Unknown to the majority of the world, on August 30, 1969, several dozen of researchers, engineers, and students at UCLA managed to send the first bits of information between an SDS Sigma 7 computer acting as a host and a Honeywell DDP-516 acting as an IMP (Interface Message Processor). The IMP was the first node of the ARPANET, the network that would later become the Internet. At the time, little did these pioneers realize that they were giving birth to a reality that would change the lives of millions of people and become one of the most relevant sociological phenomena of the end of the 20th century and beginning of the third millennium.

At the very beginning, ARPANET was conceived as a tool for resource sharing among scientists and researchers, and the first nodes of the network were deployed in universities. However, it soon became apparent that the nature of the traffic was different from what ARPA (which funded the project) had expected. The scientists in the network were using it predominantly for personal communication via electronic messages. In 1972, Ray Tomlinson, an engineer at Bolt Beranek and Newman (today known as BBN Technologies),^[2] wrote the first email program. Tomlinson is regarded as the inventor of email. He designed an addressing format that required a symbol in order to separate the name and the location of the computer of the recipient. When trying to discern what the best choice was for such a symbol, Tomlinson looked at the keyboard in front of him and came up with a quick decision: “The one that was most obvious was the @ sign, because this person was @ this other computer,” he later explained. “At the time, there was nobody with an @ sign in their name that I was aware of.”

By 1978, 75% of the ARPANET traffic was email and when ARPANET became Internet at the beginning of the 1980s with the adoption of the TCP/IP protocol suite, email became popular even among nonscientists. The first TCP/IP-based email protocol standard, SMTP (Simple Mail Transfer Protocol), which was instrumental to the success of email, was published in 1981. It is still in use today.

Email was the first revolution within the Net, and it is a revolution whose effects have stretched to the present day.

The second revolution in the Internet started in 1990, when Tim Berners-Lee, a British physicist working at the CERN (European Organization for Nuclear Research) facility near Geneva, Switzerland, conceived the World Wide Web (WWW), an application running on top of TCP/IP. Before that, the Internet provided only screens full of text—and not in a very aesthetic way.

The Web was made out of four key building blocks:

It was actually this last concept that gave the web its name. As Berners-Lee said, “There was a power in arranging ideas in an unconstrained, web-like way.”

In fact, though the web revolution started in 1990, it truly materialized three years later, when Marc Andreessen and Eric Bina created a program called Mosaic, which made web browsing very easy and graphically intuitive. Also in 1993, CERN declared that they would claim no fees for the use of the technology. This proved to be a visionary decision—by 1994, there were a million browser copies in use.

That year signaled the beginning of the Internet era. From that timeonward, the Internet has experienced a massive growth, the likes of which had never been seen before.

Soon huge amounts of information started to move to servers. Initially, people used the web just to display information—but then came the search engines, followed by applications such as online shopping. Yahoo! and Amazon started in 1994, and Google in 1998.

In 1994, the web grew by an incredible 2,300 percent. By 1998, there already existed 750,000 commercial sites. From then, the Internet started to change people’s lives—the way they worked, communicated, managed their bank accounts, bought cinema tickets, arranged trips, looked for jobs, enjoyed themselves, went to the supermarket, just to name a few examples. And even today, with the advent of Web 2.0, a new cultural shift around the web is being materialized. Web-based communities and social networking sites are positioning end users as key players in the generation of content. The new environment based on collaboration and sharing of information between end users is boosting innovation on the web. The web revolution was the second Internet revolution.

By that time, Vinton Cerf, one of the Internet’s principal designers, said: “Revolutions like this don’t come along very often.” However, the third Internet revolution was already being conceived. In December 1996, the first Internet Draft (I-D) of the Session Initiation Protocol (SIP)—authored by Mark Handley, Henning Schulzrinne, and Eve Schooler—was submitted to the IETF in San Jose, California. SIP was conceived as a signaling protocol for inviting users to multimedia conferences. With this, the third Internet revolution silently started. This was the revolution that would end up converting the Internet into a Total Communications system that would allow people to talk to each other, see each other, work collaboratively, or send messages in real-time. Internet telephony and, in general, Internet multimedia, is the new revolution today, and SIP is, behind the scenes, the key protocol that is materializing this revolution.

In the same way that SMTP and HTTP made Internet email and the web, respectively, possible, so SIP is also the key enabler for the new era of Internet multimedia. Before SIP, multimedia transmission on IP networks was already a reality; however, SIP brought all the power of the signaling concept to IP networks in order to enhance the multimedia communications experience. To quote Vinton Cerf again: “SIP is probably the third great protocol of the Internet, after TCP/IP and HTTP.”

One of the interesting things about the SIP standard and Internet multimedia ideas is that the revolution that they entail is even more far-fetched than the previous ones. It is not only changing the way people use the Internet, but is also driving the evolution of the telecommunications industry. The concepts behind Internet multimedia and SIP are revolutionizing the traditionally closed telecom environment.

