Internet Communications Using SIP: Delivering VoIP and Multimedia Services with Session Initiation Protocol

Introduction

The second edition of Internet Communications Using SIP had to be rewritten almost from the ground up, because of the dramatic changes in the industry in the five years that have passed since the first edition. Some of the developments had been envisaged in the first edition, but naturally, some have not.

The Internet Has Replaced the Telephone System and the Telecommunication Networks

Since the publication in 2001 of the first edition of this book, Internet Communications Using SIP, Voice over IP (VoIP) has developed from an emerging technology to the recognized replacement of existing global telephone systems based on Time Division Multiplex (TDM) circuit switching. The Internet has also replaced the proposed connection-oriented offsprings of TDM, such as the Integrated Services Digital Network (ISDN) and the Asynchronous Transfer Multiplex (ATM) based broadband version BISDN, envisaged for the telecommunications industry by the International Telecommunications Union ITU-T standards body. TDM, ATM, ISDN, and BISDN are now history.

All wired and wireless communications are instead migrating to the Internet standards developed by the Internet Engineering Task Force (IETF). The legacy telecommunication networks, while still dominant, are recognized as a present-day cash cow only and are scheduled for replacement by IP networks.

The end-to-end nature of the Internet that places intelligence in the applications running in the endpoints and gives control to the user at the endpoints has indeed replaced TDM-based telephony with central control. The Internet has also proven to be the home network for other types of communications, information, entertainment, and data applications. To quote Jon Peterson, area director of the IETF:

The Internet is the service.

The Session Initiation Protocol Is the Standard for VoIP and Multimedia Communications

Another change from the first edition of this book is the Session Initiation Protocol (SIP), which has been adopted by practically all public VoIP service providers for wired and wireless communications. The discussions about SIP versus H.323 standardized by the ITU-T are over as well. The installed base of H.323 is considered a liability and planned for replacement by SIP sooner or later.

A global industry has emerged to take advantage of SIP and its associated IETF standards for real-time communications. More than 560 VoIP service providers have been reported [1] in early 2006, most of them using SIP-based networks. The list of SIP-based equipment (such as SIP phones, software for PCs, and mobile devices, servers, gateways, and so on) is now large and still growing. Actually, all equipment and system vendors are now supporting SIP.

Presence and Instant Messaging Are Mainstream Communications

Presence and instant messaging (IM) are now mainstream with consumers and, in the enterprise, complementing or sometimes replacing voice communications in specific situations (such as in circumstances where silence is required). Even for VoIP, presence has emerged not only as a valuable enhancement, but presence may be the dial tone of the twenty-first century.

Presence and event-based communications have enabled the integration of communications with applications. Presence and IM are discussed in Chapter 13, Presence and Instant Messaging.

The so-called IM services provided by large Internet companies, such as AOL, Apple, Google, IBM, Microsoft, Skype (not SIP-based), and Yahoo!, actually carry at present most of the public VoIP traffic between end users around the globe.

It is not far-fetched to see the IM Internet companies replacing the former telephone companies in the voice communication business. Many legacy telecommunication companies are also using VoIP to replace the internal TDM voice networks, but their VoIP services may not survive the advanced technologies deployed by the IM Internet companies and the challenge posed by peer-to-peer (P2P) communications.

Redefining Communications: Mobility, Emergency and Equal Access for the Disabled

Internet communications have been known not to be dependent on the location on the Internet. Application-level mobility based on SIP is a key component to seamless mobile communications, as discussed in Chapter 15, SIP Application Level Mobility.

Emergency calling services by users in distress using the Internet (such as 911 in the United States or 112 in Europe) are far more powerful and cost less than the Public Switched Telephone Network (PSTN) based emergency services. Internet-based emergency calling is indeed in the design stage in a number of countries. Chapter 16, Emergency and Preemption Communication Services, discusses Internet-based emergency services.

The multimedia nature of Internet communications gives hearing- and speech-impaired people the opportunity to fully participate in rich communications for work and in personal life. Chapter 17, Accessibility for the Disabled, discusses access to communications for disabled people.

