The IETF’s Session Initiation Protocol (SIP) started out innocuously enough, as an end-to-end application layer signaling protocol used to control real-time multimedia sessions (packet exchanges of voice, video, etc.) between participants across an IP network, operating independently of any of the underlying transport protocols (typically RTP, UDP and IP) and with no dependency on the type of session. SIP is now the most popular call control standard for VoIP in both wireline and wireless 3G networks, as defined by the Third Generation Partnership Projects (3GPP). SIP is the principal protocol in IMS (IP-based Multimedia Subsystem), the up coming common service architecture for fixed-line and wireless networks.
Eric Winsborrow (News
), Chief Marketing Officer at Sipera, says, “With our company’s name, Sipera
, we obviously believe this is the era of SIP. At a high level, we’re seeing trends in enterprises and service providers in terms of trying to open up the network and extend capabilities beyond just plain old tip-and-ring and replacing Ma Bell-like traditional phone companies. We’re really opening up this era of unified communications, be it in the enterprise or as advanced services and architectures in the service provider space. To do that, you need SIP. I worked at various companies on UC and VoIP technology in the 1990s and loved SIP when it first appeared, but for some reason we went with H.323 at Avaya and Cisco went with something else, and none of the non-SIP protocols gave us the flexibility that we wanted, quite frankly. So even back in the 1990s, we realized that SIP would eventually be used for next-gen services. Clearly, SIP is where it’s at for what enterprises and service providers want to do in terms of services or just future communications capabilities.”
“Experts and analysts track line usage, and SIP, in 2008, will easily surpass even H.323 in number of lines deployed,” says Winsborrow. “By next year, there should be close to 70 million SIP lines actually deployed, and it won’t be long – probably a year or two after that – when SIP overcomes all of the different protocols, at least enterprise-wide, for VoIP or these new UC services. So, it took about 10 years, but SIP is now really becoming the dominant protocol. Even though SCCP has a large installed base, Cisco is moving toward SIP. Nortel is doing the same. Their CS2100, a high-end service provider box used in large Fortune 500 companies, is now SIP-enabled. Even their mainstream mid-level box, the CS1000, will be moved to SIP by the end of the year. Avaya is doing much the same thing. All major IP PBX vendors are standardizing on SIP. And Microsoft OCS, which, by 2010, will be really disruptive in the VoIP space, is SIP-enabled.”
Developing SIP-enabled products can be a challenge for some, though the protocol is actually less complex overall than its predecessor, H.323. For those willing to take the plunge, there are many hardware and software packages available.
Product Manager Ian Colville, says, “We have a SIP stack that works with our products such as our Host Media Processing-based product, which is Prosody S, or Prosody X, which is our media processing board. It’s much like the way our SS7 stack or TDM-based protocols work with our board products.”
Aculab’s Prosody X also works with an extended SIP API called the “SIP Bridge” that coexists with the generic call control API. SIP Bridge gives you the same kind of third party call control that you experience with a TDM PBX having a sophisticated computer telephony integration. SIP Bridge can achieve this without a CTI interface, or PBX for that matter. One result of this is that the Aculab SIP Bridge has become an inexpensive way to build complex IP contact center and IP PBX type products, with rich media and call control features, using APIs common across a range of IP and TDM protocols and formats. Moreover, SIP Bridge’s IP calls enable many call control scenarios not possible with TDM trunking and protocols.
“We’re not currently selling or licensing a SIP stack as a separate product,” says Colville. “Nor do we do that with the SS7 stack. But our stacks stand in comparison to anything that you could get from a stack vendor with regards to what you would want to do with, say, a media processing board.”
“We’ve been to the SIPit events, and we’re members of the SIP Forum (News
) and SIP Center, and we’ve done work regarding interoperability, and our SIP stack performs very well in those tests,” says Colville. “Our stack has been used by our customers in environments incorporating the Cisco Call Manager, for example, Avaya’s Communications Manager, Siemens (News
) HiPath 4800 Series, and various gateways including those of Cisco and Audiocodes. We also work with a range of SIP servers and SIP phones, and so forth.”
