In these days of rabid polarization, where even the facts seem to have two sides, it's comforting to know that there are still some things that nearly everyone can agree on. When it comes to choosing an architectural framework for building the next service delivery infrastructure, for example, the IP multimedia subsystem, or IMS, at least to this point, is running unopposed.

While the polls are still open and an alternative candidate could still emerge, IMS is the clear front runner, brandishing what is essentially universal support from an alphabet soup of standards bodies and industry associations. The 3rd Generation Partnership Program (3GPP) — where IMS originated — the Internet Engineering Task Force (IETF), CableLabs, ETSI, the ITU and others are coalescing around the IMS framework.

That the IMS framework is being embraced by nearly all quarters should not come as a shock to even casual followers of the telecommunications industry. Service delivery models of the past decade or so have basically relied on the construction of an overlay network, fortified with dedicated resources, for each service a carrier offers. This one-to-one relationship between services and dedicated networks has long been the poison service providers have been forced to swallow. IMS is the antidote. It simultaneously unites fixed and mobile operations, enables rapid delivery of new services and reduces capital and operational expenses. Though RBOCs, PTTs and other Tier 1 carriers and service providers have rarely been accused of decisiveness or sagacity, even telecommunications behemoths, when confronted with what amounts to network Nirvana, are smart enough to embrace it.

And so are equipment makers. Nearly every major, minor and niche player in the telecommunications equipment market has at this point unveiled its IMS strategy and portions of its road map. The majority of large incumbent equipment makers, such as Alcatel, Ericsson, Lucent and Siemens, is shipping IMS-compliant functional modules (though release 6 of the 3GPP specification is not yet finished) — and a few have even announced deployments. Despite the progress that seems to be taking place in the advancement toward an IMS-based service delivery model, the truth is that the party hasn't even started. On a maturity scale from 1 to 10, IMS would be considered a fraction. If you're partial to cradle-to-grave time-lines, with “Infant” representing the far left of the lifespan and “Geezer” the far right, then IMS would be plotted somewhere around the second trimester.

Get it? It's still really early. And given the precocious state of the market, getting the lay of the competitive landscape or forecasting the course that IMS adoption will eventually take is a tricky proposition.

QUANTITY, NOT QUALITY

At this stage of the game, a measured evaluation of the competing equipment vendors is essentially impossible. Though nearly all competitors, a list that includes the world's largest telecommunications equipment makers, such as Alcatel, Cisco, Ericsson, Lucent, Nortel, Nokia, Motorola, Siemens and others have announced or delivered multiple IMS-based functional modules, little if any information regarding performance is currently available. This shortage of raw performance data, such as session set ups per second or simultaneous sessions, makes it impossible to provide a meaningful comparison between, say, a Call Session Control Function (CSCF) from Ericsson and a CSCF from Alcatel.

In the absence of quantifiable performance data, the only reasonable way to compare relative IMS products at this point is to grade each competitor on the breadth and availability of their portfolio and their early traction with service providers and carriers. While this evaluation methodology is obviously flawed and offers little value in terms of predicting long-term success, it's technically accurate to characterize equipment makers that offer comprehensive wireless product portfolios as being off to a strong start. Ericsson and Siemens, for example, have announced several deployments — both with wireline and wireless carriers — of their respective CSCFs, which is the functional module at the heart of the IMS architecture.

It's not a coincidence that mobile-oriented equipment makers are among the first to fill out their IMS-based portfolios. IMS has it's origin in the wireless world, giving Alcatel, Ericsson, Motorola, Nokia and Siemens a head start over the more wireline-oriented competitors on their IMS road maps. Consequently, wireline-oriented players, such as Cisco, Nortel, Sonus and others, are still finishing up work on major portions of their IMS-based product portfolios. Lucent stands out as a bit of an exception. It was among the first of the equipment makers to embrace the concept of uniting wireline and wireless call control mechanism under the IMS architecture and introduced a comprehensive IMS-based portfolio late last year.