The Rise of Peer-to-Peer Communications

P2P traffic has risen in the Internet since around 2000 and became the dominant part of Internet traffic by 2004. Since 2004, Skype (which is based on P2P VoIP, IM, and presence) has also become by far the dominant VoIP provider worldwide. Since P2P SIP standards work is just emerging as of this writing, Skype can be considered a prestandard P2P Internet communication service.

The reasons for the emergence of overlay networks and P2P applications and their nature are discussed in Chapter 20, Peer-to-Peer SIP, and also in Chapter 6, SIP Overview. Though the present VoIP industry is built on client-server (CS) SIP, this may significantly change. To quote David Bryan from p2p.org:

P2P SIP may change VoIP to the same extent that VoIP has changed telecommunications.

VoIP and Multimedia Communications Services Are Still Fragmented

In spite of all the technological progress, VoIP, IM, presence, and multimedia services are still a highly fragmented industry:

Telephone services based on VoIP operate as islands and can interconnect (as of this writing) using mostly the legacy Public Switched Telephone Network (PSTN). The service model is giving broadband users access to the legacy telephone system, actually a voice gateway service between the Internet and TDM. The business model of most VoIP service providers is just lower cost for legacy-style telephone service, also called PSTN over IP. The PSTN gateway services are using IP inside their networks, but users are not exposed to the rich IP services, except when all parties are on the same network.

The most successful public voice, IM, and presence service is Skype, which is not standards-based.

Walled gardens: The fragmentation of communications is still actively pursued by most mobile service providers by deploying systems where their users can get rich IP multimedia services only on their own networks. The fees to communicate between mobile service providers are a significant part of the business model, and open connectivity to the Internet (Internet neutrality) is still a hotly debated issue. Internet neutrality is also still debated by many broadband Internet access providers (such as DSL and cable companies), although we believe that enlightened government regulators in the developed countries will weigh in favor of users and open network access in general.

The proliferation of islands for communications makes them less useful the more there are, since this proliferation is in denial of Metcalf’s law that the value of a network increases with the square of the number of points attached to the network. The Internet with more than 1 billion attached endpoints has thus the highest value for communications. By contrast, the mobile phone industry boasts 3 billion users, but in many fragmented networks.

Past Obsessions and Present Dangers: QoS and Security

Network-based quality of service (QoS) for voice and the reliability of the legacy telephone network have long been used by telephone industry marketers to scare users away from VoIP. In the meantime, all public VoIP services have proven that Internet best-effort QoS works just fine, as long network congestion is avoided. Internet-based voice can actually be much better than the 3.1 kHz voice over the PSTN. As for reliability, all recent major man-made and natural disasters have proven the Internet and VoIP to be more resilient than the existing wireline and wireless telephone networks.

Chapter 18, Quality of Service for Real-Time Internet Communications, is aimed at a balanced approach for QoS, and Chapter 16, Emergency and Preemption Communication Services, discusses the Emergency Services based on SIP.

The security threats on the Internet have provided well-justified concerns about the security of VoIP, and even more, the security of IM. As a result, a new industry niche, that of VoIP and IM security, has sprung up and, as usual, marketers are first drumming up the vulnerabilities of Internet communications to prepare the sell for all kinds of security products. Though no significant security breaks have been reported so far for Internet communications, security for VoIP and IM is still work in progress. Chapter 9, SIP Security, deals with SIP security.

References

[1] A list of VoIP companies is provided at www.myvoipprovider.com.

Chapter 1

Introduction

The telecommunications, television, and information technology (IT) network industries are all transformed by the Internet. The transformation is driven by the need for growth based on new services, more complete global coverage, and consolidation. In this chapter, we will explore some of the problems and solutions for end users and every type of business because of the profound disruptions caused by the Internet.

Problem: Too Many Public Networks

Before the emergence of the Internet, users and service providers were generally accustomed to thinking in terms of four distinct network types: Networks for IT (data), networks for voice, mobile networks, and networks for television. Each of these dedicated network types could, in turn, be divided into many incompatible regional and even country-specific flavors with different protocol variants.