Developers Shouldn’t Re-Invent the Wheel
As service providers enter the new era of IP communications, Tekelec provides solutions that enable operators to offer new, revenue-generating multimedia services while leveraging their existing infrastructure and taking a stepped approach to all-IP networks.
Ajay Deo, AVP of Engineering at Tekelec
, says, “The approach we’ve taken is that we’re leveraging as much open source technology as we can. In terms of core IMS development we are using an open source technology called the OpenSER SIP Server, and then we have developed the IMS extensions on top of that. In terms of applications, such as presence, instant messaging, or any other application-level product we have, we also use another open source project called reSIProcate which has well-defined APIs and performance characteristics, and further it supports the latest features worked on by the IETF [The reSIProcate project develops an object-oriented SIP stack written in C++ and intended to serve as the SIP reference implementation. Its design objective is to create a well-documented and easy-to-use SIP stack for use in phones, gateways, SIP proxies, back-to-back user agents as well as instant messaging and presence applications.] We use that for most of our app server and app development.”
Tekelec’s Principal Engineer, Adam Roche, says, “Because SIP is a feature-rich protocol, and because the protocol itself is somewhat complex, it’s kind of daunting when you first approach it in terms of programming, because it looks like there’s so much to do. From that perspective, the best thing to do is to start with a developed stack, as we’ve done. If you wrote a program that uses TCP, you wouldn’t write your own TCP stack too, would you? You’d use an already-written stack. There’s not a whole lot of sense in having a many SIP different stacks out there. If you just have a small handful of them that do what they do extremely well, then the application developers and go and use those without having to worry about the underlying protocols too much. One nice thing about the reSIProcate project is that it was started and is maintained by most of the core contributors to the SIP protocol standardization efforts in the IETF. In fact, it is where we tend to try out the new ideas that we are working on in the IETF to make sure that they’re feasible and that they work the way that they’re supposed to. So, from that perspective, it is pretty cutting edge in terms of the features that is supports.”
“One you’ve got a SIP stack and software to work with, there are a number of protocol specific features and questions that might arise,” says Roche. “There’s actually a very good resource by Henning Schulzrinne, called the SIP Implementer’s Mailing List. You can send off questions to it and you usually get pretty coherent answers. There are a number of people on that mailing list who are also involved in SIP standardization, so not only do the answers arrive quickly, but they’re pretty accurate, too.”
Instead of having to deal with gateways from your IP phone system and network to the old circuit-switched PSTN, SIP trunking moves these gateways from your corporate premises to your Internet Telephony Service Provider (ITSP). You can now use your IP PBX to communicate over IP within the enterprise, but also outside the enterprise to people on the PSTN via a SIP trunk over the Internet as provided by your ITSP, thus eliminating not just local PSTN gateways, but also a bundle of physical phone lines (Basic and Primary Rate Interfaces) coming into your premises.
To successfully deploy SIP trunks you need a PBX with a SIP-enabled trunk side, a SIP-capable enterprise edge device understanding SIP and an Internet telephony or SIP trunking service provider.
is well-known for its SIP trunking solutions for voice communications services. Their solutions are designed to handle high volume telecom users worldwide. These include enterprises of all size, call centers, and Interactive Voice Response (IVR) environments. BandTel ensures that its customers have continuous service via its fully redundant “N-Plus Matrix” VoIP network. BandTel deploys its SIP softswitch technology as N-active switching nodes, so any end-user can receive service from any of the N BandTel switching nodes in the BandTel switching matrix, at any given time. If a node in the “N-Plus Matrix” fails, the other surviving nodes in the matrix take over and service continues immediately.
BandTel’s Vice President of Marketing, Joel Maloff, says, “We’re a pure SIP trunking provider. We don’t provide hosted IP PBXs or conferencing services. We strictly provide SIP trunking and we do that through distribution channels, such as VARs, distributors, resellers, and so forth. The company has created its own patent-pending architecture built both on SIP standards but also incorporating some ‘philosophical components’ from the telephony SS7 environment. The idea is that when BandTel is providing SIP trunking, because of the ‘N-Plus’ distributed architecture, the nodes can be located anywhere and failover can occur amongst them.”
