When I first became aware of "SIP Trunking", probably over 8 years ago while I was a Program Manager in the Microsoft UC group, I was distinctly underwhelmed by the whole idea. Although I completely understood the need for, and value of, connecting a UC system with the PSTN and mobile networks, there existed a myriad of IP-PSTN gateways (e.g. Dialogic, AudioCodes, and about 25 others at the time) that were capable of connecting a UC system to a PRI trunk. Therefore, the value of connecting to a Telco via SIP/RTP, including dealing with all the commensurate provisioning, security and authentication issues at the enterprise edge, only to have to hit another IP-PSTN gateway at the Telco network edge, was something I just didn't understand. I was much more interested in the concept of end-to-end, multi-modal communications that we now call "federation". Naturally, I was punished for giving forth on this point of view by being assigned to create the Microsoft SIP Trunking partner program.
Part of my program strategy was to minimize the "surface area" of the SIP Trunking interfaces that would have to be addressed by tackling the interoperability standards problem: so I got Microsoft involved in the SIP Forum's SIPconnect v1.1 initiative. As it turned out, the big Telcos (ILECs & PTTs) were not interested in participating in that exercise, partly because they saw SIP Trunking as cannibalizing their PRI revenues (i.e., it breathed new life into the CLEC business) and partly because they disagreed with some of the requirements that, in their view, added cost to their end of the implementation (see here for more on this topic).
Despite these objections from the big Telcos, their initial lack of uptake on SIP Trunking was baffling to me. Since, as I have suggested above, "dialing digits over SIP" was so underwhelming, I was sure that SIP Trunking would rapidly transition to a multimedia service, and a new enterprise federation network would emerge. As we all now know, the Telcos, rather than embracing this opportunity, chose to ignore UC and focus on consumer mobile technology, international expansion and, in the US, the re-assembling of "Ma Bell" via massive M&A activity. I have already written at length about why I think that this was a poor choice on their part, not least of which was that it gave the UC vendors the opportunity to launch federation as an alternate multi-modal communications network that completely bypasses the core value proposition of the Telcos.
Now that I have had chance to think about it more, it turns out that a "scope reduction" decision in the SIPconnect v1.1 process may have had a larger impact than anticipated. SIPconnect v1.1 was aimed at the enterprise-to-Telco interface, and not inter-carrier connectivity. SIP Trunking was primarily intended for PSTN termination and origination, so it was assumed that the PSTN would be the inter-carrier network. A number of Telcos today, including Verizon with their VIPER service, offer end-to-end SIP calling among their own subscribers; however, I remain unaware of any inter-carrier offerings for SIP Trunking. As a consequence of these assumptions and limitations, the opportunity for end-to-end UC communications was ignored, to the detriment of the overall value proposition for SIP Trunking.
While I was doing research for two of my sessions (SIP Trunking and Federation) at Enterprise Connect 2013, I discovered that two companies have recently taken the initiative to try to address this gap. BroadSoft, a vendor of next-generation communications switching elements, and Intelepeer, a cloud-communications service provider, have collaborated to deploy routing technology and a directory service that will allow SIP Trunk service providers to offer their customers inter-enterprise end-to-end RTP audio and video. In other words, SIP Trunking is now morphing into federation.
The graphic above is fairly self-explanatory, but are a few things that I should point out that aren't included:
1. Broadsoft's BroadWorks switches are deployed in over 500 service providers in 61 countries around the world; including 20 of the global top 25 Telcos.
2. The SuperRegistry is essentially a private ENUM directory which currently contains 450 million phone numbers that can be routed by Intelepeer either directly or via a redirect to the originating service provider.
3. Although this service is apparently free to the originating service provider, it is the service provider's choice as to whether to configure this routing option in its own BroadWorks elements.
The latter issue is critical: if Telcos choose not to configure this option, for whatever reason, then clearly it won't be available. Furthermore, the theory of "network effect" tells us that for each Telco that chooses not to participate, the overall value proposition becomes exponentially less attractive for those who choose to participate.
Next Page: Enabling wideband audio
The service is currently focused on voice, with video capabilities emerging over time. Furthermore, the addressing mechanism is "e.164" (i.e. phone numbers); therefore, it is not strictly federation. However, it facilitates end-to-end SIP and RTP sessions, and therefore enables wideband or "HD" communications; which is, in my opinion, the compelling part of the federation concept. Over time, modalities such as instant messaging and presence could be added.
The other exciting part of this service is that it is not restricted to the UC platforms that currently support multimedia federation (i.e. Cisco and Microsoft); any enterprise that has deployed a SIP Trunk can potentially benefit from this service.
It is my understanding that the Intelepeer service is primarily focused on North America and Europe, which in any case represents 90% of global UC deployment, according to my UC Deployment Market Data. Also, the BroadSoft announcement implies that Intelepeer is potentially the first of multiple partners that could provide such central directory and routing services.
It is worth noting that IP communications are currently unregulated in the US and that this general function in the PSTN world is provided by the Local Number Portability system that is a regulated service, currently operated by Neustar. However, LNP is implemented on a national basis and this is arguably a service that should be implemented with international scope.
It is also worthy of note that this general "SIP Trunking/Federation" concept had already been implemented prior to the BroadSoft/Intelepeer collaboration, by a UK-based company called XConnect. I don't have much information about their service or its uptake by customers; however it does appear to require an interoperability verification, which takes us back to the issue of who specifies the interface and what impact that has on the implementations of others. The advantage of the BroadSoft/Intelepeer alliance is that the interoperability is already built in by the two vendors; providing near global reach.
Overall, I see this as a very positive development for enterprise customers. The Telcos may see myriad issues with this scheme and therefore may choose not to "opt-in". However, these are the same Telcos who have long been lobbying to be relieved of their government-mandated PSTN support obligations. I am not aware of any significant progress on the definition of an all-IP public communications network, so this service seems to neatly get around that issue and could potentially become a de facto standard. It will be fascinating to see how this service develops over time and what the uptake will be from both enterprise customers and the Telcos.