From their inception, the telecommunication networks (e.g., the PSTN,^[3] or Public Switched Telephone Network) were designed to carry primarily voice. When data standards were born, different data networks, such as X.25 and frame relay, were deployed alongside those already existing for voice. This led to the telecommunication operators having to design, deploy, and maintain separate networks for voice and for data. In the same way, the end users needed separate infrastructures for accessing each of these disparate networks.

A first attempt to provide a converged (voice and data) access for the end user came with the ISDN (Integrated Services Digital Network), which, in addition, brought forward the digitalization of the telephone network. The ISDN never became widespread, and the new applications that it provided, such as video-telephony, never gained extensive adoption. Interestingly enough, at the end of the 1990s, ISDN technology did raise some interest, but as a mere access method to the Internet.

In the early 1990s, the GSM (Global System for Mobile communications)^[4] standard was introduced. GSM was based extensively on ISDN and used circuit switching. This had a tremendous success that continues to this day. In some western European countries, mobile subscriptions now exceed the number of fixed lines. Thanks to GSM, mobile telephony has changed our lives and has become a sociological phenomenon—most people think of the mobile phone as an indispensable element in their lives. The flourishing business within the GSM industry has driven the evolution of mobile technology significantly faster than that of the fixed networks. In order to demonstrate this statement, the reader need only compare the GSM terminals of 1992 with the fancy smart phones of today, which feature a number of technologies such as GSM, GPRS (General Packet Radio Service),^[5] UMTS (Universal Mobile Telecommunications System),^[6] XHTML (Extensible Hypertext Markup Language), MMS (Multimedia Messaging Service), WLAN (Wireless Local Area Network), Bluetooth, MP3, and IMS (Internet Multimedia Subsystem). In no other field of engineering has such a huge technological leap taken place in such a short period as in the technology for the mobile terminals.

Following GSM, a number of technological advances have occurred in the remit of mobile networks, such as GPRS and UMTS. With UMTS, the so-called third generation of mobile networks, we have seen a number of significant improvements—for example, an important increase of bandwidth in the radio access through the utilization of CDMA (Code Division Multiple Access)^[7] techniques or the adoption of a converged packet-switched network to converge the transport of both voice and data traffic through the UMTS Release 4 split architecture. The UMTS Release 4 architecture for the core network moves away from the circuit-switched technology that GSM embraced, and adopts the convergence ideas based on a packet-switched backbone.

As important a step as this may be, it does not represent a revolution, but rather, an evolution where the fundamental concepts regarding central intelligence in the network and coupling between call-control signaling and media description still remain mostly unchanged. From PSTN to ISDN, from ISDN to GSM, and from GSM to UMTS Release 4, the paradigm continued to be basically the same, and the new network was always designed with the clear objective of smoothing the integration with the existing one. With UMTS Release 5, which includes the IMS, this trend is broken, and a true revolution ensues. The IMS is designed with the Internet paradigm in mind, and its objective is to bring the Internet flexibility for delivering multimedia services into the mobile handsets. It is believed that by merging the two most successful technologies of the past 15 years, synergies will be created and the amount of business greatly increased. Moreover, 3GPP (3rd Generation Partnership Project), the standardization body behind UMTS and IMS, took the design decision to base the IMS architecture around the SIP structure, and it is impossible to understand the next-generation IP multimedia networks without understanding what functionality SIP provides. Thus the new IMS telecommunication network no longer resembles the previous telecom networks, but now resembles the Internet, and uses SIP as the basis for that convergence!

Meanwhile, what about the fixed-operator networks? Several attempts to evolve their fixed networks into the packet-switched domain, such as the ATM (Asynchronous Transfer Mode)^[8]–based B-ISDN (Broadband Integrated Services Digital Network),^[9] have failed ostentatiously. These approaches still attempted to use the paradigm of old telecom networks even though they were using a packet-based transport. In 2005, the ITU (International Telecommunication Union) declared the principles for the NGN (Next Generation Network), embracing the end-to-end IP connectivity and the Internet principles [ITU Y.2001] and [ITU Y.2011]. Today, ETSI (European Telecommunications Standards Institute), after merging the TIPHON (Telecommunications and Internet Protocol Harmonization Over Networks) and SPAN (Services and Protocols for Advanced Networks) Working Groups and creating the TISPAN (Telecoms and Internet converged Services and Protocols for Advanced Networks) group, is materializing the NGN concept. As it happens, Release 1 of the TISPAN NGN architecture has already been published, and is based largely on the IMS core network as defined by 3GPP. Thus, also in the remit of fixed networks, we see SIP being adopted as the base protocol and the architecture for delivering the next-generation multimedia services, but also for replacing the still-existing PSTN (in the so-called ETSI TISPAN PSTN emulation approach).

