and the B/E bits of all other DATA chunks in the series to '00'. The data receiver MUST recognize the segmented DATA chunks, by examining the B/E bits in each of the received DATA chunks, and queue the segmented DATA chunks for re-assembly. Then, it shall pass the re-assembled user message to the specific stream for possible re-ordering and final dispatching. Note, if the data receiver runs out of buffer space while still waiting for more segments to complete the re-assembly of the message, it should dispatch part of its inbound message through a partial delivery API (see Section 9), freeing some of its receive buffer space so that the rest of the message may be received. Stewart, et al [Page 57] Internet Draft Simple Control Transmission Protocol March 2000 5.10 Bundling and Multiplexing An SCTP sender achieves data bundling by simply including multiple DATA chunks in one outbound SCTP datagram. Note that the total size of the resultant IP datagram, including the SCTP datagram and IP headers, MUST be less or equal to the current MTU. Note, if the data receiver is multi-homed, the sender shall choose a size no larger than the latest MTU of the current primary destination address. When multiplexing control chunks with DATA chunks, control chunks have the priority and MUST be placed first in the outbound SCTP datagram and be transmitted first. The transmitter MUST transmit DATA chunks within a SCTP datagram in increasing order of TSN. Partial chunks MUST NOT be placed in a SCTP datagram. The receiver MUST process the chunks in order in the datagram. The receiver uses the chunk length field to determine the end of a chunk and beginning of the next chunk taking account of the fact that all chunks end on a thirty-two-bit word boundary. If the receiver detects a partial chunk, it MUST drop the chunk. 6. Congestion control Congestion control is one of the basic functions in the SCTP protocol. For some applications, it may be likely that adequate resources will be allocated to SCTP traffic to assure prompt delivery of time-critical SCTP data - thus it would appear to be unlikely, during normal operations, that SCTP transmissions encounter severe congestion condition. However SCTP must prepare itself for adverse operational conditions, which can develop upon partial network failures or unexpected traffic surges. In such situations SCTP must follow correct congestion control steps to recover from congestion quickly in order to get data delivered as soon as possible. In the absence of network congestion, these preventive congestion control algorithms should show no impact on the protocol performance. IMPLEMENTATION NOTE: as far as its specific performance requirements are met, an implementation is always allowed to adopt a more conservative congestion control algorithm than the one defined below. The congestion control algorithms used by SCTP are based on RFC 2581 [3], "TCP Congestion Control". This section describes how the algorithms defined in RFC 2581 are adopted for use in SCTP. We first list differences in protocol designs between TCP and SCTP, and then describe SCTP's congestion control scheme. The description will use the same terminology as in TCP congestion control whenever appropriate. Note: SCTP congestion control is always applied to the entire association, and NOT to individual streams. Stewart, et al [Page 58] Internet Draft Simple Control Transmission Protocol March 2000 6.1 SCTP Differences from TCP Congestion control One difference between SCTP and TCP is that the Selective Acknowledgment function (SACK) is designed into SCTP, rather than an enhancement that is added to the protocol later as is the case for TCP. SCTP SACK carries the same semantic meaning with that of TCP SACK. TCP and SCTP considers the information carried in the SACK as advisory information only. In SCTP, any DATA chunk that has been acknowledged by SACK, including DATA that arrived at the receiving end out of order, are NOT considered fully delivered until the Cumulative Acknowledgement point passes the acknowledged DATA chunk. Consequently, the value of cwnd controls the amount of outstanding data, rather than the upper bound between the highest acknowledged sequence number and the latest DATA chunk that can be sent within the congestion window, as is the case in non-SACK TCP. SCTP SACK leads to different implementations of fast-retransmit and fast-recovery from that of non-SACK TCP. As an example see [16]. The biggest difference between SCTP and TCP, however, is multi-homing. SCTP is designed to establish robust communication associations between two end points each of which may be reachable by more than one transport address. Potentially different addresses may lead to distinguished data paths between the two points, thus ideally one may need a separate set of congestion control parameters for each of the paths. The treatment here of congestion control for multi-homed receivers is new with SCTP and may require refinement in the future. The current algorithms make the following assumptions: o The sender always uses the same destination address until being instructed by the upper layer otherwise. o The sender keeps a separate congestion control parameter set for each of the destination addresses it can send to (NOT each source-destination pair but for each destination) . The parameters should decay if the address is not used for a long enough time period. o For each of the destination addresses, do slow-start upon the first transmission to that address. 6.2 SCTP Slow-Start and Congestion Avoidance The slow start and congestion avoidance algorithms MUST be used by a SCTP sender to control the amount of outstanding data being injected into the network. The congestion control in SCTP is employed in regard to the association, not to an individual stream. In some situations it may be beneficial for an SCTP sender to be more conservative than the algorithms allow, however an SCTP sender MUST NOT be more aggressive than the following algorithms allow. Like TCP, an SCTP sender uses the following three control variables on a per destination basis to regulate its transmission rate. Stewart, et al [Page 59] Internet Draft Simple Control Transmission Protocol March 2000 o Receiver advertised window size (rwnd, in octets), which is set by the receiver based on its available buffer space for incoming packets. o Congestion control window (cwnd, in octets), which is adjusted by the sender based on observed network conditions. o Slow-start threshold (ssthresh, in octets), which is used by the sender to distinguish slow start and congestion avoidance phases. SCTP also requires one additional control variable, partial_bytes_acked, which is used during congestion avoidance phase to facilitate cwnd adjustment. Unlike TCP, an SCTP sender MUST keep a set of these control variables for EACH destination address of its peer (when its peer is multi-homed). 6.2.1 Slow-Start Beginning data transmission into a network with unknown conditions requires SCTP to probe the network to determine the available capacity. The slow start algorithm is used for this purpose at the beginning of a transfer, or after repairing loss detected by the retransmission timer. o The initial cwnd before data transmission or after a sufficiently long idle period MUST be <= 2*MTU. o The initial cwnd after a retransmission timeout MUST be no more than 1*MTU. o The initial value of ssthresh MAY be arbitrarily high (for example, some implementations use the size of the receiver advertised window). o Whenever cwnd is greater than zero, the sender is allowed to have cwnd octets of data outstanding on that transport address. o When cwnd is less than or equal to ssthresh an SCTP sender MUST use the slow start algorithm to increase cwnd (assuming the current congestion window is being fully utilized). If the incoming SACK advances the cumulative TSN, cwnd MUST be increased by at most the lesser of 1) the total size of the previously outstanding DATA chunk(s) acknowledged, and 2) the destinations path MTU. This prevents against the ACK-Splitting attack outlined in [15]. NOTE: In instances where the data receiver endpoint is multi-homed, if a SACK arrives at the data sender that advances the sender's cumulative TSN point, then the data sender should update its cwnd (or cwnds) apportioned to the destination addresses where the data was transmitted to. However if the SACK does not advance the cumulative TSN point, the data sender MUST not adjust the cwnd of any of the destination addresses. Stewart, et al [Page 60] Internet Draft Simple Control Transmission Protocol March 2000 NOTE: because an SCTP data sender's cwnd is not tied to its cumulative TSN point, as duplicate SACKs come in, even though they may not advance the cumulative TSN point an SCTP sender can still use them to clock out new data. That is, the data newly acknowledged by the SACK diminishes the amount of data now in flight to less than cwnd; and so the current, unchanged value of cwnd now allows new data to be sent. On the other hand, the increase of cwnd must be tied to the cumulative TSN advancement as specified above. Otherwise the duplicate SACKs will not only clock out new data, but also will adversely clock out *more* new data than what has just left the network, during a time of possible congestion. o When the sender does not transmit data on a given transport address, the cwnd of the transport address should be adjusted to max(cwnd / 2, 2*MTU) per RTO. 6.2.2 Congestion Avoidance When cwnd is greater than ssthresh, cwnd should be incremented by 1*MTU per RTT if the sender has cwnd or more octets of data outstanding on the corresponding transport address. In practice an implementation can achieve this goal in the following way: o partial_bytes_acked is initialized to 0. o Whenever cwnd is greater than ssthresh, upon each SACK arrival, increase partial_bytes_acked by the total number of octets of all new chunks acknowledged in that SACK. o When partial_bytes_acked is equal or greater than cwnd and before the arrival of the SACK the sender has cwnd or more octets of data outstanding, increase cwnd by MTU, and reset partial_bytes_acked to (partial_bytes_acked - cwnd). o Same as in the slow start, when the sender does not transmit data on a given transport address, the cwnd of the transport address should be adjusted to max(cwnd / 2, 2*MTU) per RTO. o When all of the data transmitted by the sender has been acknowledged by the receiver, partial_bytes_acked is initialized to 0. 