- •Contents
- •Foreword to the First Edition
- •Preface to the Third Edition
- •Preface to the Second Edition
- •Preface to the First Edition
- •1 SIP and the Internet
- •1.1 Signaling Protocols
- •1.2.2 Data/Link Layer
- •1.2.3 Network Layer
- •1.2.4 Transport Layer
- •1.2 Internet Multimedia Protocol Stack
- •1.2.1 Physical Layer
- •1.2.5 Application Layer
- •1.2.6 Utility Applications
- •1.2.7 Multicast
- •1.3 Internet Names
- •1.4 URLs, URIs, and URNs
- •1.5 Domain Name Service
- •1.5.1 DNS Resource Records
- •1.5.2 Address Resource Records (A or AAAA)
- •1.5.3 Service Resource Records (SRV)
- •1.5.4 Naming Authority Pointer Resource Records (NAPTR)
- •1.5.5 DNS Resolvers
- •1.6 Global Open Standards
- •1.7 Internet Standards Process
- •1.8 A Brief History of SIP
- •1.9 Conclusion
- •References
- •2 Introduction to SIP
- •2.1 A Simple Session Establishment Example
- •2.2 SIP Call with a Proxy Server
- •2.3 SIP Registration Example
- •2.4 SIP Presence and Instant Message Example
- •2.5 Message Transport
- •2.5.1 UDP Transport
- •2.5.2 TCP Transport
- •2.5.3 TLS Transport
- •2.5.4 SCTP Transport
- •2.6 Transport Protocol Selection
- •2.7 Conclusion
- •2.8 Questions
- •References
- •3 SIP Clients and Servers
- •3.1 SIP User Agents
- •3.2 Presence Agents
- •3.3 Back-to-Back User Agents
- •3.4 SIP Gateways
- •3.5 SIP Servers
- •3.5.1 Proxy Servers
- •3.5.2 Redirect Servers
- •3.5.3 Registrar Servers
- •3.6 Uniform Resource Indicators
- •3.7 Acknowledgment of Messages
- •3.8 Reliability
- •3.9 Multicast Support
- •3.10 Conclusion
- •3.11 Questions
- •References
- •4 SIP Request Messages
- •4.1 Methods
- •4.1.1 INVITE
- •4.1.2 REGISTER
- •4.1.5 CANCEL
- •4.1.6 OPTIONS
- •4.1.7 SUBSCRIBE
- •4.1.8 NOTIFY
- •4.1.9 PUBLISH
- •4.1.10 REFER
- •4.1.11 MESSAGE
- •4.1.12 INFO
- •4.1.13 PRACK
- •4.1.14 UPDATE
- •4.2 URI and URL Schemes Used by SIP
- •4.2.1 SIP and SIPS URIs
- •4.2.2 Telephone URLs
- •4.2.3 Presence and Instant Messaging URLs
- •4.3 Tags
- •4.4 Message Bodies
- •4.5 Conclusion
- •4.6 Questions
- •References
- •5 SIP Response Messages
- •5.1 Informational
- •5.1.1 100 Trying
- •5.1.2 180 Ringing
- •5.1.3 181 Call is Being Forwarded
- •5.1.4 182 Call Queued
- •5.1.5 183 Session Progress
- •5.2 Success
- •5.2.2 202 Accepted
- •5.3 Redirection
- •5.3.1 300 Multiple Choices
- •5.3.2 301 Moved Permanently
- •5.3.3 302 Moved Temporarily
- •5.3.4 305 Use Proxy
- •5.3.5 380 Alternative Service
- •5.4.3 402 Payment Required
- •5.4.4 403 Forbidden
- •5.4.5 404 Not Found
- •5.4.6 405 Method Not Allowed
- •5.4.7 406 Not Acceptable
- •5.4.8 407 Proxy Authentication Required
- •5.4.9 408 Request Timeout
- •5.4.11 410 Gone
- •5.4.12 411 Length Required
- •5.4.13 412 Conditional Request Failed
- •5.4 Client Error
- •5.4.1 400 Bad Request
- •5.4.2 401 Unauthorized
- •5.4.14 413 Request Entity Too Large
- •5.4.15 414 Request-URI Too Long
- •5.4.16 415 Unsupported Media Type
- •5.4.17 416 Unsupported URI Scheme
- •5.4.18 417 Unknown Resource Priority
- •5.4.19 420 Bad Extension
- •5.4.20 421 Extension Required