Thus, we find many types of telephony numbering plans, signaling, and audio encodings; several TV standards; and various types and flavors of what the telecom industry calls data networks—all of them incompatible and impossible to integrate into one single global network.

The mobile telephone networks have converged on a smaller number of standards in the second generation (2G) networks and in the emerging third generation (3G) mobile networks. It may turn out, however, that with the proliferation of new radio technologies for the so-called 4th generation (4G), such as Wi-Fi and WiMAX, all modern mobile networks will become just a wireless access mechanism to the Internet, where all public communications, entertainment, and applications will reside anyhow.

Data networks that originated in the telecom industry came in many forms, such as digital private lines, X.25, Integrated Services Digital Network (ISDN), Switched Multimegabit Data Service (SMDS), Frame Relay, and Asynchronous Transfer Mode (ATM) networks. These so-called data networks were mostly inspired by circuit-switched telephony concepts. Their names are meant to suggest that they were not designed primarily to carry voice.

Voice networks are still used for data and fax because of their general availability, though less and less so. However, these networks have come to the end of their evolution, since they are fundamentally optimized for voice only. TV networks were designed and optimized for the distribution of entertainment video streams.

Needless to say, all network types (data, voice, TV, and mobile) have specific end-user devices that cannot be ported to other service providers or network types, and most often cannot be globally deployed.

The impact of the Internet has made the wired and wireless phone companies and the TV cable companies look for new business models that can take advantage of Internet technologies and protocols, among them the Session Initiation Protocol (SIP) for real-time communications, such as Voice over IP (VoIP), instant messaging (IM), video, conferencing/collaboration, and others. Examples of the various categories and their business models are illustrated in Table 1.1. We assume that most readers are familiar with the acronyms used in the table, and we also explain these acronyms and terms in the book. They can also be found in the index.

Table 1.1 Internet Communications in 2005 with Examples from North America

The proliferation of isolated communication islands as shown in Table 1.1 makes them less useful as their number keeps increasing (think of many more communication islands all over the world). Building communication islands (also called walled gardens) is in conflict with Metcalfe’s law that the value of the network increases by the square of the number of connected endpoints. Last, but not least, in case of an emergency, having many networks that cannot communicate directly is not very helpful.

Closed networks are an impediment for innovation, since innovators must work (technology and legal agreements) with every closed network separately to bring a new service or product to market. By contrast, the Internet extends the reach for new applications and services instantly to the whole world.

Another observation from Table 1.1 is that the strongest financing available is at present for closed networks (walled gardens), the ones that are most limited in reach and usefulness. This raises business issues and regulatory questions (what are the public interest obligations, if any?) that are beyond the scope of this book.

Incompatible Enterprise Communications

Enterprise communication systems are often an even greater mix of incompatible and disjoint systems and devices:

Proprietary PBX and their phones. Phones from one PBX cannot be used by another.

Instant messaging is a separate system from the PBX.

Various IM systems don’t talk to each other.

Voice conferencing and web-based collaboration use yet other systems.

Maintaining various incompatible and nonintegrated proprietary enterprise systems is quite costly and reduces the overall productivity of the workforce.

Network Consolidation: The Internet

The Internet has benefited from a number of different fundamentals compared to legacy networks, such as the tremendous progress of computing technology and the open standard Internet protocols that define it. This progress can be attributed to the expertise of the research, academic, and engineering communities whose dedication to excellence and open collaboration on a global basis have surpassed the usual commercial pressure for time-to-market and competitive secrecy.

The result is an Internet that uses consistent protocols on a global basis, and is equally well suited to carry data, transactions, and real-time communications, such as instant messaging (IM), voice, video, and conferencing/collaboration. Actually, the Internet is the dumb network, designed for any application, even those not yet invented. This is in stark contrast to the isolated walled gardens with central control of all services illustrated in Table 1.1.