6.2.3 Congestion Control Upon detection of packet losses from SACK reports (see section 6.2.4), the sender should do the following: ssthresh = max(cwnd/2, 2*MTU) cwnd = ssthresh Stewart, et al [Page 61] Internet Draft Simple Control Transmission Protocol March 2000 Basically, a packet loss causes cwnd to be cut in half. When the T3-rxt timer expires on an address, SCTP should perform slow start by: ssthresh = max(cwnd/2, 2*MTU) cwnd = 1*MTU and assure that no more than one data packet will be in flight on that address until the sender receives acknowledgment for successful delivery of data to that address. 6.2.4 Fast Retransmit on Gap Reports In the absence of data losses, a SCTP receiver performs delayed acknowledgment. However, whenever a receiver notices a hole in the arriving TSN sequence, it should start sending a SACK back every time a packet arrives carrying data. At the sender end, whenever the sender receives a SACK that indicate some TSN(s) missing, it SHOULD wait for 3 further miss indications (via subsequent SACKs) on the same TSN(s) before taking action. When the TSN(s) is reported as missing in consecutive SACKs for the 4th time, the sender shall: 1) Mark the missing DATA chunk(s) for retransmission, 2) Adjust the ssthresh and cwnd of the destination address(es) where the missing data chunks were last sent, according to the formula described in Section 6.2.3. 3) Determine how many of the earliest (i.e., lowest TSN) missing Data chunks will fit into a single packet, subject to constraint of the path MTU of the destination transport address to which the packet is being sent. Call this value K. Retransmit those K data chunks in a single packet. 4) Restart T3-rxt timer ONLY IF the last SACK acknowledged the lowest outstanding TSN number sent to that address, or we are retransmitting the first outstanding Data chunk sent to that address. Note, before the above adjustments, if the received SACK also acknowledges new data chunks and advances the cumulative TSN point, the cwnd adjustment rules defined in Sections 6.2.1 and 6.2.2 must be applied first. A straightforward implementation of the above requires that the sender keeps a counter for each TSN hole first reported by a SACK; the counter keeps track of whether 3 subsequent SACKs have reported the same hole. Stewart, et al [Page 62] Internet Draft Simple Control Transmission Protocol March 2000 Because cwnd in SCTP indirectly bounds the number of outstanding TSN's, the effect of TCP fast-recovery is achieved automatically with no adjustment to the congestion control window size. 6.3 Path MTU Discovery RFC 1191 [11] discusses "Path MTU Discovery", whereby a sender maintains an estimate of the maximum transmission unit (MTU) along a given Internet path and refrains from sending datagrams along that path which exceed the MTU, other than occasional attempts to probe for a change in the path MTU. RFC 1191 is thorough in its discussion of the MTU discovery mechanism and strategies for determining the current end-to-end MTU setting as well as detecting changes in this value. RFC 1981 [12] discusses applying the same mechanisms for IPv6. An SCTP sender SHOULD apply these techniques, and SHOULD do so on a per-destination-address basis. There are 4 ways in which SCTP differs from the description in RFC 1191 of applying MTU discovery to TCP: 1) SCTP associations can span multiple set of addresses. Per the above comment, an SCTP sender MUST maintain separate MTU estimates for each destination address of its peer. 2) Elsewhere in this document, when the term "MTU" is discussed, it refers to the MTU associated with the destination address corresponding to the context of the discussion. 3) Unlike TCP, SCTP does not have a notion of "Maximum Segment Size". Accordingly, the MTU for each destination address SHOULD be initialized to a value no larger than the link MTU for the local interface to which datagrams for that remote destination address will be routed. 4) Since data transmission in SCTP is naturally structured in terms of TSNs rather than bytes (as is the case for TCP), the discussion in section 6.5 of RFC 1191 applies: when retransmitting a datagram to a remote address for which the datagram appears too large for the path MTU to that address, the datagram SHOULD be retransmitted without the DF bit set, allowing it to possibly be fragmented. Transmissions of new datagrams MUST have DF set. Other than these differences, the discussion of TCP's use of MTU discovery in RFCs 1191 and 1981 applies to SCTP, too, on a per-destination-address basis. Stewart, et al [Page 63] Internet Draft Simple Control Transmission Protocol March 2000 7. Fault Management 7.1 Endpoint Failure Detection The data sender shall keep a counter on the total number of consecutive retransmissions to its peer (including retransmissions to ALL the destination transport addresses of the peer if it is multi-homed). If the value of this counter exceeds the limit indicated in the protocol parameter 'Association.Max.Retrans', the data sender shall consider the peer endpoint unreachable and shall stop transmitting any more data to it (and thus the association enters the CLOSED state). In addition, the data sender shall report the failure to the upper layer, and optionally report back all outstanding user data remaining in its outbound queue. The association is automatically terminated when the peer endpoint becomes unreachable. The counter shall be reset each time a datagram sent to that destination address is acknowledged by the peer endpoint, or a HEARTBEAT-ACK is received from the peer endpoint. 7.2 Path Failure Detection When the remote endpoint is multi-homed, the data sender should keep a 'retrans.count' counter for each of the destination transport addresses of the remote endpoint. Each time the T3-rxt timer on any address, or when a HEARTBEAT sent to an idle address is not acknowledged, the 'retrans.count' counter of that destination address will be incremented. When the value in 'retrans.count' exceeds the protocol parameter 'Path.Max.Retrans' of that destination address, the data sender should mark the destination transport address as inactive, and a notification SHOULD be sent to the upper layer. When an outstanding TSN is acknowledged or a HEARTBEAT sent to that address is acknowledged with a HEARTBEAT-ACK, the data sender shall clear the 'retrans.count' counter of the destination transport address to which the datagram was last sent (or HEARTBEAT was sent). Note, when the data receiver is multi-homed and the last sent was a retransmission to an alternate address of the receiver, there exists an ambiguity as to whether or not the acknowledgment should be credited to the address of the last sent. However, this ambiguity does not seem to bear any significant consequence to SCTP behavior. If this ambiguity is undesirable, the data sender may choose not to clear the 'retrans.count' counter if the last sent was a retransmission. Stewart, et al [Page 64] Internet Draft Simple Control Transmission Protocol March 2000 Note, when configuring the SCTP endpoint, the user should avoid having the value of 'Association.Max.Retrans' larger than the summation of the 'Path.Max.Retrans' of all the destination addresses for the remote endpoint. Otherwise, all the destination addresses may become inactive while the endpoint still considers the peer endpoint reachable. When this condition occurs, how the SCTP chooses to function is implementation specific. Note, when the primary destination address is marked inactive (due to excessive retransmissions, for instance), the sender MAY automatically transmit new datagrams to an alternate destination address if one exists and is active. This is, however, an implementation option. 