Two main signaling standards exist for multimedia communications:
- H.323, the ITU's umbrella standard for audio, video and data-sharing over packet (IP) networks, is already widely deployed and has significant market share.
- SIP,the IETF's multimedia signaling protocol, is gathering attention for its simplicity and modularity, but has less commercial deployment to date.
Both H.323 and SIP were originally developed to provide multimedia services across IP networks. Both run over IP, and use TCP and UDP sessions for signaling and the Real Time Protocol (RTP) for transmitting the voice/video stream itself. Neither invented new coding/decoding methods, but instead, leveraged other existing protocols (like G.711 and G.729).
Both typically use a server to act as a middleman for setting up calls. In H.323, a gatekeeper sends and receives keep alive and signaling packets to terminals that in turn set up their media streams to PSTN gateways. With SIP, a proxy server can process and forward requests from user agents to set up calls directly to other user agents, or through gateways for calls to traditional PSTN numbers. That being said, SIP can be implemented in a point-to-point fashion on a limited scale, as user agents can setup sessions directly without an intermediary server, as in the one-X Quick Edition platform.
While the protocols are conceptually similar, they're quite different in their architectures and in the services they provide. H.323, first expanded for VoIP in 1996 and now in its fifth version, was based on telephony protocols like ISDN Q.931. IETF visionaries first conceived of SIP in the mid-90's and have published two RFCs since then, the latest (RFC 3261) in 2002. SIP is based on text-based protocols like HTTP and SMTP, which are well understood by many programmers who find SIP fairly simple to code and troubleshoot. H.323 is written in binary code, making it less friendly to a programmer without significant experience and development tools.
A major difference between protocols is that SIP is simply used to set up and tear down media sessions, while H.323 specifies in detail which underlying protocols will be used to provide the media service. With SIP, the media itself is independent of the signaling protocol. In fact, SIP relies on another Session Description Protocol to define, negotiate, and handle the media streams. As a result, SIP can just as easily be used to set up a voice or video call as it can set up a gaming or instant messaging session. In other words, SIP is not a VoIP protocol. Since SIP is usable in so many areas, developers will continue to become more familiar and get more creative with SIP than they will with H.323.
SIP has allowed developers to create more new services than they could with H.323. SIP is a part of the IETF toolkit, meaning there are clearly defined requirements to interface and extend it. Extensions have already been added to use SIP beyond the setup and teardown of traditional media streams. New types of media sessions like some push-to-talk services offered by cellular carriers are based on SIP. Instant messaging services are made possible with a SIP extension called SIMPLE (SIP for Instant Messaging and Presence Leveraging Extensions). SIMPLE provides an architecture for the implementation of a traditional buddy list-based instant messaging and presence application with a standards-based core.
SIP's native support of presence will propagate into other devices, enabling exciting new features like single-number contact and visibility of presence across multiple devices (i.e. PC, desk phone, and cell phone). This will allow a caller to contact the called party using the best device at the right time, instead of only viewing the status of the IM client and guessing if the person is on their phone. This should reduce the frequently typed instant message, "can I call u?"
More interestingly, SIP and open Web services provide options to integrate presence and communications into traditional enterprise applications. For instance, events that occur within a business application (like an inventory shortage) can instigate a conference call to the relevant stakeholders (like the line manager and parts supplier) using presence status (on their PDAs, desktops or cell phones) to base the decision of how best to reach the stakeholder.
H.323 still predominant in the enterprise
Clearly, no such services have been created with H.323, and it is unlikely that they will be retrofitted. Yet the most scalable and feature-rich enterprise deployments today are using H.323 systems. While SIP matures, H.323 systems will still continue to be deployed, just as some TDM systems are still deployed today amidst the popularity of H.323. Most vendors continue to develop H.323 and SIP solutions in order to leverage each protocol most effectively and to provide options for enterprise communications.
It should be noted that H.323, while much more mature as a protocol, actually has less interoperable features across multi-vendor equipment. Even so, H.323 systems have a significant head start, with extended feature sets reaching into multiples of hundreds. Thus, in enterprises with call centers and highly customized dial plans, H.323's head start will not be relinquished soon. The vast majority of new VoIP deployments in enterprises today use H.323, but SIP is expected to grow rapidly in small office environments where telephony needs are simple and as service providers develop hosted telephony solutions using SIP.
SIP gaining momentum
While SIP is not as broadly deployed as H.323 in enterprise environments, it has rapidly gained momentum for service providers to carry enterprise and consumer VoIP traffic across their backbones. Service providers are now requesting SIP technologies in their RFPs for new equipment, and are migrating to softswitches that nimbly set up and tear down calls between proxies, SIP clients and PSTN gateways. SIP also allows service providers to offer outsourced services that were not possible with TDM or H.323 technologies. Components like voicemail or conferencing systems can be integrated into enterprise PBXs using SIP across network boundaries, providing service providers with new revenue and enterprises with options to outsource non-strategic components.
Although SIP is considered a mature protocol, it is not 100% complete as a standard. Standards are still being developed, especially in the privacy and security areas, and enhancements to the feature set will continue for the foreseeable future. SIMPLE still has more fundamental architecture decisions being ironed out in the IETF.
While H.323 will continue to play a major role in the future of VoIP, SIP deployments will begin to gain popularity. The underlying protocol itself has not been the driver for implementation decisions -- decisions are more effectively made based on the business and application requirements. Depending on the application, one or the other may make more sense, and in many environments, there will be a place for both!
Early adopters have seen value in SIP as an open-standard, device-independent and flexible protocol for the enterprise environment and many other customers are complementing their H.323 environments, using SIP where it makes sense.
About the author
Christian Stegh is the IP telephony practice leader for Avaya's North American region. His experience includes deploying and maintaining IP networks for an enterprise IT staff, designing converged networks, and integrating Avaya IP Telephony solutions into hundreds of client networks. His current responsibilities include providing Avaya's product developers with requirements, direction, and input from Avaya clients. He sits on Avaya's Security Steering Committee, multiple product decision teams, and speaks regularly at IEEE and industry events. He has a Bachelor's in Electrical Engineering and his MBA is from the University of Iowa.
For more information on SIP, visit the Avaya SIP portal.