Where is it written that a good concept somehow loses its innate qualities if the world doesn't rush to embrace it? And who says that standards handed down from the technology mountaintop are sacred writ and inviolate? Nowhere and nobody. By all indications, the adoption of an IP multimedia subsystem architecture continues to be a slow and evolving process. So what?

Some have said recently that telecom service providers are backing off plans to deploy an IMS infrastructure. But the folks consulted for this article generally agree with Sanjay Mewada, vice president of strategy for NetCracker Technology, who said “backing off” is too strong a term. “Granted, service providers aren't going to try to boil the ocean with IMS and move all their services onto it, but that's the wrong way to look at IMS,” he said.

The right way, he said, is to look at IMS as a framework or a set of platforms for delivering converged services. Those services will likely be added one at a time. No matter what angle you view the IMS market from — as the test and monitoring, infrastructure, software and middleware companies you'll hear from here see it — the right way to react to the snail's pace of adoption is with patience.

“It's going to be a slow process,” said Joe McGarvey, principal analyst for Current Analysis. He said vendors have progressed in positioning their IMS solutions since last year, when it was all about describing the destination. McGarvey likened the approach to John Steinbeck's novel “Of Mice and Men,” in which Lenny would constantly tell George about the rabbits and how good things were going to be.

“Well, service providers are sick of hearing about the rabbits,” McGarvey said. “In 2007, they want to hear about the journey, about how to get from where they are now to this nirvana that has been talked about for the last couple of years called IMS.”

There is no shortage of Lennys willing to provide directions, but McGarvey said service providers won't budge in a meaningful way until they are forced to do so by the likes of Google and Yahoo! He said traditional service providers are already looking past voice over IP (VoIP) as a threat because they can't compete with the low price point. They are looking to the next competitive step, and that is with converged voice and content services.

The problem with this, from an adoption point of view, is that service providers must get this transition right. The services they hope to deliver must be delivered faster and more accurately than previous services they have rolled out. And they have to be rolled out across two platforms: legacy time division multiplex (TDM) and next-generation session initiation protocol (SIP). Is there any question why they are taking their time?

Another perceived lull in adoption comes from what McGarvey calls the big new wrinkle for IMS: the service delivery platform (SDP). Always considered outside of, or a complement to, IMS, the SDPs coming to market today are being fitted with IMS-like attributes, such as enabling new network-agnostic services.

In addition to the three network layers typically associated with IMS — transport, control and application — there appears to be a call for a fourth layer between application and control that can act as a service layer, McGarvey said. This is being answered with the SDP, and a number of vendors are heeding the call, including BEA, HP, IBM, NetCracker, Oracle and Telcordia.

With a mission to extend its Fusion Middleware and the service-oriented architecture from the enterprise world to telecom, Oracle recently launched the latest solution on its SDP road map: the Oracle Communication and Mobility Server (OCMS), which includes a third-generation, programmable SIP and presence server, proxy and registrar services, and an SIP-based VoIP softphone client.

And if the adoption of SDPs is perceived to slow the transition to IMS, so be it, said Indu Kodukula, vice president of product management for Oracle.

“That is one possible outcome, but the real value of the SDP is that it more effectively monetizes IMS deployments,” Kodukula said. “Perhaps there is some initial latency or delay in rolling out a full-scale IMS network, but when that rollout happens, it will be with a much better understanding of what services are likely to be successful and revenue-generating when the IMS network is fully available.”

Many service providers are looking at SDPs as a way to have the best of both worlds by getting IMS-like service availability without the full IMS investment. The SDPs extend the life of the legacy network by enabling new services across both networks. Still, it is not the aim of SDP vendors such as Oracle to delay IMS deployment. The design of its J2EE-based SIP application server — and the 3 million Java developers free to write applications for it — is meant to “bring Internet speed to telecommunications services,” Kodukula said.

It features an IMS presence server that is SIP/Simple and Open Mobile Alliance-compliant and supported by an embedded Oracle TimesTen In-Memory Database. It includes a SIP-based softphone client and support for authentication, authorization and accounting.

Kodukula sees IMS as just one of the network types supported by an SDP, rather than the SDP simply being a compliment to IMS. “In our view, it is a horizontal, standards-based platform that allows the delivery and execution of a number of different and interesting services that work across multiple networks,” he said.

It is good Oracle has its own definition of an SDP because not everyone does, or more accurately, not everyone agrees on what the definition is. The TeleManagement Forum created a working group last year to produce a common, functional SDP definition. Oracle is involved in the effort.

Kodukula said traditional TMF models around operations support system and business support system integration have been static and focused on building resource models that map services to resources in a predefined manner. “The SDP fundamentally changes that equation,” he said.

With basic issues such as an SDP definition still being hammered out, it is not hard to see why IMS adoption is proceeding slowly. And the SDP is only part of the issue.

Test and measurement companies have a bird's eye view of the progress of any new technology because of their involvement in vendor development. Empirix is no exception. Through its participation in various plugfests and other industry events such as GMI2006, the company has a good sense of where IMS readiness stands.

“I do think that people believed things would have moved much more aggressively and be at the point of testing millions of subscribers and feature quality, but we're not there yet,” said Duane Sword, vice president of product management and marketing for Empirix.

Vendors are still testing basic call management and interoperability between their own network elements, and they are still conducting what Sword calls positive testing, where Empirix is asked to emulate proper sessions with clean registrations and voice quality. They hold the hard testing for their own labs, where they ask Empirix and other test companies to introduce malicious attacks and other bad data into the mix.

Sword doesn't expect this to progress far in 2007, as people come face to face with the realities of delivering stable, real-time session management. One of the reasons for this is the lack of a conformance standard against which vendors can test — a result of the downside to the open standard to which IMS is being built.

“People need a conformance standard where there is an abstract test suite that gets converted into an executable test schema that goes from feature testing to load testing and from positive through negative testing, and that people agree on,” Sword said.

In the absence of a standard and regulatory mandates, companies are moving ahead and choosing partners based on convenience and political expediency. “There is a pretty good reason why Verizon created a consortium privately to go after IMS to fix the things that the standards don't address or describe clearly enough,” Sword said.

Despite the issues, Sword isn't complaining. IMS has definitely contributed to more test gear purchases, and that's good, he said. So far, North American equipment makers are investing more in test tools and in their labs. In light of the latest CableLabs 2.0 specification clarifying questions about the infrastructure, cable companies are investing more than expected.

There are other signs that the pace may soon pick up. Oracle points to its growing partner ecosystem. When the company announced its J2EE-based OCMS, it announced 30 new development partners ranging from HP and Sun Microsystems to developers of wireless RSS feeds and gaming applications.

“In general, we are expecting generally available products from them in the next six to nine months,” Oracle's Kodukula said.

And other vendors are still entering the field with new products. Stratus Technologies, for one, has turned the Emerging Network Solutions enterprise products it acquired six months ago into carrier-grade products. Stratus combined Emerging's Entice infrastructure environment with its Converged Service Broker (CSB) to fulfill the need of the IMS service capability interaction manager and service switching function.

Nathan Franzmeier, vice president of Emerging Network Solutions, said the CSB allows services on different platforms and networks to be applied to the same call. “You can inject a service anywhere our product is installed,” he said.

Franzmeier said Stratus plans to participate in upcoming tests and interoperability events.

Given the testing that remains, the standards that must be agreed upon and the SDPs that allow service providers to gain access to IMS-like service on their TDM networks, 2007 might not be the year for broad IMS adoption.

However, “It will be the year all the hard work gets done,” McGarvey said.