7.3 Path Heartbeat By default, an SCTP endpoint shall monitor the reachability of the idle destination transport address(es) of its peer by sending HEARTBEAT messages periodically to the destination transport address(es). A destination transport address is considered "idle" if no new chunk which can be used for updating path RTT (usually including first transmission DATA, INIT, COOKIE, etc.) and no heartbeat has been sent to it within the current heartbeat period of that address. This applies to both active and inactive destination addresses. The upper layer can optionally initiate the following functions: A) disable heart beat on a specific destination transport address of a given association, B) re-enable heart beat on a specific destination transport address of a given association, and, C) request an on-demand heartbeat on a specific destination transport address of a given association. The endpoint should increment the respective 'retrans.count' counter of the destination transport address each time a HEARTBEAT is sent to that address. When the value of this counter reaches the protocol parameter 'Path.Max.Retrans', the endpoint should mark the corresponding destination address as inactive if it is not so marked, and may also optionally report to the upper layer the change of reachability of this destination address. After this, the endpoint should continue heartbeat on this destination address but should stop increasing the counter. The sender of the HEARTBEAT message should include in the Heartbeat Information field of the message the current time when the message is sent out and the information on the destination address to which the message is sent. Stewart, et al [Page 65] Internet Draft Simple Control Transmission Protocol March 2000 The receiver of the HEARTBEAT should immediately respond with a HEARTBEAT ACK that contains the Heartbeat Information field copied out from the received HEARTBEAT message. Upon the receipt of the HEARTBEAT ACK, the sender of the HEARTBEAT should clear the 'retrans.count' counter of the destination transport address to which the HEARTBEAT was sent, and mark the destination transport address as active if it is not so marked. The endpoint may optionally report to the upper layer when an inactive destination address is marked as active due to the reception of the latest HEARTBEAT ACK. The receiver of the HEARTBEAT ACK should also perform an RTT measurement for that destination transport address using the time value carried in the HEARTBEAT ACK message. On an idle destination address that is allowed to heartbeat, HEARTBEAT messages is RECOMMENDED to be sent once per RTO of that destination address, with jittering of +/- 50%, and exponential back-off if the previous HEARTBEAT is unanswered. A primitive is provided for the SCTP user to change the heart beat interval and turn on or off the heart beat on a given destination address. Note, the heartbeat interval set by the SCTP user on any of the idle destination addresses SHOULD be no smaller than the RTO of that destination address. Separate timers may be used to control the heartbeat transmission for different idle destination addresses. 7.4 Handle "Out of the blue" Packets An SCTP datagram is called an "out of the blue" (OOTB) datagram if it is correctly formed, i.e., passed the receiver's Adler-32 check (see Section 5.8), but the receiver is not able to identify the association to which this datagram belongs. The receiver of an OOTB datagram MUST do the following: 1) check if the OOTB datagram contains an ABORT chunk. If so, the receiver MUST silently discarded the OOTB datagram and take no further action. Otherwise, 2) the receiver should respond the sender of the OOTB datagram with an ABORT. When sending the ABORT, the receiver of the OOTB datagram MUST fill in the Verification Tag field of the outbound datagram with the value found in the Verification Tag field of the OOTB datagram. After sending this ABORT, the receiver of the OOTB datagram shall discard the OOTB datagram and take no further action. Stewart, et al [Page 66] Internet Draft Simple Control Transmission Protocol March 2000 7.5 Verification Tag The Verification Tag rules defined in this section apply when sending or receiving SCTP datagrams which do NOT contain an INIT, SHUTDOWN ACK, or ABORT chunk. The rules for sending and receiving SCTP datagrams containing one of these chunk types are discussed separately in Section 7.5.1. When sending an SCTP datagram, the sender MUST fill in the Verification Tag field of the outbound datagram with the tag value of the peer endpoint to which this SCTP datagram is destined. When receiving an SCTP datagram, the receiver MUST ensure that the value in the Verification Tag field of the received SCTP datagram matches its own Tag. If the received tag value does not match the receiver's own tag value, the receiver shall silently discard the datagram and shall not process it any further. 7.5.1 Exceptions in Verification Tag Rules A) Rules for datagram carrying INIT: - The sender MUST set the Verification Tag of the datagram to 0. - The receiver, when noticing an incoming SCTP datagram with the Verification Tag set to 0, should continue to process the datagram only if an INIT chunk is present. Otherwise, the receiver MUST silently discard the datagram and take no further action. B) Rules for datagram carrying ABORT: - The sender shall always fill in the Verification Tag field of the outbound datagram with the destination endpoint's tag value if it is known. - If the ABORT is sent in response to an OOTB datagram, the sender MUST follow the procedure described in Section 7.4. - The receiver MUST accept the datagram IF the Verification Tag matches either its own tag, OR the tag of its peer. Otherwise, the receiver MUST silently discard the datagram and take no further action. C) Rules for datagram carrying SHUTDOWN ACK: - When sending a SHUTDOWN ACK, the sender is allowed to either use the destination endpoint's tag or set the Verification Tag field of the outbound datagram to 0. - The receiver of a SHUTDOWN ACK shall accept the datagram IF the Verification Tag field of the datagram matches its own tag OR is set to 0. Otherwise, the receiver MUST silently discard the datagram and take no further action. NOTE: the receiver of the SHUTDOWN ACK MUST ignore the chunk if it is not in the SHUTDOWN SENT state. Stewart, et al [Page 67] Internet Draft Simple Control Transmission Protocol March 2000 8. Termination of Association All existing associations should be terminated when an endpoint exits from service. An association can be terminated by either close or shutdown. 8.1 Close of an Association When an endpoint decides to close down an existing association, it shall send an ABORT message to its peer endpoint. The sender MUST fill in the peer's Verification Tag in the outbound datagram and MUST NOT bundle any DATA chunk with the ABORT. No acknowledgment is required for an ABORT message. In any circumstances, an endpoint MUST NOT respond to any received datagram that contains an ABORT with its own ABORT (also see Section 7.4). The receiver shall apply the special Verification Tag check rules described in Section 7.5.1 when handling the datagram carrying an ABORT. After checking the Verification Tag, the peer shall remove the association from its record, and shall report the termination to its upper layer. 8.2 Shutdown of an Association Using the TERMINATE primitive (see Section 9.1), the upper layer of an endpoint in an association can gracefully shutdown the association. This will guarantee that all outstanding datagrams from the peer of the shutdown initiator be delivered before the association terminates. Upon receipt of the TERMINATE primitive from its upper layer, the initiator endpoint enters SHUTDOWN-PENDING state and remains there until all outstanding TSNs have been acknowledged by the far end. It accepts no new data from its upper layer, but retransmits data to the far end if necessary to fill gaps. Once all outstanding TSNs have been acknowledged, the initiator endpoint shall send a SHUTDOWN message to the peer of the association, and shall include the last cumulative TSN it has received from the peer in the 'Cumulative TSN ACK' field. It shall then start the T2-shutdown timer and enter the Shutdown-sent state. If the timer expires, the initiator must re-send the SHUTDOWN with the updated last TSN received from its peer. The same rules in 5.3 SHALL be followed to determine the proper timer value for T2-shutdown. The sender of the SHUTDOWN message may also optionally include a SACK to indicate any gaps by bundling both the SACK and SHUTDOWN message together. Stewart, et al [Page 68] Internet Draft Simple Control Transmission Protocol March 2000 Note the sender of a shutdown should limit the number of retransmissions of the shutdown message to the protocol parameter 'Association.Max.Retrans'. If this threshold is exceeded the endpoint should destroy the TCB and may report the endpoint unreachable to the upper layer (and thus the association enters the CLOSED state). Upon the reception of the SHUTDOWN, the peer shall enter the Shutdown-received state, and shall verify, by checking the TSN ACK field of the message, that all its outstanding datagrams have been received by the initiator. If there are still outstanding datagrams left, the peer shall mark them for retransmission and start the retransmit procedure as defined in Section 5.3. While in Shutdown-sent state, the initiator shall immediately respond to each inbound SCTP datagram containing user data from the peer with a SACK and restart the T2-shutdown timer. If there is no more outstanding datagrams, the peer shall send a SHUTDOWN ACK and then remove all record of the association. Upon the receipt of the SHUTDOWN ACK, the initiator shall stop the T2-shutdown timer and remove all record of the association. Note: that it should be the responsibility of the initiator to assure that all the outstanding datagrams on its side have been resolved before it initiates the shutdown procedure. Note: an endpoint shall reject any new data request from its upper layer if it is in Shutdown-sent or Shutdown-received state until completion of the sequence. Note: if an endpoint is in Shutdown-sent state and receives an INIT message from its peer, it should discard the INIT message and retransmit the shutdown message. The sender of the INIT should respond with a stand-alone SHUTDOWN ACK in an SCTP datagram with the Verification Tag field of its common header set to 0, and let the normal T1-init timer cause the INIT message to be retransmitted and thus restart the association. Note: if an endpoint is in Shutdown-sent state and receives a SHUTDOWN message from its peer, the endpoint shall respond immediately with a SHUTDOWN ACK and shall stop the T2-shutdown timer and remove all record of the association. 