- •5.4.21 422 Session Timer Interval Too Small
- •5.4.22 423 Interval Too Brief
- •5.4.23 428 Use Identity Header
- •5.4.24 429 Provide Referror Identity
- •5.4.25 430 Flow Failed
- •5.4.26 433 Anonymity Disallowed
- •5.4.27 436 Bad Identity-Info Header
- •5.4.29 438 Invalid Identity Header
- •5.4.30 439 First Hop Lacks Outbound Support
- •5.4.31 440 Max-Breadth Exceeded
- •5.4.32 470 Consent Needed
- •5.4.33 480 Temporarily Unavailable
- •5.4.34 481 Dialog/Transaction Does Not Exist
- •5.4.35 482 Loop Detected
- •5.4.36 483 Too Many Hops
- •5.4.37 484 Address Incomplete
- •5.4.38 485 Ambiguous
- •5.4.39 486 Busy Here
- •5.4.40 487 Request Terminated
- •5.4.41 488 Not Acceptable Here
- •5.4.42 489 Bad Event
- •5.4.43 491 Request Pending
- •5.4.44 493 Request Undecipherable
- •5.4.45 494 Security Agreement Required
- •5.5 Server Error
- •5.5.1 500 Server Internal Error
- •5.5.2 501 Not Implemented
- •5.5.3 502 Bad Gateway
- •5.5.4 503 Service Unavailable
- •5.5.5 504 Gateway Timeout
- •5.5.6 505 Version Not Supported
- •5.5.7 513 Message Too Large
- •5.5.8 580 Preconditions Failure
- •5.6 Global Error
- •5.6.1 600 Busy Everywhere
- •5.6.2 603 Decline
- •5.6.3 604 Does Not Exist Anywhere
- •5.6.4 606 Not Acceptable
- •5.7 Questions
- •References
- •6 SIP Header Fields
- •6.1 Request and Response Header Fields
- •6.1.1 Accept
- •6.1.2 Accept-Encoding
- •6.1.3 Accept-Language
- •6.1.4 Alert-Info
- •6.1.5 Allow
- •6.1.6 Allow-Events
- •6.1.7 Answer-Mode
- •6.1.8 Call-ID
- •6.1.9 Contact
- •6.1.10 CSeq
- •6.1.11 Date
- •6.1.12 Encryption
- •6.1.13 Expires
- •6.1.14 From
- •6.1.15 History Info
- •6.1.16 Organization
- •6.1.17 Path
- •6.1.19 Record-Route
- •6.1.20 Recv-Info
- •6.1.21 Refer-Sub
- •6.1.22 Retry-After
- •6.1.23 Subject
- •6.1.24 Supported
- •6.1.25 Timestamp
- •6.1.27 User-Agent
- •6.2 Request Header Fields
- •6.2.1 Accept-Contact
- •6.2.2 Authorization
- •6.2.3 Call-Info
- •6.2.4 Event
- •6.2.5 Hide
- •6.2.6 Identity
- •6.2.7 Identity-Info
- •6.2.8 In-Reply-To
- •6.2.9 Info-Package
- •6.2.10 Join
- •6.2.11 Priority
- •6.2.12 Privacy
- •6.2.13 Proxy-Authorization
- •6.2.14 Proxy-Require
- •6.2.15 P-OSP-Auth-Token
- •6.2.16 P-Asserted-Identity
- •6.2.17 P-Preferred-Identity
- •6.2.18 Max-Breadth
- •6.2.19 Max-Forwards
- •6.2.20 Reason
- •6.2.21 Refer-To
- •6.2.22 Referred-By
- •6.2.23 Reply-To
- •6.2.24 Replaces
- •6.2.25 Reject-Contact
- •6.2.26 Request-Disposition
- •6.2.27 Require
- •6.2.28 Resource-Priority
- •6.2.29 Response-Key
- •6.2.30 Route
- •6.2.31 RAck
- •6.2.32 Security-Client
- •6.2.33 Security-Verify
- •6.2.34 Session-Expires
- •6.2.35 SIP-If-Match
- •6.2.36 Subscription-State
- •6.2.37 Suppress-If-Match
- •6.2.38 Target-Dialog
- •6.2.39 Trigger-Consent
- •6.3 Response Header Fields
- •6.3.1 Accept-Resource-Priority
- •6.3.2 Authentication-Info
- •6.3.3 Error-Info
- •6.3.4 Flow-Timer
- •6.3.5 Min-Expires
- •6.3.7 Permission-Missing
- •6.3.8 Proxy-Authenticate
- •6.3.9 Security-Server
- •6.3.10 Server
- •6.3.11 Service-Route
- •6.3.12 SIP-ETag
- •6.3.13 Unsupported
- •6.3.14 Warning