Voice over IP

Although the Internet has quickly established itself as the preeminent network for data, commercial transactions, and audio-video distribution, the use of voice over the Internet has been slower to develop. This has less to do with the capability of the Internet to carry voice with equal or higher quality than the telephone network but rather with the complex nature of signaling in voice services, as you will see in Chapter 6, SIP Overview.

There are various approaches for voice services over the Internet, based on different signaling and control design. Some examples include the following:

Use signaling concepts from the telephone industry—H.323, MGCP, MEGACO/H.248.

Use control concepts from the telephone industry—central control and softswitches.

Use the Internet-centric protocol—Session Initiation Protocol (SIP), the topic of this book.

The movement from such concepts as telephony call models to discovery/rendezvous and session setup between any processes on any platform anywhere on the Internet is opening up completely new types of communication services.

The use of SIP for establishing voice, video, and data sessions places telephony as just another application on the Internet, using similar addressing, data types, software, protocols, and security as found, for example, on the World Wide Web or e-mail.

Separate networks for voice are no longer necessary, and this is of great consequence for all wired and wireless telephone companies.

Complete integration of voice with all other Internet services and applications probably provides the greatest opportunity for innovation. The open and distributed nature of this service and the dumb network model will empower many innovators, similar to what has happened with other industries on the Internet and the resulting online economy.

Most IM systems on the Internet already have voice and telephony capability as well, though if it is proprietary, they cannot intercommunicate without IM gateways, although IM gateways inevitably cannot translate all the features from one system to another. IM gateways are also transitory in nature, since any changes to a proprietary IM protocol may render the gateway close to useless. By contrast, SIP-based communications offer a global standards-based approach for interoperability for presence, IM, voice, and video, as we will show in the following chapters.

Presence—The Dial Tone for the Twenty-First Century?

Unsuccessful telephone calls are a serious drag on productivity and a source of frustration, since both parties waste time and talk to voicemail instead to each other. Also, the timing of the phone call may not be appropriate or not reach the called party in a suitable location. The advent of presence, so well-known from IM systems, can provide much more rich information before trying to make a call in the first place, compared to just hearing the dial tone. Another convenience of SIP and presence is that many contact addresses may reside beneath a buddy icon, so the caller need not to know or worry about picking the right phone number or URI. Presence may, therefore, replace the dial tone used in telephony for well over 100 years.

The Value Proposition of SIP

SIP is not just another protocol. SIP redefines communications and is impacting the telecom industry to a similar or greater degree than other industries. This has been recognized by all telecom service providers and their vendors for wired and wireless services, as well as by all IT vendors. Chapter 2 will provide an overview of how the Internet and SIP are redefining communications.

SIP Is Not a Miracle Protocol

As discussed in Chapter 2, Internet Communications Enabled by SIP, SIP is not a miracle protocol and is not designed to do more than discover remote users and establish interactive communication sessions. SIP is not meant to ensure quality of service (QoS) all by itself or to transfer large amounts of data. It is not applicable for conference floor control. Neither is it meant to replace all known telephony features, many of which are caused by the limitations of circuit-switched voice or to the regulation of voice services. And such a list can go on.

Various other Internet protocols are better suited for other functions. As for legacy telephony, not all telephone network features lend themselves to replication on the Internet.

The Short History of SIP [1]

By 1996, the Internet Engineering Task Force (IETF) had already developed the basics for multimedia on the Internet (see Chapter 14, SIP Conferencing) in the Multi-Party, Multimedia Working Group. Two proposals, the Simple Conference Invitation Protocol (SCIP) by Henning Schulzrinne and the Session Initiation Protocol (SIP) by Mark Handley, were announced and later merged to form Session Initiation Protocol. The new protocol also preserved the HTTP orientation from the initial SCIP proposal that later proved to be crucial to the merging of IP communications on the Internet.

Schulzrinne focused on the continuing development of SIP with the objective of re-engineering the telephone system from ground up, an opportunity that appears only once in 100 years, as we heard him argue at a time when few believed this was practical.