9. Interface with Upper Layer The Upper Layer Protocols (ULP) shall request for services by passing primitives to SCTP and shall receive notifications from SCTP for various events. Stewart, et al [Page 69] Internet Draft Simple Control Transmission Protocol March 2000 The primitives and notifications described in this section should be used as a guideline for implementing SCTP. The following functional description of ULP interface primitives is shown for illustrative purposes. We must warn readers that different SCTP implementations may have different ULP interfaces. However, all SCTPs must provide a certain minimum set of services to guarantee that all SCTP implementations can support the same protocol hierarchy. 9.1 ULP-to-SCTP The following sections functionally characterize a ULP/SCTP interface. The notation used is similar to most procedure or function calls in high level languages. The ULP primitives described below specify the basic functions the SCTP must perform to support inter-process communication. Individual implementations must define their own exact format, and may provide combinations or subsets of the basic functions in single calls. A) Initialize Format: INITIALIZE ([local port], [local eligible address]) -> local SCTP instance name This primitive allows SCTP to initialize its internal data structures and allocate necessary resources for setting up its operation environment. Note that once SCTP is initialized, ULP can communicate directly with other endpoints without re-invoking this primitive. A local SCTP instance name will be returned to the ULP by the SCTP. Mandatory attributes: None. Optional attributes: The following types of attributes may be passed along with the primitive: o local port - SCTP port number, if ULP wants it to be specified; o local eligible address - A single address that the local SCTP endpoint should bind. By default all transport interface cards should be used by the local endpoint. IMPLEMENTATION NOTE: if this optional attribute is supported by an implementation, it will be the responsibility of the implementation to enforce that the IP source address field of any SCTP datagrams sent out by this endpoint MUST contain the IP addresses indicated in the local eligible address. Stewart, et al [Page 70] Internet Draft Simple Control Transmission Protocol March 2000 B) Associate Format: ASSOCIATE(local SCTP instance name, destination transport addr, outbound stream count) -> association id [,destination transport addr list] [,outbound stream count] This primitive allows the upper layer to initiate an association to a specific peer endpoint. The peer endpoint shall be specified by one of the transport addresses which defines the endpoint (see section 1.4). If the local SCTP instance has not been initialized, the ASSOCIATE is considered an error. An association id, which is a local handle to the SCTP association, will be returned on successful establishment of the association. If SCTP is not able to open an SCTP association with the peer endpoint, an error is returned. Other association parameters may be returned, including the complete destination transport addresses of the peer as well as the outbound stream count of the local endpoint. One of the transport address from the returned destination addresses will be selected by the local endpoint as default primary destination address for sending SCTP datagrams to this peer. The returned "destination transport addr list" can be used by the ULP to change the default primary destination address or to force sending a datagram to a specific transport address. IMPLEMENTATION NOTE: If ASSOCIATE primitive is implemented as a blocking function call, the ASSOCIATE primitive can return association parameters in addition to the association id upon successful establishment. If ASSOCIATE primitive is implemented as a non-blocking call, only the association id shall be returned and association parameters shall be passed using the COMMUNICATION UP notification. Mandatory attributes: o local SCTP instance name - obtained from the INITIALIZE operation. o destination transport addr - specified as one of the transport addresses of the peer endpoint with which the association is to be established. o outbound stream count - the number of outbound streams the ULP would like to open towards this peer endpoint. Optional attributes: None. Stewart, et al [Page 71] Internet Draft Simple Control Transmission Protocol March 2000 C) Terminate Format: TERMINATE(association id) -> result Gracefully terminates an association. Any locally queued user data will be delivered to the peer. The association will be terminated only after the peer acknowledges all the messages sent. A success code will be returned on successful termination of the association. If attempting to terminate the association results in a failure, an error code shall be returned. Mandatory attributes: o association id - local handle to the SCTP association Optional attributes: None. D) Abort Format: ABORT(association id [, cause code]) -> result Ungracefully terminates an association. Any locally queued user data will be discarded and an ABORT message is sent to the peer. A success code will be returned on successful abortion of the association. If attempting to abort the association results in a failure, an error code shall be returned. Note: If possible the SCTP should attempt to return all un-acknowledged data to the upper layer, however this behavior is implementation dependent. Mandatory attributes: o association id - local handle to the SCTP association Optional attributes: o cause code - reason of the abort to be passed to the peer. None. Stewart, et al [Page 72] Internet Draft Simple Control Transmission Protocol March 2000 E) Send Format: SEND(association id, buffer address, byte count [,context] [,stream id] [,life time] [,destination transport address] [,un-order flag] [,no-bundle flag]) -> result This is the main method to send user data via SCTP. Mandatory attributes: o association id - local handle to the SCTP association o buffer address - the location where the user message to be transmitted is stored; o byte count - The size of the user data in number of octets; Optional attributes: o context - optional information that will be carried in the sending failure notification to the ULP if the transportation of this datagram fails. o stream id - to indicate which stream to send the data on. If not specified, stream 0 will be used. o life time - specifies the life time of the user data. The user data will not be sent by SCTP after the life time expires. This parameter can be used to avoid efforts to transmit stale user messages. SCTP notifies the ULP, if the data cannot be initiated to transport (i.e. sent to the destination via SCTP's send primitive) within the life time variable. However, the user data will be transmitted if a TSN has been assigned to the user data before the life time expired. o destination transport address - specified as one of the destination transport addresses of the peer endpoint to which this message should be sent. Whenever possible, SCTP should use this destination transport address for sending the datagram, instead of the current primary destination transport address. o un-order flag - this flag, if present, indicates that the user would like the data delivered in an un-ordered fashion to the peer. o no-bundle flag - instructs SCTP not to bundle the user data with other outbound DATA chunks. Note: SCTP may still bundle even when this flag is present, when faced with network congestion. Stewart, et al [Page 73] Internet Draft Simple Control Transmission Protocol March 2000 F) Set Primary Format: SETPRIMARY(association id, destination transport address) -> result Instructs the local SCTP to use the specified destination transport address as primary destination address for sending datagrams. The result of attempting this operation shall be returned. If the specified destination transport address is not present in the "destination transport address list" returned earlier in an associate command or communication up notification, an error shall be returned. Mandatory attributes: o association id - local handle to the SCTP association o destination transport address - specified as one of the transport addresses of the peer endpoint, which should be used as primary address for sending datagrams. This overrides the current primary address information maintained by the local SCTP endpoint. Stewart, et al [Page 74] Internet Draft Simple Control Transmission Protocol March 2000 G) Receive Format: RECEIVE(association id, buffer address, buffer size [,stream id]) -> byte count [,transport address] [,stream id] [,stream sequence number] [,partial flag] [, delivery number] This primitive shall read the first user message in the SCTP in-queue to ULP, if there is one available, into the specified buffer. The size of the message read, in octets, will be returned. It may, depending on the specific implementation, also return other information such as the sender's address, the stream id on