- •6.3.15 WWW-Authenticate
- •6.3.16 RSeq
- •6.4 Message Body Header Fields
- •6.4.1 Content-Encoding
- •6.4.2 Content-Disposition
- •6.4.3 Content-Language
- •6.4.4 Content-Length
- •6.4.5 Content-Type
- •6.4.6 MIME-Version
- •6.5 Questions
- •References
- •7 Wireless, Mobility, and IMS
- •7.1 IP Mobility
- •7.2 SIP Mobility
- •7.4 IMS Header Fields
- •7.5 Conclusion
- •7.6 Questions
- •References
- •8 Presence and Instant Messaging
- •8.1 Introduction
- •8.2 History of IM and Presence
- •8.3 SIMPLE
- •8.4 Presence with SIMPLE
- •8.4.1 SIP Events Framework
- •8.4.2 Presence Bodies
- •8.4.3 Resource Lists
- •8.4.4 Filtering
- •8.4.6 Partial Publication
- •8.4.7 Presence Documents Summary
- •8.5 Instant Messaging with SIMPLE
- •8.5.1 Page Mode Instant Messaging
- •8.5.4 Message Composition Indication
- •8.5.5 Multiple Recipient Messages
- •8.5.6 Session Mode Instant Messaging
- •8.6 Jabber
- •8.6.1 Standardization as Extensible Messaging and Presence Protocol
- •8.6.2 Interworking with SIMPLE
- •8.6.3 Jingle
- •8.6.4 Future Standardization of XMPP
- •8.7 Conclusion
- •8.8 Questions
- •References
- •9 Services in SIP
- •9.1 Gateway Services
- •9.2 SIP Trunking
- •9.3 SIP Service Examples
- •9.4 Voicemail
- •9.5 SIP Video
- •9.6 Facsimile
- •9.7 Conferencing
- •9.7.1 Focus
- •9.7.2 Mixer
- •9.8 Application Sequencing
- •9.9 Other SIP Service Architectures
- •9.9.1 Service Oriented Architecture
- •9.9.2 Servlets
- •9.9.3 Service Delivery Platform
- •9.10 Conclusion
- •9.11 Questions
- •References
- •10 Network Address Translation
- •10.1 Introduction to NAT
- •10.2 Advantages of NAT
- •10.3 Disadvantages of NAT
- •10.4 How NAT Works
- •10.5 Types of NAT
- •10.5.1 Endpoint Independent Mapping NAT
- •10.5.2 Address Dependent Mapping NAT
- •10.5.3 Address and Port Dependent Mapping NAT
- •10.5.4 Hairpinning Support
- •10.5.5 IP Address Pooling Options
- •10.5.6 Port Assignment Options
- •10.5.7 Mapping Refresh
- •10.5.8 Filtering Modes
- •10.6 NAT Mapping Examples
- •10.7 NATs and SIP
- •10.8 Properties of a Friendly NAT or How a NAT Should BEHAVE
- •10.9 STUN Protocol
- •10.10 UNSAF Requirements
- •10.11 SIP Problems with NAT
- •10.11.1 Symmetric SIP
- •10.11.2 Connection Reuse
- •10.11.3 SIP Outbound
- •10.12 Media NAT Traversal Solutions
- •10.12.1 Symmetric RTP
- •10.12.2 RTCP Attribute
- •10.12.3 Self-Fixing Approach
- •10.13 Hole Punching
- •10.14 TURN: Traversal Using Relays Around NAT
- •10.15 ICE: Interactive Connectivity Establishment
- •10.16 Conclusion
- •10.17 Questions
- •References
- •11 Related Protocols
- •11.1 PSTN Protocols
- •11.1.1 Circuit Associated Signaling
- •11.1.2 ISDN Signaling
- •11.1.3 ISUP Signaling
- •11.2 SIP for Telephones
- •11.3 Media Gateway Control Protocols
- •11.4.1 Introduction to H.323
- •11.4.2 Example of H.323
- •11.4.3 Versions
- •References
- •12 Media Transport
- •12.1 Real-Time Transport Protocol (RTP)
- •12.2 RTP Control Protocol (RTCP)
- •12.2.1 RTCP Reports
- •12.2.2 RTCP Extended Reports
- •12.3 Compression
- •12.4.1 Audio Codecs
- •12.4.2 Video Codecs