Summarizing, SIP was not created to evolve the telecom networks nor to make them converge. SIP was born in a pure Internet environment—at the beginning, mainly as a medium to invite users to participate in multimedia conferences. However, SIP has evolved, and what we are seeing today is not only how SIP is revolutionizing the Internet, but also how SIP is about to change for all time the old telecom environment.

The recent adoption of the Voice over Internet Protocol (VoIP) technology in the Internet by the mass market might lead many people to believe that the technology that allows the transmission of voice on packet-based networks is relatively new. That is, however, not the case.

Soon after the birth of ARPANET in 1969, several students and researchers who were already familiar with the tools to transmit text and data on that network started to explore the possibility of using it to carry real-time voice. In August 1974, the first real-time packet voice transmission was demonstrated in ARPANET between the University of Southern California (USC) and the Massachussets Institute of Technology (MIT). This was not yet “Voice over IP” because the Internet Protocol (IP) had not yet been invented; however, these experiments demonstrated that voice communication over packet networks was feasible. The protocol used for this experiment was called NVP (Network Voice Protocol), and was built on top of the protocol that ARPANET used by that time, which was known as “BBN Report 1822.” The NVP specification would later, in 1977, become [RFC741].

More experiments of this sort followed during the rest of the 1970s between universities connected to ARPANET in the United States. In 1978, the first intercontinental experiment of Voice over Packet (VoP) networks was conducted between the United States, the United Kingdom, and Norway. Also by that time, the first patents were granted in connection with this technology. Although these experiments were successful, they gave birth to no product that might be widely used.

During the 1980s, ARPANET became Internet, and workstations increased their processing power so as to allow voice transmission between two workstations connected to the Net. Researchers at universities created a number of tools for Internet telephony; however, these tools were never widespread. At the time, few people had access to workstations connected to the Internet, and the tools were too difficult to use. On top of that, the Internet of the 1980s did not yet have the necessary capacity (the web revolution had not yet occurred), and voice quality was an issue.

By the beginning of the next decade, things started to change. New tools were released: the VAT (Visual Audio Tool), VT (Voice Terminal), and VTALK. VAT was using version 0 of a protocol, RTP (Real-time Transport Protocol), that would later become the standard for voice and video transport on IP networks. And in 1991, the first audiocast in MBone (Multicast Backbone)^[10] took place in San Diego, California.

From then onward, the experiments in MBone helped to improve RTP, a new version of which, version 1, was produced with the appearance of the first drafts in the IETF AVT Working Group.

At that time, the web revolution (with the release of Mosaic), together with the commercialization of the Internet and the creation of new ISPs (Internet Service Providers), made the Internet more and more popular and increased the capacity of the Net. This was crucial for the success of Internet telephony because in order to obtain acceptable voice quality in packet networks, these must have excess capacity.^[11] Another key enabling factor for the take-up of Internet telephony was the growth in the processing power of PCs and the proliferation of multimedia-enabled PCs and high-speed modems.

A key year in the history of Internet telephony was 1995. Before that time, no commercial product had really yet appealed to the market, but in 1995, more than a dozen commercial products were released. The first of these was the “Internet Phone,” released by the Israeli company VocalTec in March 1995. In less than one year, more than 600,000 downloads of VocalTec’s trial software took place. Ordinary people started to believe that the Internet could also be used to provide real-time voice calls with acceptable sound quality. Other products were Maven, CU-SeeMe, and or Netphone.

In the remit of the IETF, 1995 saw version 2 of the RTP specification become RFC 1889. Also in 1995, Eve Schooler developed the Multimedia Conference Control (MMCC), a tool that provided point-to-point and multipoint teleconferences, including audio, video, and whiteboarding. Three years before, in 1992, Thierry Turletti had developed the INRIA Videoconferencing System (IVS). It was a software system for audio- and videoconferencing on the Internet, and incorporated PCM (Pulse-Code Modulation), ADPCM (Adaptive Differential Pulse-Code Modulation), and H.261 software-based codecs.

These first multimedia systems gave way to the design of the first version of the Session Invitation Protocol (SIPv1) by Mark Handley and Eve Schooler, in February 1996. By that time, the ideas around Internet multimedia were becoming clear with the growing usage of MBone, and new protocols were being defined. One particular bit that was still missing was a mechanism to invite people to participate in an MBone session. In MBone, if som one wanted to know what conferences would be multicast on a particular day, there already existed the SAP (Session Announcement Protocol) [RFC 2974], which announced the sessions (one could compare it with a TV program guide). However, if someone was listening to an interesting videoconference in the MBone and wanted to invite a friend, there was no mechanism for doing that. That was, really, the original purpose of the Session Invitation Protocol (hence the name) when it was submitted to the IETF in 1996.