which it is received, whether there are more messages available for retrieval, etc. For ordered messages, their stream sequence number may also be returned. Depending upon the implementation, if this primitive is invoked when no message is available the implementation should return an indication of this condition or should block the invoking process until data does become available. Mandatory attributes: o association id - local handle to the SCTP association o buffer address - the memory location indicated by the ULP to store the received message. o buffer size - the maximum size of data to be received, in octets. Optional attributes: o stream id - to indicate which stream to receive the data on. o stream sequence number - the stream sequence number assigned by the sending SCTP peer. o partial flag - if this returned flag is set to 1, then this message is a partial delivery of the whole message. When this flag is set, the stream id and stream sequence number MUST accompany this receive. When this flag is set to 0, it indicates that no more deliveries will be received for this stream sequence number. Stewart, et al [Page 75] Internet Draft Simple Control Transmission Protocol March 2000 H) Status Format: STATUS(association id) -> status data This primitive should return a data block containing the following information: association connection state, destination transport address list, destination transport address reachability state, current receiver window size, current congestion window sizes, number of DATA chunks awaiting acknowledgment, number of DATA chunks pending receipt, primary destination transport address, SRTT on primary destination address, RTO on primary destination address, SRTT and RTO on other destination addresses, etc. Mandatory attributes: o association id - local handle to the SCTP association Optional attributes: None. I) Change Heartbeat Format: CHANGEHEARTBEAT(association id, destination transport address, new state [,interval]) -> result Instructs the local endpoint to enable or disable heart beat on the specified destination transport address. The result of attempting this operation shall be returned. Note, even when enabled, heart beat will not take place if the destination transport address is not idle. Mandatory attributes: o association id - local handle to the SCTP association o destination transport address - specified as one of the transport addresses of the peer endpoint. o new state - the new state of heart beat for this destination transport address (either enabled or disabled). Optional attributes: o interval - if present, indicates the frequency of the heart beat if this is to enable heart beat on a destination transport address. Default interval is the RTO of the destination address. Stewart, et al [Page 76] Internet Draft Simple Control Transmission Protocol March 2000 J) Request HeartBeat Format: REQUESTHEARTBEAT(association id, destination transport address) -> result Instructs the local endpoint to perform a HeartBeat on the specified destination transport address of the given association. The returned result should indicate whether the transmission of the HEARTBEAT message to the destination address is successful. Mandatory attributes: o association id - local handle to the SCTP association o destination transport address - the transport address of the association on which a heartbeat should be issued. K) Get SRTT Report Format: GETSRTTREPORT(association id, destination transport address) -> srtt result Instructs the local SCTP to report the current SRTT measurement on the specified destination transport address of the given association. The returned result can be an integer containing the most recent SRTT in milliseconds. Mandatory attributes: o association id - local handle to the SCTP association o destination transport address - the transport address of the association on which the SRTT measurement is to be reported. Stewart, et al [Page 77] Internet Draft Simple Control Transmission Protocol March 2000 L) Set Failure Threshold Format: SETFAILURETHRESHOLD(association id, destination transport address, failure threshold) -> result This primitive allows the local SCTP to customize the reachability failure detection threshold 'Path.Max.Retrans' for the specified destination address. Mandatory attributes: o association id - local handle to the SCTP association o destination transport address - the transport address of the association on which the failure detection threshold is to be set. o failure threshold - the new value of 'Path.Max.Retrans' for the destination address. M) Set Protocol Parameters Format: SETPROTOCOLPARAMETERS(association id, [,destination transport address,] protocol parameter list) -> result This primitive allows the local SCTP to customize the protocol parameters. Mandatory attributes: o association id - local handle to the SCTP association o protocol parameter list - The specific names and values of the protocol parameters (e.g., Association.Max.Retrans [see Section 13]) that the SCTP user wishes to customize. Optional attributes: o destination transport address - some of the protocol parameters may be set on a per destination transport address basis. 9.2 SCTP-to-ULP It is assumed that the operating system or application environment provides a means for the SCTP to asynchronously signal the ULP process. When SCTP does signal an ULP process, certain information is passed to the ULP. IMPLEMENTATION NOTE: in some cases this may be done through a seperate socket or error channel. Stewart, et al [Page 78] Internet Draft Simple Control Transmission Protocol March 2000 A) DATA ARRIVE notification SCTP shall invoke this notification on the ULP when a user message is successfully received and ready for retrieval. The following may be optionally be passed with the notification: o association id - local handle to the SCTP association o stream id - to indicate which stream the data is received on. B) SEND FAILURE notification If a message can not be delivered SCTP shall invoke this notification on the ULP. The following may be optionally be passed with the notification: o association id - local handle to the SCTP association o data - the location ULP can find the un-delivered message. o cause code - indicating the reason of the failure, e.g., size too large, message life-time expiration, etc. o context - optional information associated with this message (see D in section 9.1). C) NETWORK STATUS CHANGE notification When a destination transport address is marked down (e.g., when SCTP detects a failure), or marked up (e.g., when SCTP detects a recovery), SCTP shall invoke this notification on the ULP. The following shall be passed with the notification: o association id - local handle to the SCTP association o destination transport address - This indicates the destination transport address of the peer endpoint affected by the change; o new-status - This indicates the new status. Stewart, et al [Page 79] Internet Draft Simple Control Transmission Protocol March 2000 D) COMMUNICATION UP notification This notification is used when SCTP becomes ready to send or receive user messages, or when a lost communication to an endpoint is restored. IMPLEMENTATION NOTE: If ASSOCIATE primitive is implemented as a blocking function call, the association parameters are returned as a result of the ASSOCIATE primitive itself. In that case, COMMUNICATION UP notification is optional at the association initiator's side. The following shall be passed with the notification: o association id - local handle to the SCTP association o status - This indicates what type of event that has occurred o destination transport address list - the complete set of transport addresses of the peer o outbound stream count - the maximum number of streams allowed to be used in this association by the ULP o inbound stream count - the number of streams the peer endpoint has requested with this association (this may not be the same number has 'outbound stream count'). Stewart, et al [Page 80] Internet Draft Simple Control Transmission Protocol March 2000 E) COMMUNICATION LOST notification When SCTP loses communication to an endpoint completely or detects that the endpoint has performed an abort or graceful shutdown operation, it shall invoke this notification on the ULP. The following shall be passed with the notification: o association id - local handle to the SCTP association o status - This indicates what type of event that has occurred; The following may be optionally passed with the notification: o unsent-messages - The number and location of un-sent messages still in hold by SCTP; o unacknowledged-messages - The number and location of messages that were attempted to be transported to the destination, but were not acknowledged when the loss of communication was detected. o last-acked - the sequence number last acked by that peer endpoint; o last-sent - the sequence number last sent to that peer endpoint; o received-but-not-delivered - messages that were received by SCTP but not yet delivered to the ULP. Note: the un-send data report may not be accurate for those user messages which are segmented by SCTP during transmission. F) COMMUNICATION ERROR notification When SCTP receives an ERROR chunk from its peer and decides to notify its ULP, it can invoke this notification on the ULP. The following can be passed with the notification: o association id - local handle to the SCTP association o error info - this indicates the type of error and optionally some additional information received through the ERROR chunk. Stewart, et al [Page 81] Internet Draft Simple Control Transmission Protocol March 2000 10. Security Considerations 10.1 Security Objectives As a common transport protocol designed to reliably carry time- sensitive user messages, such as billing or signaling messages for telephony services, between two networked endpoints, SCTP has the following security objectives. - availability of reliable and timely data transport services - integrity of the user-to-user information carried by SCTP 10.2 SCTP Responses To Potential Threats It is clear that SCTP may potentially be used in a wide variety of risk situations. It is important for operator(s) of the systems concerned to analyze their particular situations and decide on the appropriate counter-measures. Where the SCTP system serves a group of users, it is probably operating as part of a professionally managed corporate or service provider network. It is reasonable to expect that this management includes an appropriate security policy framework. [RFC 2196, "Site Security Handbook", B. Fraser Ed., September 1997] should be consulted for guidance. The case is more difficult where the SCTP system is operated by a private user. The service provider with whom that user has a contractual arrangement SHOULD provide help to ensure that the user's site is secure, ranging from advice on configuration through downloaded scripts and security software. 