- •12.5 Conferencing
- •12.6 ToIP—Conversational Text
- •12.7 DTMF Transport
- •12.8 Questions
- •References
- •13 Negotiating Media Sessions
- •13.1 Session Description Protocol (SDP)
- •13.1.1 Protocol Version
- •13.1.2 Origin
- •13.1.3 Session Name and Information
- •13.1.5 E-Mail Address and Phone Number
- •13.1.6 Connection Data
- •13.1.7 Bandwidth
- •13.1.8 Time, Repeat Times, and Time Zones
- •13.1.9 Encryption Keys
- •13.1.10 Media Announcements
- •13.1.11 Attributes
- •13.2 SDP Extensions
- •13.3 The Offer Answer Model
- •13.3.1 Rules for Generating an Offer
- •13.3.2 Rules for Generating an Answer
- •13.3.3 Rules for Modifying a Session
- •13.3.4 Special Case—Call Hold
- •13.4 Static and Dynamic Payloads
- •13.5 SIP Offer Answer Exchanges
- •13.6 Conclusion
- •13.7 Questions
- •References
- •14 SIP Security
- •14.1 Basic Security Concepts
- •14.1.1 Encryption
- •14.1.2 Public Key Cryptography
- •14.1.4 Message Authentication
- •14.2 Threats
- •14.3 Security Protocols
- •14.3.1 IPSec
- •14.3.3 DNSSec
- •14.3.4 Secure MIME
- •14.4 SIP Security Model
- •14.4.1 SIP Digest Authentication
- •14.4.2 SIP Authentication Using TLS
- •14.4.3 Secure SIP
- •14.4.4 Identity
- •14.4.5 Enhanced SIP Identity
- •14.6 Media Security
- •14.6.1 Non-RTP Media
- •14.6.2 Secure RTP
- •14.6.3 Keying SRTP
- •14.6.4 Best Effort Encryption
- •14.6.5 ZRTP
- •14.7 Questions
- •References
- •15 Peer-to-Peer SIP
- •15.1 P2P Properties
- •15.2 P2P Properties of SIP
- •15.3 P2P Overlays
- •15.4 RELOAD
- •15.5 Host Identity Protocol
- •15.6 Conclusion
- •15.7 Questions
- •References
- •16 Call Flow Examples
- •16.1 SIP Call with Authentication, Proxies, and Record-Route
- •16.2 SIP Call with Stateless and Stateful Proxies with Called Party Busy
- •16.3 SIP to PSTN Call Through Gateways
- •16.4 PSTN to SIP Call Through a Gateway
- •16.5 Parallel Search
- •16.6 Call Setup with Two Proxies
- •16.7 SIP Presence and Instant Message Example
- •References
- •17 Future Directions
- •17.2 More Extensions
- •17.3 Better Identity
- •17.4 Interdomain SIP
- •17.5 Making Features Work Better
- •17.6 Emergency Calling
- •17.7 More SIP Trunking
- •17.9 Improved NAT Traversal
- •17.10 Security Deployment
- •17.11 Better Interoperability
- •References
- •A.1 ABNF Rules
- •A.2 Introduction to XML
- •References
- •About the Author
Preface to the Second Edition
Much has changed in the 2.5 years since the first edition of SIP: Understanding the Session Initiation Protocol was published. In 2001, SIP was a relatively unknown quantity, an upstart in the voice over IP (VoIP) and multimedia communications industry. Today, SIP is seen as the future of call signaling and telephony. It has been widely deployed by service providers and enterprises and is used casually every day by users of the dominant PC operating system. The full range of possibilities enabled by SIP is just now being glimpsed, and many more possibilities are yet to come.