On February 22, 1996, Henning Schulzrinne submitted an I-D to the IETF specifying the Simple Conference Invitation Protocol (SCIP), which was also conceived as a mechanism for inviting users to point-to-point and multicast sessions.

At the 35th IETF meeting in Los Angeles, it was decided to merge the two protocols into a new protocol—one that kept the SIP acronym, but now stood for Session Initiation Protocol. Also, the second version of the protocol, SIPv2—co-authored by Hanley, Schulzrinne, and Schooler—was submitted to the IETF at the 37th meeting in December 1996.

That same year, another relevant event took place. The first version of H.323 was released by the ITU. H.323 was meant to be a packet-based, LAN-only standard for audiovisual conferences. Unlike SIP, which was quite limited in scope and followed from the very beginning the Internet principles and ideas, H.323 had a very big scope and took a lot of ideas from existing ITU protocols from the ISDN remit, which resulted in unnecessary heaviness.

One area where the International Telecommunication Union adopted the Internet Standards was in the transport protocols, where the ITU embraced IETF RTP within the H.323 framework.

In the years from 1996 to 2000, several important things occurred. First of all, interest in H.323 grew, especially after VocalTec and Cisco founded the Voice Over IP Forum in order to set the standards for VoIP products. The industry tried to tune H.323 v1 into the specificities of the most popular WAN (Wide Area Network) environment: the Internet. In February 1998, version 2 of H.323, now renamed Packet-Based Multimedia Communications Systems, was approved by the ITU.

In the meantime, in the IETF, the multimedia concepts on the Internet were well understood, and work started at the same time to extend the Internet multimedia architecture for use in telephony—and, by doing so, to re-engineer the telephone system. Due to the enormous complexity and richness of the voice services on the PSTN, this was not an easy job. It was not until the end of 2000 that this matter was well understood. The years from 1997 to 2000 saw new SIP drafts being submitted before March 17, 1999, when it became a proposed standard and was published as RFC 2543 in the remit of the MMUSIC Working Group. In addition to the SIP specification itself, many other drafts were published during this period. These drafts were aimed at extending SIP in order to address many of the voice services that were already present in the PSTN, and also new ideas such as presence, Instant Messaging (IM), and so on.

By 2000, it was clear to the research and academic community that “H.323-based VoIP networks could not deliver the IP telephony service on a par with feature rich existing networks” [draft-tiphon-background]. On the other hand, interest in SIP was growing, especially when the 3GPP adopted SIP as the main protocol for the establishment of multimedia sessions within the IMS. According to the first version of 3GPP TS 22.228, which set up the requirements for the new IMS concept:

From then, SIP has grown considerably, evolving from the initial MMUSIC Working Group. Two additional WGs devoted to SIP have been created in IETF (SIP and SIPPING). In June 2002, a revised version of the standard was published as RFC 3261 by the SIP WG. Since then, the SIP capabilities have been expanded to incorporate not only voice and video, but also presence, IM, data sharing, and so forth, so that SIP has become the key IP signaling protocol for enabling a true real-time peer-to-peer total communications experience for end users.

A completely new paradigm for VoIP, different from H.323 and SIP, was born in 1999, when some new and incumbent carriers declared their urgent need to converge their voice and data infrastructure. The industry response was the IETF MGCP (Media Gateway Control Protocol), later to become the IETF and ITU MEGACO (Media Gateway Control). MEGACO gave birth to the concept of the call server, which has now been implemented in the networks of quite a few operators.

Whereas H.323 is already considered legacy, and SIP clearly represents the future-proof approach that is already a reality today, the MEGACO concepts have found their way into some interim architectures—for example, the 3GPP R4 core network split architecture conceived as an evolution of the circuit-switched domain for GSM/UMTS operators.

Already in 2007, there are millions of Internet users subscribed to SIP-based VoIP or messaging services. Very popular communicator applications such as Microsoft’s or Yahoo!’s are based on the Session Initiation Protocol, and software providers are clearly betting on SIP. Also, most companies are looking into this protocol in order to cover their enterprise communication needs—not only voice, but also video, presence, IM, whiteboarding, and so on. Moreover, open-source initiatives in the SIP domain have been extremely successful—for example, the Asterisk IP PBX (Private Branch Exchange), which is being adopted by more and more corporate customers every year.

What’s more, the traditional fixed-network providers are now starting to move toward SIP in order to offer their corporate and residential customers advanced services at the same time that they pave the way for the PSTN replacement.

Even though starting with VoIP and IM, SIP is advancing at a rapid pace in order to deliver the promise of unified IP communications.

This review of the past history of Internet and multimedia communications has allowed us to better comprehend the state of today’s technology. Now our focus will turn toward the technology itself. In the next chapter, we will try to give the reader the global picture of multimedia communications, highlighting the different Internet Protocols that contribute to making it possible. This will set the scene for the rest of the book.