10.2.1 Countering Insider Attacks The principles of the Site Security Handbook [13] should be applied to minimize the risk of theft of information or sabotage by insiders. These include publication of security policies, control of access at the physical, software, and network levels, and separation of services. Stewart, et al [Page 82] Internet Draft Simple Control Transmission Protocol March 2000 10.2.2 Protecting against Data Corruption in the Network Where the risk of undetected errors in datagrams delivered by the lower layer transport services is considered to be too great, additional checksum protection may be required. The question is whether this is appropriately provided as an SCTP service because it is needed by most potential users of SCTP, or whether instead it should be provided by the SCTP user application. (The SCTP protocol overhead, as opposed to the signaling payload, is protected adequately by the Adler-32 checksum and measures taken in SCTP to prevent replay attacks and masquerade.) In any event, the checksum must be specifically designed to ensure that it detects the errors left behind by the Adler-32 checksum. 10.2.3 Protecting Confidentiality In most cases, the risk of breach of confidentiality applies to the signaling data payload, not to the SCTP or lower-layer protocol overheads. If that is true, encryption of the SCTP user data only may be considered. As with the supplementary checksum service, user data encryption may be performed by the SCTP user application. Particularly for mobile users, the requirement for confidentiality may include the masking of IP addresses and ports. In this case IPSEC ESP should be used instead of application-level encryption. Similarly, where other reasons prompt the use of the IPSEC ESP service, application-level encryption is unnecessary. It will be up to the SCTP system operators to configure the application appropriately. Regardless of which level performs the encryption, the IPSEC ISAKMP service should be used for key management. Operators should consult [RFC 2401, "Security Architecture for the Internet Protocol", S. Kent, R. Atkinson, November 1998] for information on the configuration of IPSEC services between hosts with and without intervening firewalls. 10.2.4 Protecting against Blind Denial of Service Attacks A blind attack is one where the attacker is unable to intercept or otherwise see the content of data flows passing to and from the target SCTP node where it is not a party to the association. Blind denial of service attacks may take the form of flooding, masquerade, or improper monopolization of services. Stewart, et al [Page 83] Internet Draft Simple Control Transmission Protocol March 2000 10.2.4.1 Flooding The objective of flooding is to cause loss of service and incorrect behavior at target systems through resource exhaustion, interference with legitimate transactions, and exploitation of buffer-related software bugs. Flooding may be directed either at the SCTP node or at resources in the intervening IP Access Links or the Internetwork. Where the latter entities are the target, flooding will manifest itself as loss of network services, including potentially the breach of any firewalls in place. In general, protection against flooding begins at the equipment design level, where it includes measures such as: - avoiding commitment of limited resources before determining that the request for service is legitimate - giving priority to completion of processing in progress over the acceptance of new work - identification and removal of duplicate or stale queued requests for service. Network equipment should be capable of generating an alarm and log if a suspicious increase in traffic occurs. The log should provide information such as the identity of the incoming link and source address(es) used which will help the network or SCTP system operator to take protective measures. Procedures should be in place for the operator to act on such alarms if a clear pattern of abuse emerges. The design of SCTP is resistant to flooding attacks, particularly in its use of a four-way start-up handshake, its use of a cookie to defer commitment of resources at the responding SCTP node until the handshake is completed, and its use of a verification tag to prevent insertion of extraneous messages into the flow of an established association. 10.2.4.2 Masquerade Masquerade can be used to deny service in several ways: - by tying up resources at the target SCTP node to which the impersonated node has limited access. For example, the target node may by policy permit a maximum of one SCTP association with the impersonated SCTP node. The masquerading attacker may attempt to establish an association purporting to come from the impersonated node so that the latter cannot do so when it requires it. - by deliberately allowing the impersonation to be detected, thereby provoking counter-measures which cause the impersonated node to be locked out of the target SCTP node. - by interfering with an established association by inserting extraneous content such as a SHUTDOWN request. Stewart, et al [Page 84] Internet Draft Simple Control Transmission Protocol March 2000 SCTP prevents masquerade through IP spoofing by use of the four-way startup handshake. Because the initial exchange is memoryless, no lockout mechanism is triggered by masquerade attacks. SCTP protects against insertion of extraneous messages into the flow of an established association by use of the verification tag. Logging of received INIT requests and abnormalities such as unexpected INIT ACKs might be considered as a way to detect patterns of hostile activity. However, the potential usefulness of such logging must be weighed against the increased SCTP startup processing it implies, rendering the SCTP node more vulnerable to flooding attacks. Logging is pointless without the establishment of operating procedures to review and analyze the logs on a routine basis. 10.2.4.3 Improper Monopolization of Services Attacks under this heading are performed openly and legitimately by the attacker. They are directed against fellow users of the target SCTP node or of the shared resources between the attacker and the target node. Possible attacks include the opening of a large number of associations between the attacker's node and the target, or transfer of large volumes of information within a legitimately- established association. Such attacks take advantage of policy deficiencies at the target SCTP node. Defense begins with a contractual prohibition of behavior directed to denial of service to others. Policy limits should be placed on the number of associations per adjoining SCTP node. SCTP user applications should be capable of detecting large volumes of illegitimate or "no-op" messages within a given association and either logging or terminating the association as a result, based on local policy. 10.3 Protection against Fraud and Repudiation The objective of fraud is to obtain services without authorization and specifically without paying for them. In order to achieve this objective, the attacker must induce the SCTP user application at the target SCTP node to provide the desired service while accepting invalid billing data or failing to collect it. Repudiation is a related problem, since it may occur as a deliberate act of fraud or simply because the repudiating party kept inadequate records of service received. Potential fraudulent attacks include interception and misuse of authorizing information such as credit card numbers, blind masquerade and replay, and man-in-the middle attacks which modify the messages passing through a target SCTP association in real time. Stewart, et al [Page 85] Internet Draft Simple Control Transmission Protocol March 2000 The interception attack is countered by the confidentiality measures discussed in section 10.2.3 above. Section 10.2.4.2 describes how SCTP is resistant to blind masquerade attacks, as a result of the four-way startup handshake and the validation tag. The validation tag and TSN together are protections against blind replay attacks, where the replay is into an existing association. However, SCTP does not protect against man-in-the-middle attacks where the attacker is able to intercept and alter the messages sent and received in an association. Where a significant possibility of such attacks is seen to exist, or where possible repudiation is an issue, the use of the IPSEC AH service is recommended to ensure both the integrity and the authenticity of the messages passed. SCTP also provides no protection against attacks originating at or beyond the SCTP node and taking place within the context of an existing association. Prevention of such attacks should be covered by appropriate security policies at the host site, as discussed in section 10.2.1. 