One reason for this rapid acceptance is that SIP is an incredibly powerful call control protocol. It allows intelligent end points to implement the entire suite of telephony, Private Branch Exchange (PBX), Class, and Centrex services without a service provider, and without a controller or switch, for example.
The biggest driver for SIP on the Internet, however, has less to do with SIP’s signaling and call control capabilities. Instead, it is due to the extensions of SIP that turn it into a powerful “rendezvous” protocol that leverages mobility and presence to allow users to communicate using different devices, modes, and services anywhere they are connected to the Internet. SIP applications provide support for presence—the ability to find out the status or location of a user without attempting to set up a session.
Another major change in the past few years is the adoption of wireless SIP to enable multimedia IP communications. As described in the chapter on wireless, SIP is now being used both in its standard form over 802.11 wireless networks and in planned commercial Third Generation Partnership Project (3GPP) rollouts in the coming years. SIP is ideally suited for this key application.
Since 2001 SIP has also grown in terms of the specification itself. Initially, SIP was described by a single RFC with a few related RFCs and a couple of RFC extensions. In Chapter 6 alone, more than 20 SIP-related RFCs are referenced. This book attempts to put all those documents together and provide a single
xxv
xxvi |
SIP: Understanding the Session Initiation Protocol |
reference for the protocol and all its extensions. Even SIP headers and responses that were standardized in the past but are now removed (deprecated) are listed in this text, providing useful context and background. Many others are discussed that are in the final stages of standardization prior to publication as RFCs, providing an up-to-date insider’s view of the future of the protocol. In closing, I again thank my colleagues in the Internet Engineering Task Force (IETF) and at MCI for all their contributions to the development of this protocol—it has been a privilege to be a part of a group of people that have created the SIP industry. Finally, I’d like to tip my hat to two of the key inventors of SIP who continue to develop and propel its implementation: Henning Schulzrinne and Jonathan Rosenberg.
Preface to the First Edition
When I began looking into the Session Initiation Protocol (SIP) in October 1998, I had prepared a list of a half dozen protocols relating to Voice over IP and Next Generation Networking. It was only a few days into my study that my list narrowed to just one: SIP. My background was in telecommunications, so I was familiar with the complex suite of protocols used for signaling in the Public Switched Telephone Network. It was readily apparent to me that SIP would be revolutionary in the telecommunications industry. Only a few weeks later I remember describing SIP to a colleague as the “SS7 of future telephony”—quite a bold statement for a protocol that almost no one had heard of, and that was not even yet a proposed standard!