One thing that all competitors in this burgeoning market have in common is that they are still trying to figure out the best way to bundle the various functional modules and the most effective way to differentiate their products from competitive products.

IMS PACKAGING

As several vendors, including Nortel and Alcatel, have announced plans to eventually move their IMS portfolios to ATCA platforms, it appears that most equipment makers envision IMS functionally being packaged in some sort of server blade technology. The major benefit of a server blade approach is that different functional modules could be mixed and matched as needed, creating a scenario in which a single chassis — managed as an integrated system — could house multiple IMS modules, such as the CSCF, MGCF, MR< PDA and so on. Lucent is one of the first equipment makers to provide details about its IMS packaging strategy. The Lucent Session Manager, which is now in trials, is a chassis-based platform that can be configured to offer multiple IMS functional modules, including the CSCF, SCIM, BCFE, etc.

While locating multiple IMS-based modules in a single chassis is appropriate for small installations, server blade technology also enables service providers to neatly scale their IMS infrastructures to accommodate millions, if not potentially billions of simultaneous sessions. It may turn out that some network operators will need to dedicate entire racks of blade servers to a specific function, such as the CSCF or the MGCF. The point is that putting functions on blades will enable carriers to mix and match a variety of independent functional modules but still managed their infrastructures as an integrated system.

With all this talk of blade servers, it doesn't take long to recognize a strong correlation between this type of telephony architecture and traditional Internet server architecture. The similarity is not a coincidence. As real time communications move to an all-IP environment and session control and application servers become software processes stored on industry standard hardware, carriers will naturally leverage the existing and familiar Internet architecture to delivery real-time communications services. It is only when millions of communications sessions can be pulled off of Internet-like servers, instead of being associated with hardware ports or nailed-up connections, that carriers and service provider will be able to enjoy dramatic drops in capital and operational expenses.

VIVA LA DIFFERENCE

One of the major unfinished pieces of business in the IMS framework is an agreement on the way quality of service (QOS) and policy enforcement mechanisms will be implemented in real world deployments. At this point, vendors disagree on whether this technology can be implemented in existing infrastructure gear or if it requires a dedicated piece of equipment, such as a bandwidth manager.

It shouldn't come as a surprise, then, that several equipment makers see QOS and policy enforcement as areas in which they can add value beyond the basic IMS specification. Cisco, for example, is looking to leverage the deep packet inspection capabilities of the session control technology it acquired with the purchase of P-Cube to provide its customers with the tools to provide dynamic QOS. Only by looking closely at each signaling and media stream that passes through the network can carriers control the delivery of services that originate on their network and, more importantly, those that do not.

Near the top of carriers' wish lists is for IMS to provide them with the knobs and the dials that will allow them to prevent their last mile networks from becoming dumb conduits for bandwidth-hogging applications sold by third-party content providers. Mechanisms built into the IMS specification, and potentially enhanced by proprietary bandwidth management tools, are designed to enable carriers to have a say in how content is delivered across their networks and to charge ASPs for ensuring that the services they provide to subscriber are of an acceptable quality. Carriers understand that they must learn to live with the Skypes, Vonages and AT&T CallAdvantages of the world. Still, they want to change the relationship they have with these ASPs from host-parasite to one that is mutually beneficial. An IMS-based infrastructure — if adoption goes according to plan — will provide carriers with the capability to serve as a link in the value chain. Owners of the last mile pipe will provide their ASP “partners” with guarantees that their content will have enough bandwidth and the proper priority. In return, facilities-based operators will grab a piece of the action.

PARTIAL PARADISE

Just as all overly hyped products or technologies never seem to live up to market expectations, IMS, when adopted, will fall somewhere short of delivering the utopian-like infrastructure that service providers long to adopt. In the end, Nirvana always turns into “Nearvana” and Shangri-La ends up being more “Shangri-Like.” The difference this time around is that even a muted IMS will bring about substantial improvements to current service delivery models. IMS may not be the Promised Land. But it is at least a big step in the right direction.

Joe McGarvey is Current Analysis' senior analyst for carrier IP telephony infrastructure.

IMS Evolutionary Path