11. Recommended Transmission Control Block (TCB) Parameters This section details a recommended set of parameters that should be contained within the TCB for an implementation. This section is for illustrative purposes and should not be deemed as requirements on an implementation NOR as an exhaustive list of all parameters inside an SCTP TCB. Each implemenation may need its own additional parameters to optimize their implemenation. 11.1 Parameters necessary for the SCTP instance Associations A list of current associations and mappings to the data consumers for each association. This may be in the form of a hash table or other implementation dependent structure. The data consumers may be process identification information such as file descriptors, named pipe pointer, or table pointers dependent on how SCTP is implemented. Secret Key A secret key used by this endpoint to sign all cookies. This SHOULD be a cryptographic quality random number with a sufficient length. Discussion in RFC 1750 [1] can be helpful in selection of the key. Address List The list of IP addresses that this instance has bound. This information is passed to one's peer(s) in INIT and INIT-ACK messages. SCTP Port The local SCTP port number the endpoint is bound to. Stewart, et al [Page 86] Internet Draft Simple Control Transmission Protocol March 2000 Peer Tag value to be sent in every datagram and is received Verification in the INIT or INIT ACK message. Tag My Tag expected in every inbound datagram and sent in the Verification INIT or INIT ACK message. Tag 11.2 Parameters necessary per association (i.e. the TCB) State A state variable indicating what state the association is in, i.e . COOKIE_WAIT, COOKIE_SENT, ESTABLISHED, SHUTDOWN_PENDING, SHUTDOWN_SENT, SHUTDOWN_RECEIVED. Note: No "CLOSED" state is illustrated since if a association is "CLOSED" its TCB SHOULD be removed. Peer Transport A list of SCTP transport addresses that the peer is Address List bound to. This information is derived from the INIT or INIT-ACK and is used to associate an inbound datagram with a given association. Normally this information is hashed or keyed for quick lookup and access of the TCB. Primary This is the current primary destination transport Destination address of the peer endpoint. Overall The overall association error count. Error Count Overall Error The threshold for this association that if the Overall Threshold Error Count reaches will cause this association to be torn down. Peer Rwnd Current calculated value of the peer's rwnd. Next TSN My next TSN number I will assign. This is sent in the INIT or INIT-ACK message to the peer and incremented each time a DATA chunk is assigned a TSN (normally just prior to transmit or during segmentation). Last Rcvd TSN This is the last TSN I received and is the current cumulative TSN point. This value is set initially by taking the peers initial TSN, received in the INIT or INIT-ACK message, and subtracting one from it. Mapping Array An array of bits or bytes indicating which out of order TSN's have been received (relative to the cumulative TSN i.e. Last Rcvd TSN). If no GAP's exist, i.e. no out of order messages have been received, this array will be set to all zero. This structure may be in the form of a circular buffer or bit array. Stewart, et al [Page 87] Internet Draft Simple Control Transmission Protocol March 2000 Ack State This flag indicates if the next received datagram is to be responded to with a SACK. This is initialized to 0, when a datagram is received it is incremented. If this value reaches 2, a SACK is sent and the value is reset to 0. Note: this is used only when no datagrams are received out of order, when DATA chunks are out of order SACK's are not delayed (see Section 5). Inbound An array of structures to track the inbound streams. Streams Normally including the next sequence number expected and possibly the stream number. Outbound An array of structures to track the outbound streams. Streams Normally including the next sequence number to be sent on the stream. Reasm Queue A re-assembly queue. 11.3 Per Transport Address Data For each destination transport address in the peer's address list derived from the INIT or INIT ACK message, a number of data elements needs to be maintained including: Error count The current error count for this destination. Error Current error threshold for this destination i.e. Threshold what value marks the destination down if Error count reaches this value. cwnd The current congestion window. ssthresh The current ssthresh value. RTO The current retransmission timeout vaule. SRTT The current smoothed round trip time. RTTVAR The current RTT variation. partial bytes The tracking method for increase of cwnd when in acked congestion avoidance mode (see section 6.2.2) state The current state of this destionation, i.e. DOWN, UP, ALLOW-HB, NO-HEARTBEAT, etc. P-MTU The current known path MTU. Per A timer used by each destination. Destination Timer Stewart, et al [Page 88] Internet Draft Simple Control Transmission Protocol March 2000 RTO-Pending A flag used to track if one of the datagrams sent to this address is currently being used to compute a RTT. If this flag is 0, the next datagram sent to this destination should be used to compute a RTT and this flag should be set. Every time the RTT calcualtion completes (i.e. the datagram is SACK'd) clear this flag. last-time The time this destination was last sent to. This can be used used to determine if a HEARTBEAT is needed. 11.4 General Parameters Needed Out Queue A queue of outbound datagrams. In Queue A queue of inbound datagrams. 12. IANA Consideration This protocol will require port reservation like TCP for the use of "well known" servers within the Internet. It is suggested that all current TCP ports should be automatically reserved in the SCTP port address space. New requests should follow IANA's current mechanisms for TCP. This protocol may also be extended through IANA in three ways: -- through definition of additional chunk types, -- through definition of additional parameter types, or -- through definition of additional cause codes within Operation Error chunks In the case where a particular ULP using SCTP desires to have its own ports, the ULP should be responsible for registering with IANA for getting its ports assigned. 12.1 IETF-defined Chunk Extension The appropriate use of specific chunk types is an integral part of the SCTP protocol. In consequence, the intention is that new IETF-defined chunk types MUST be supported by standards-track RFC documentation. As a transitional step, a new chunk type MAY be introduced in an Experimental RFC. Chunk type codes MUST remain permanently associated with the original documentation on the basis of which they were allocated. Thus if the RFC supporting a given chunk type is deprecated in favor of a new document, the corresponding chunk type code value is also deprecated and a new code value is allocated in association with the replacement document. Stewart, et al [Page 89] Internet Draft Simple Control Transmission Protocol March 2000 The documentation for a new chunk code type must include the following information: (a) a long and short name for the new chunk type; (b) a detailed description of the structure of the chunk, which MUST conform to the basic structure defined in section 2.2; (c) a detailed definition and description of intended use of each field within the chunk, including the chunk flags if any; (d) a detailed procedural description of the use of the new chunk type within the operation of the protocol. If the primary numbering space reserved for IETF use (0x00 to 0xFD) is exhausted, new codes shall subsequently be allocated in the extension range 0x0000 through 0xFFFF. Chunks allocated in this range MUST conform to the following structure: First word (32 bits): as shown in section 2.2, with chunk type code equal to 0xFF. Second word: first octet MUST be all 1's (0xFF). Next octet MUST be all 0's (0x00). Final two octets contain the allocated extension code value. