Well, we're about a month away from the Enterprise Connect conference in Orlando. It's 82 degrees in Southern California (where I am as I write this) as I watch the news showing the blizzard conditions in the Northeast. Although not a perfect analogy, the wide differences in telecom carrier offerings and support are somewhat similar to the variations in the weather.
SIP Trunking has proven to be a popular topic in my consulting practice, at last year's conference, and as well at the four-city Enterprise Connect Road Show and Virtual Event on SIP Trunking. The tour attracted many enterprise users who were considering this "new" technology but hadn't yet deployed it. This article is a preview of what Jim Allen and I will be presenting at this year's conference on Tuesday, March 18. Our SIP Trunking Tutorial will cover:
* Current State
* SIP Trunking Architecture options
* Session Border Controllers and their role
* SIP Trunking Implementation pearls and nuggets
* SIP Trunking Support
* Fortune 500 company case study (insider's view)
The Tutorial will also address each of these elements:
Truth or Myth
SIP is new: Myth (mostly)--SIP has existed in some form for almost two decades (earliest RFC I could find is 1996). Most in the industry are familiar with implementations for endpoints (i.e. SIP phones) or as the protocol choice of integrating UC components. RFC 4904 (June 2007) largely defines SIP Trunking, so it's been around at least for six years.
SIP Trunking is standardized: Partial Myth--Although there are several RFCs supporting SIP, as well the SIPconnect 1.1 standard, significant differences exist in implementations from various carriers and equipment manufacturers (i.e. IP-PBX, gateways, SBCs). In addition to the technical differences, carriers have bundled features and options differently so as to make comparisons among their SIP trunks more challenging (when compared with PRIs).
SIP Trunking is "mature" Partly True--It's kind of like a teenager--behaving the way you expect it to sometimes and surprising you often. This is particularly true in the provisioning and implementation phases.
When compared to the system supporting PRIs, the tribal knowledge of SIP Trunking by personnel, and actual systems at the carrier's, isn't quite fully baked. Recently-published surveys show continued enterprise customer dissatisfaction with carrier support. Although this is improving, we'll describe some "war stories" of actual implementation issues.
SIP Trunking offers cost savings: True--Though as the saying goes, "your mileage may vary". Savings in the range of 30% to 50% are possible for large distributed environments that adopt a centralized SIP Trunking model. Smaller savings in the 10% to 15% range are possible for smaller environments doing a like-for-like SIP/PRI replacement. Some carriers appear to be aggressively pricing PRIs in an attempt to retain installed base.
I don't need a Session Border Controller (SBC): False--Almost every IP-PBX provider is able to support SIP trunking natively via supported gateways and Session Management. In addition, carriers provide SBCs on their side of the SIP trunk. Nevertheless, enterprises are encouraged to procure their own highly available SBC for the following reasons:
* Security--Enterprise SBC acts as an Application Aware Firewall for SIP communications
* Protocol Interoperability--Between different implementations of SIP (and possibly other protocols such as H.323)
* Handling multiple carriers
• Dial Plan Support
SIP Trunking is widely available: True--SIP trunking is available in most metropolitan areas directly from Tier 1 and Tier 2 carriers. In addition, SIP Trunking is available almost everywhere from "over the top" implementations.
There is one model to implement SIP Trunking: False--A couple of years ago, this might have been considered a true statement, as "centralized" was touted as the way to go. This still is the most popular approach, offering the highest cost savings as well as dial plan centralization, but two other models have come into the picture: distributed and hybrid.
The distributed model allows for quicker implementations by keeping the dial plan somewhat intact. It also offers a better use of network resources if video will be used heavily by the enterprise (i.e. prevents hairpinning and trunk saturation). Hybrid of course combines properties of both centralized and distributed models, customized for the specific needs of an organization.
Below are diagrams highlighting the centralized and distributed approaches:
CENTRALIZED:
DISTRIBUTED:
Summary
SIP Trunking is maturing as an industry and continues to improve year over year. The lack of knowledgeable carrier support staff and processes to go along with standard implementations has slowed adoption, though I fully expect the pace to accelerate given the cost savings and support of technologies such as UC and video. The use of SBCs allows the enterprise to mitigate the current shortcomings. Piloting and rigorous testing is still required, as is the case with most new technologies.
The Society of Telecommunications Consultants is an international organization of independent information and communication technology (ICT) professionals serving clients in all business sectors and government worldwide.
Check out Dave Stein and Jim Allen's SIP Trunking tutorial at Enterprise Connect Orlando!