Nearly 2 years later, I have continued to work almost exclusively with SIP since that day in my position with WorldCom, giving seminars and teaching the protocol to others. This book grew out of those seminars and my work on various Internet-Drafts.
This revolutionary protocol was also the discovery of a radical standards body—the Internet Engineering Task Force (IETF). Later, I attended my first IETF meeting, which was for me a career changing event. To interact with this dedicated band of engineers and developers, who have quietly taken the Internet from obscurity into one of the most important technological developments of the late 20th century, for the first time was truly exciting.
Just a few short years later, SIP has taken the telecommunications industry by storm. The industry press contains announcement after announcement of SIP product and service support from established vendor startups, and from established carriers. As each new group and company joins the dialog, the protocol has been able to adapt and grow without becoming unwieldy or overly complex. In the future, I believe that SIP, along with a TCP/IP stack, will find its way into practically every intelligent electronic device that has a need to communicate with the outside world.
xxvii
xxviii |
SIP: Understanding the Session Initiation Protocol |
With my telecommunications background, it is not surprising that I rely on telephone examples and analogies throughout this book to explain and illustrate SIP. This is also consistent with the probability that telecommunications is the first widely deployed use of the protocol. SIP stacks will soon be in multimedia PCs, laptops, palmtops, and in dedicated SIP telephones. The protocol will be used by telephone switches, gateways, wireless devices, and mobile phones. One of the key features of SIP, however, is its flexibility; as a result, the protocol is likely to be used in a whole host of applications that have little or nothing to do with telephony. Quite possibly one of these applications, such as instant messaging, may become the next “killer application” of the Internet. However, the operation and concepts of the protocol are unchanged regardless of the application, and the telephone analogies and examples are, I feel, easy to follow and comprehend.
The book begins with a discussion of the Internet, the IETF, and the Internet Multimedia Protocol Stack, of which SIP is a part. From there, the protocol is introduced by examples. Next, the elements of a SIP network are discussed, and the details of the protocol in terms of message types, headers, and response codes are covered. In order to make up a complete telephony system, related protocols, including Session Description Protocol (SDP) and Real-Time Transport Protocol (RTP), are covered. SIP is then compared to another signaling protocol, H.323, with the key advantages of SIP highlighted. Finally, the future direction of the evolution of the protocol is examined.
Two of the recurring themes of this book are the simplicity and stateless nature of the protocol. Simplicity is a hallmark of SIP due to its text-encoded, highly readable messages, and its simple transactional model with few exceptions and special conditions. Statelessness relates to the ability of SIP servers to store minimal (or no) information about the state or existence of a media session in a network. The ability of a SIP network to use stateless servers that do not need to record transactions, keep logs, fill and empty buffers, etc., is, I believe, a seminal step in the evolution of communications systems. I hope that these two themes become apparent as you read this book and learn about this exciting new protocol.
The text is filled with examples and sample SIP messages. I had to invent a whole set of IP addresses, domain names, and URLs. Please note that they are all fictional—do not try to send anything to them.
I would first like to thank the group of current and former engineers at WorldCom who shared their knowledge of this protocol and gave me the opportunity to author my first Internet-Draft document. I particularly thank Henry Sinnreich, Steve Donovan, Dean Willis, and Matt Cannon. I also thank Robert Sparks, who I first met at the first seminar on SIP that I ever presented. Throughout the whole 3-hour session I kept wondering about the guy with the pony tail who seemed to know more than me about this brand new protocol! Robert
Preface to the First Edition |
xxix |
and I have spent countless hours discussing fine points of the protocol. In addition, I would like to thank him for his expert review of this manuscript prior to publication—it is a better book due to his thoroughness and attention to detail. I also thank everyone on the IETF SIP list who has assisted me with the protocol and added to my understanding of it.
A special thanks to my wife Lisa for the terrific cover artwork and the cool figures throughout the book.
Finally, I thank my editor Jon Workman, the series editor and reviewer, and the whole team at Artech for helping me in this, my first adventure in publishing.