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 1 1 1 1 1 1|Chunk Flags | Chunk Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 1 1 1 1 1 1|0 0 0 0 0 0 0 0| Extension Type Code | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Value / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 12.2 IETF-defined Chunk Parameter Extension The allocation of a new chunk parameter type code from the IETF numbering space MUST be supported by RFC documentation. As with chunk type codes, parameter type codes are uniquely associated with their supporting document and MUST be replaced if new documentation is provided. This documentation may be Informational, Experimental, or standards-track at the discretion of the IESG. It MUST contain the following information: (a) Name of the parameter type. (b) Detailed description of the structure of the parameter field. This structure MUST conform to the general type-length-value format described in section 2.2.1. (c) Detailed definition of each component of the parameter value. (d) Detailed description of the intended use of this parameter type, and an indication of whether and under what circumstances multiple instances of this parameter type may be found within the same chunk. Stewart, et al [Page 90] Internet Draft Simple Control Transmission Protocol March 2000 Additional parameter type codes may be allocated initially from the range 0x0000 through 0xFFFD. If this space is exhausted, extension codes shall be allocated in the range 0x0000 through 0xFFFF. Where an extension code has been allocated, the format of the parameter must conform to the following structure: First word (32 bits): contains the parameter type code 0xFFFF and parameter length as described in section 2.2.1. Second word: first octet MUST be all 1's (0xFF). Next octet MUST be all 0's (0x00). Final two octets contain the allocated extension code value. The Value portion of the parameter, if any, follows the second word. 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1| Length | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |1 1 1 1 1 1 1 1|0 0 0 0 0 0 0 0| Extension Type Code | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ \ \ / Value / \ \ +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 12.3 IETF-defined Additional Error Causes Additional cause codes may be allocated in the range 0x0004 to 0xFFFF upon receipt of any permanently-available public documentation containing the following information: (a) Name of the error condition. (b) Detailed description of the conditions under which an SCTP endpoint should issue an Operation Error with this cause code. (c) Expected action by the SCTP endpoint which receives an Operation Error chunk containing this cause code. (d) Detailed description of the structure and content of data fields which accompany this cause code. The initial word (32 bits) of a cause code parameter MUST conform to the format shown in section 2.3.9, i.e.: -- first two octets contain the cause code value -- last two octets contain length of the cause parameter. Stewart, et al [Page 91] Internet Draft Simple Control Transmission Protocol March 2000 12.4 Payload Protocol Identifiers Except for value 0x00000000 which is reserved by SCTP to indicate the absence of a payload protocol identifier in a DATA chunk, SCTP will not be responsible for standardizing or verifying any payload protocol identifiers; SCTP simply receives the identifier from the upper layer and carries it with the corresponding payload data. The upper layer, i.e, the SCTP user, SHOULD standardize any specific protocol identifier with IANA if it is so desired. The use of any specific payload protocol identifier is out of the scope of SCTP. 13. Suggested SCTP Protocol Parameter Values The following protocol parameters are RECOMMENDED: RTO.Initial - 3 seconds RTO.Min - 1 second RTO.Max - 60 seconds RTO.Alpha - 1/8 RTO.Beta - 1/4 Valid.Cookie.Life - 5 seconds Association.Max.Retrans - 10 attempts Path.Max.Retrans - 5 attempts (per destination address) Max.Init.Retransmits - 8 attempts 'retrans.count' - counter (per destination address) 'receiver.buffer' - variable (per peer endpoint) IMPLEMENTATION NOTE: The SCTP implementation may allow ULP to customize some of these protocol parameters (see Section 9). 14. Acknowledgments The authors wish to thank Mark Allman, Richard Band, Scott Bradner, Steve Bellovin, Ram Dantu, R. Ezhirpavai, Sally Floyd, Matt Holdrege, Henry Houh, Christian Huetima, Gary Lehecka, John Loughney, Daniel Luan, Lyndon Ong, Shyamal Prasad, Kelvin Porter, Heinz Prantner, Jarno Rajahalme, Ivan Raymond E. Reeves, Renee Revis, Arias Rodriguez, A. Sankar, Greg Sidebottom, Brian Wyld, and many others for their invaluable comments. Stewart, et al [Page 92] Internet Draft Simple Control Transmission Protocol March 2000 15. Authors' Addresses Randall R. Stewart Tel: +1-847-632-7438 Motorola, Inc. EMail: rstewar1@email.mot.com 1501 W. Shure Drive, #2315 Arlington Heights, IL 60004 USA Qiaobing Xie Tel: +1-847-632-3028 Motorola, Inc. EMail: qxie1@email.mot.com 1501 W. Shure Drive, #2309 Arlington Heights, IL 60004 USA Ken Morneault Tel: +1-703-484-3323 Cisco Systems Inc. EMail: kmorneau@cisco.com 13615 Dulles Technology Drive Herndon, VA. 20171 USA Chip Sharp Tel: +1-919-472-3121 Cisco Systems Inc. EMail:chsharp@cisco.com 7025 Kit Creek Road Research Triangle Park, NC 27709 USA Hanns Juergen Schwarzbauer Tel: +49-89-722-24236 SIEMENS AG Hofmannstr. 51 81359 Munich Germany EMail: HannsJuergen.Schwarzbauer@icn.siemens.de Tom Taylor Tel: +1-613-736-0961 Nortel Networks 1852 Lorraine Ave. Ottawa, Ontario Canada K1H 6Z8 EMail:taylor@nortelnetworks.com Ian Rytina Tel: +61-3-9301-6164 Ericsson Australia EMail:ian.rytina@ericsson.com 37/360 Elizabeth Street Melbourne, Victoria 3000 Australia Malleswar Kalla Tel: +1-973-829-5212 Telcordia Technologies MCC 1J211R 445 South Street Morristown, NJ 07960 USA EMail: kalla@research.telcordia.com Stewart, et al [Page 93] Internet Draft Simple Control Transmission Protocol March 2000 Lixia Zhang Tel: +1-310-825-2695 UCLA Computer Science Department EMail: lixia@cs.ucla.edu 4531G Boelter Hall Los Angeles, CA 90095-1596 USA Vern Paxson Tel: +1-510-642-4274 x 302 ACIRI EMail: vern@aciri.org 1947 Center St., Suite 600, Berkeley, CA 94704-1198 USA 16. References [1] Eastlake , D. (ed.), "Randomness Recommendations for Security", RFC 1750, December 1994. [2] Deutsch, P., and Gailly, J-L., "ZLIB Compressed Data Format Specification version 3.3", RFC 1950, May 1996. [3] Allman, M., Paxson, V., and Stevens, W., "TCP Congestion Control", RFC 2581, April 1999. [4] Krawczyk, H., Bellare, M., Canetti, R., "HMAC: Keyed-Hashing for Message Authentication", RFC 2104, March 1997. [5] Allman, M., and Paxson, V., "On Estimating End-to-End Network Path Properties", Proc. SIGCOMM'99, 1999. [6] Karn, P., and Simpson, W., "Photuris: Session-Key Management Protocol", RFC 2522, March 1999. [7] Bradner, S., "The Internet Standards Process -- Revision 3", RFC 2026, October 1996. [8] Postel, J. (ed.), "Transmission Control Protocol", RFC 793, September 1981. [9] Postel, J. (ed.), "User Datagram Protocol", RFC 768, August 1980. [10] Reynolds, J., and Postel, J. (ed.), "Assigned Numbers", RFC 1700, October 1994. [11] Mogul, J., and Deering, S., "Path MTU Discovery", RFC 1191, November 1990. [12] McCann, J., Deering, S., and Mogul, J., "Path MTU Discovery for IP version 6", RFC 1981, August 1996. [13] Fraser, B. (ed.), "Site Security Handbook", RFC 2196, September 1997. Stewart, et al [Page 94] Internet Draft Simple Control Transmission Protocol March 2000 [14] Kent, S., and Atkinson, R., "Security Architecture for the Internet Protocol", RFC 2401, November 1998. [15] Savage, S., Cardwell, N., Wetherall, D., and Anderson, T., "TCP Congestion Control with a Misbehaving Receiver", ACM Computer Communication Review, 29(5), October 1999. [16] Fall, K., and Floyd, S., Simulation-based Comparisons of Tahoe, Reno, and SACK TCP, Computer Communications Review, V. 26 N. 3, July 1996, pp. 5-21. Appendix A: Explicit Congestion Notification ECN (Ramakrishnan, k., Floyd, S., "Explicit Congestion Notification", RFC 2481, January 1999) describes a proposed extension to IP that details a method to become aware of congestion outside of datagram loss. This is an optional feature that an implemenation MAY choose to add to SCTP. This appendix details the minor differences an implemenator will need to be aware of if they choose to implement this feature. In general RFC 2481 should be followed with the following exceptions. Negotiation: RFC2481 details negotiation of ECN during the SYN and SYN-ACK stages of a TCP connection. The sender of the SYN sets two bits in the TCP flags, and the sender of the SYN-ACK sets only 1 bit. The reasoning behind this is to assure both sides are truely ECN capable. For SCTP this is not necessary. To indicate that an endpoint is ECN capable a endpoint MAY add to the INIT and or INIT-ACK message the TLV reserved for ECN. This TLV contains no parameters and thus has the following format: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Parameter Type = 0x000a | Parameter Length = 0x0004 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Stewart, et al [Page 95] Internet Draft Simple Control Transmission Protocol March 2000 ECN-Echo: RFC 2481 details a specific bit for a receiver to send back in its acknowledgments to notifiy the sender of the Congestion Experienced (CE) bit having arrived from the network. For SCTP this same indication is made by including the ECNE chunk. This chunk contains NO parameters or data and looks as follows: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | ID=00001100 | Flags=00000000| Chunk Length=0004 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Note: The ECNE is considered a Control chunk. CWR: RFC 2481 details a specific bit for a sender to send in its next outbound datagram to indicate to its peer that it has reduced its congestion window. This is termed the CWR bit. For SCTP the same indication is made by including the CWR chunk. This chunk contains NO parameters or data and looks as follows: 0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | ID=00001101 | Flags=00000000| Chunk Length=0004 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Note: The CWR is considered a Control chunk. This Internet Draft expires in 6 months from March, 2000 Stewart, et al [Page 96]