SIP was initially approved as RFC [2] number 2543 in the IETF in March 1999. Because of the tremendous interest and the increasing number of contributions to SIP, a separate SIP Working Group (WG) was formed in September 1999. The SIP for Instant Messaging and Presence Leveraging (SIMPLE) was formed in March 2001, followed by SIPPING for applications and their extensions in 2002. The specific needs of SIP developers and service providers have led to an increasing number of new working groups. This very large body of work attests both to the creativity of the Internet communications engineering community, and also to the vigor of the newly created industry.

We will shorten the narrative on the history of SIP by listing the related working groups (WG) in chronological order in Table 1.2. We have listed for simplicity the year of the first RFC published by the WG, though the WG was sometimes formed one to two years earlier. Years denote a new WG that has not yet produced any RFC.

Table 1.2 History of SIP-Related Working Groups

The growth of SIP-related standards in the IETF is illustrated and discussed in Chapter 21, Conclusions and Future Directions.

References in This Book

Because of the multiple developments on the Internet, SIP is being used in ever-more services, user software, and various user devices (such as in SIP phones, PCs, laptops, PDAs, and mobile phones). This is, in effect, a new industry and its participants keep making new contributions to the core SIP standards, mainly in the area of new services and new applications. This book reflects SIP developments up to and including the 64th IETF in November 2005.

We have included, by necessity, many Internet drafts that are designated work in progress, since they are the only reference source for this particular information. Some of these drafts may become standards by the time you are ready to use them; some may be a work in progress and have a higher version number than quoted as of this writing; and still others may be found only in an archive for expired drafts.

The SIP WG drafts that are work in progress can be found online at the IETF web site:

http://ietf.org/html.charters/sip-charter.html

Additional individual submissions and Internet drafts from other working groups can be found at the following site:

http://ietf.org/ID.html

SIP-related drafts that have expired (older than six months) can be found on several archives. As of this writing, following are some of the sites:

www.cs.columbia.edu/sip/drafts www.softarmor.com/sipwg

Readers may also perform a web search, such as Google, for any IETF SIP-related topic or for any Internet draft or RFC.

Several books have been published on Internet multimedia, Voice over IP, and SIP, some of which are listed here. [3], [4], [5] They focus mainly on how SIP works. This book is less about explaining how SIP works, but rather what it does and the new communications and services it enables.

We have reproduced some of the exciting services and features discussed in the IETF SIP WG and its main offsprings, the SIPPING and SIMPLE Working Groups. Also included in our discussion are some drafts from Bird of Feather (BOF) sessions that have not even made it to an accepted WG charter, such as the peer-to-peer (P2P) SIP group. [6]

Many of these expired proposals may not develop into IETF standards for various reasons, but represent good work, often backed up by running code. The references to such expired Internet drafts are intended to make you aware of these ideas that may otherwise remain buried in an archive. Such references are clearly marked as expired, so as to distinguish them from accepted work in progress items of IETF WGs that are on the path toward acceptance as standards.

SIP Open Source Code and SIP Products

There is an ever-increasing amount of open source code for SIP, and it is increasing in quality. Most or many commercial SIP products are actually based on open source SIP code. An authoritative list of SIP open source code is available from the SIP Forum:

www.sipforum.org

The SIP Forum is also an excellent source for finding commercial SIP software and products for the enterprise, for consumer products, for service providers, tools, and others.

Excellent lists of SIP products are also maintained on the web sites of Pulver.com and Ubiquity.com:

www.pulver.com/products/sip www.sipcenter.com

References for Telephony

We assume throughout this book some understanding of telephone services and of telecommunication protocols. There is a vast literature pool available on telephony and telecommunications. We refer you to Newton’s Telecommunications Dictionary [7] to brush up on various terms that will be used in the following chapters.

Summary

This chapter has discussed some of the problems and solutions to the communications industry by the Internet, and also a brief history of the SIP protocol.

During the migration from circuit-switched telephony to IP-based communications, there are too many isolated wired and wireless communication networks, even though most (but not all) are converging on SIP. SIP has undergone a 10-year development as a standard and in implementations in the marketplace.

By adopting the Internet as The Network with wired and wireless access, and SIP as the standard protocol, rich global communications are taking shape.