“We conducted a perception poll and discovered that the SIP trunking concept was not very well understood in the marketplace by the analyst community, media, or otherwise,” says Maloff. “When asked to name SIP trunking providers, people would name Ingate, Avaya or other hardware companies, which is amusing. There are actually a variety of SIP trunking companies out there that also offer a variety of other services such as hosted IP PBX, conferencing, and so forth. Doing that would be a problem for us, because our distribution channels are companies such as ShoreTel and Linksys, who offer their own services. If we also offered hosted IP PBX service, we’d be competing with our own distribution channel. We don’t do that, and they seem to like the fact that we are a pure provider of SIP trunking.”
“There are companies that are fairly well thought of in the SIP trunking space, such as Cbeyond (News
, or Bandwidth.com
,” says Maloff. “They require that their customers have a dedicated line in order to access the SIP trunking service. To us, that’s defeating the purpose of SIP trunking, which allows you to amalgamate your data and voice requirements over a single circuit. The need for another circuit from another provider is a waste and limits the regions of coverage, since they can only provide service where they have the ability to set up those leased lines. BandTel, however, is limited only by your broadband connection and, in fact, we do provide DIDs throughout the U.S. and Canada and many international locations, so we’re really serving a wide range of customer needs and services. We also have an interoperability program where we test hardware from companies with which we intend to deliver service. We assure that the SIP standards are being adhered to and they function properly, knowing how vague they can be at times. And it all seems to be working very well.”
Another SIP trunking provider, Cbeyond
, is a voice and broadband Internet provider that serves small businesses exclusively with customized packages and extensive customer support. Their local, long distance and Internet packages, anytime account management and advanced VoIP platform inexpensively deliver to small businesses the same kind of communication experience that big businesses enjoy.
Cbeyond’s CTO, Chris Gatch (News
), says, “You know how the simplicity of Ethernet has continued to gain share within the overall networking space. Ethernet started out as a LAN technology and then it just creeped all the way through the enterprise, and then into the Wide Area Network (WAN), and the Metro Area Network (MAN) and access networks of all kinds. SIP’s progress has been very similar. It continues to take over more aspects of the service provider network. When we started working with SIP seven or so years ago, it was mainly in the core of our network. We used SIP to connect our softswitches together, and to connect them to our voicemail platform. It was a protocol that we used for session initiation between different infrastructure components in our network. Since that time, SIP has really moved to different aspects of the network. We’re big and instrumental in the whole SIP trunking field, and now we have many customers connecting directly to our network via SIP, so we are using it in the edge and we are still using it to connect together core infrastructure components such as application servers with each other and trunking gateway controllers to app servers.”
“We use SIP in limited peering scenarios,” says Gatch. “SIP is being deployed in peering between carrier. We have areas in our network where we’re using competing protocols to SIP, such as MGCP, but the reality is that SIP is probably the preferred protocol in almost every aspect of the service provider network, perhaps with the exception of the trunking gateway controllers, the signaling gateways where you’re connecting out to the PSTN.”
“What’s perhaps most meaningful to our business is what’s happening with SIP trunking between the enterprise and us, the service provider,” says Gatch. “It’s very significant connecting a custome natively with a VoIP service provider via IP, because that really does to lay the foundation we’ve all talked about for a long time, with end-to-end IP feature transparency. When we’re connecting customers natively via IP, that becomes real, and we’re seeing customers use things such as videophones and high-definition and wideband audio codecs on Polycom phones. On an overall basis, the number of IP-connected customers are small, but those who are using IP and SIP trunking are experiencing the benefits.”
Start SIPing Today
SIP is unavoidable and inevitable in today’s telecom world. It’s even plausible that certain home appliances may one day understand SIP, though I wouldn’t go shopping quite yet for a SIP-enabled dishwasher or kitchen sink.
Richard Grigonis is Executive Editor of TMC’s IP Communications Group. To read more of Richard’s articles, please visit his columnist page.
Edited by Erik Linask