The old dial-tone in telephony may well be replaced by presence information, and rich multimedia will replace the narrowband voice communications used in circuit-switched telephony.

References

[1] The authors would like to thank Professor Dr. Jörg Ott, co-chair of the SIP WG and early contributor to the MMUSIC WG for helping with data on SIP history.

[2] RFC stands for Request for Comments and many of them are Internet standards.

[3] SIP: Understanding the Session Initiation Protocol, 2nd Edition, by Alan B. Johnston, Artech House, 2003.

[4] SIP Demystified by Gonzalo Camarillo, McGraw-Hill, 2001.

[5] SIP Beyond VoIP by Henry Sinnreich, Alan B. Johnston, and Robert J. Sparks, VON Publishing, 2005. www.vonmag.com/books.

[6] See the web site for P2P SIP at www.p2psip.org.

[7] Newton’s Telecommunications Dictionary, 17th edition by Harry Newton, CMP Books, March 2001.

Chapter 2

Internet Communications Enabled by SIP

This chapter provides a short overview of the topics that are discussed in more detail in Chapters 4–20.

The Internet challenges and transforms the more than one-trillion-dollar-per-year business of telecommunications. A renaissance in communications is taking place on the Internet. At its source are new communication protocols that would be impractical on the centralized control systems of ITU-T type networks used in telecommunications. Internet communications can benefit from the IP soft state and connectionless nature of the Net, and at the application layer of the IP protocol stack from its associated addressing and data representations. Users and Internet service providers (ISPs) are reaping the benefit from standards that allow interoperability with all connected parties on a global scale. The end-to-end (e2e) nature of the Internet avoids the friction of having intermediaries between the communicating parties, and also avoids the breaking of applications and security by intermediaries in the network.

While it is not possible to forecast technology and services, it is already apparent that the Internet and web technology have created an unprecedented toolkit for new applications. However, these new applications are hard to predict, just as presence and instant messaging were not predicted in the telecom world. What can be shown, however, are some of the capabilities of the technology that are presently well understood in already established services. New Internet communication services may create new revenue opportunities for Internet service providers and their suppliers of applications.

This chapter refers to many legacy telephony services.

Readers may consult Newton’s Telecommunications Dictionary [1] for definitions of the telephony and telecom services mentioned here. Chapter 11, SIP Telephony, also discusses in detail many enhanced telephone services.

The overview of SIP services provided here reflects current thinking in the community of SIP service and technology developers. Most (but not all) of them have been actually tested and implemented. Some proposed Internet drafts on SIP will make it to the level of IETF standards; some will not. It also is likely that new technologies and services will emerge that have not been made public or envisaged as of this writing.

Internet Multimedia Protocols

Networks are defined by their protocols. The global telephone network uses its own signaling and communication protocols, as do other telecom networks such as the Public Switched Telephone Network (PSTN), X.25, Integrated Services Digital Network (ISDN), Switched Multimegabit Data Services (SMDS), frame relay, Asynchronous Transfer Mode (ATM), mobile circuit-switched networks, and the (seemingly always) proposed ITU-T Next Generation public Networks (NGN). Besides legacy network protocols, there are also application-level protocols, such as those used between fax machines.

Though started with much smaller resources than the previously dominant telecom and non-IP data networks (SNA, DECnet, Novell), the Internet’s success is solely due to its well-designed architecture and protocols. The architectural principles of the Internet (covered in Chapter 3, Architectural Principles of the Internet) have made it the most effective network for any type of application, including real-time communications.

Internet telephony and the wider family of Internet communications are defined by several key application level protocols. The list of Internet protocols used for interactive communications is shown in Table 2.1.

Table 2.1 Key Standard Internet Multimedia Protocols

The nature of interactive communications and the type of service are determined by the signaling used for establishing the communication, hence the name value of signaling.

The Value of Signaling

Signaling in telephone systems is the key mechanism by which telephone calls are set up and terminated. For example, signaling from a desktop business phone tells the PBX to forward the call to another phone. In the public telephone network, signaling instructs the switching systems to forward an 800 call to a specific call