KPN brakes on IMS rollout

KPN puts the brakes on IMS rollout

KPN is drastically cutting back on development work aimed at moving to a single IMS core, Karl Heinz van der Made, KPN's director of service operations, told delegates at Informa Telecoms & Media's IMS Strategies conference in Dusseldorf, Germany.

"All our plans for IP-based services are currently being revised, as are our plans to move to a single IMS core," he said.

He added that the company might never move to a single IMS core if work on the development of services over KPN's all-IP core network were successful. "Our all-IP network is finished, and it's working well, but the business case for IMS is weak," he said. "We'll be exploring how far we can take [IP-based services] first, and in the meantime we are slowing development on IMS right down."

When Lucent was signed up to transform KPN's PSTN to an all-IP network and implement a single IMS core just 13 months ago, optimism was rife and both the operator and the vendor genuinely seemed to believe that both projects were possible.

"Lucent's IMS platform provides KPN with a future-proof, scalable and flexible replacement of the legacy voice network to further reduce our cost structure," Paul Hendriks, program director of all-IP at KPN, said in a press statement at the time. "At the same time, building an IMS core provides us the opportunity to develop the new services KPN wants to offer to business and consumer customers over KPN's all-IP network."

Just over a year later, things are very different, and although the IP backbone is in place and the first objective has largely been attained, the fact that KPN has decided to relegate the development of an IMS core toward the bottom of its list of priorities must come as a blow to both parties.

Van der Made's disappointment was evident from the downbeat tone of his presentation at IMS Strategies, and Alcatel-Lucent must also see the announcement as a major setback, even though it was attempting to implement a largely untried technology. When contacted by Informa, Alcatel-Lucent said it had no comment.

Van der Made didn't entirely rule out some sort of IMS-based services in the future, however, saying that doing things more slowly might simply mean doing it when conditions were more favourable. "We intend to work on the simple things first and do them very well," he said. "Our main priorities are IPTV and VoIP for now, because that's where we face the strongest competition. We will take on the Googles of this world in the future, but right now it's the cablecos we have to worry about."

Competition wasn't the only reason Van der Made gave for the company's strategic about-face, and he outlined what he believed were the barriers to IMS adoption. He told delegates that when KPN started its IMS initiatives, the company had great expectations. It hoped to find it easier to develop services, integrate fixed and mobile networks, blend services and implement seamless and real-time QoS management. It was also hoping to increase cost efficiency and make "zillions of euros" with IMS. But time has proved these expectations to be far too optimistic.

He said that the cost of moving toward a single IMS core was huge and hard to justify, and because IMS deployments require large volumes of traffic to make them profitable, only voice seems viable for the foreseeable future.

Malik Kamal-Saadi, a principal analyst at Informa, is broadly in agreement on this point. "IMS does not offer any cost savings when delivering a single service, such as VoIP or IPTV," he says. "A proprietary implementation of the same service is always more efficient.

"However, where IMS benefits costs is with each added service, and the cost efficiency increases with the number of value-added services the operator is willing to offer in the future. Clearly, these are not KPN ambitions."

IMS is a long-term investment designed to support a collection of blended services and to enable easy service upgrades, Kamal-Saadi says. Once installed, he says, IMS' control functions and central user database, the HSS, are reused with each new service. Unlike a proprietary system, no new switching equipment and associated user-data store need be added with each service. Only the IMS application layer is touched with the addition of an application server.

Van der Made also said he believed that the added value of converged/blended services hardly exists at the moment, and he was critical of IMS products that don't allow mixing components of different suppliers and are short on interworking possibilities because of limited flexibility. System-integration costs are also high, and lead times are long, because vendors are still learning about IMS.

"I don't see value in value-added services over IMS right now, and I don't see a lot of suppliers coming in with open interfaces," van der Made said. "Also, while pre-IMS is a great way of proving whether or not things work, I don't want to be the person that everyone learns from. It's costing me and not them."

Kamal-Saadi also agrees that the lack of killer applications is weakening IMS' business case, adding that VoIP is always presented as the killer application, because its technology might allow combinational data- and voice-service offerings to enhance the user experience. VoIP could act as an anchor service onto which IMS can be built, but the problems of interoperability are real issues.

"Vendors continue to offer IMS systems with relatively closed interfaces, which makes the system-integration process quite lengthy and costly," Kamal-Saadi said. "This is legitimate when competition among vendors is intensifying, and these vendors obviously want to protect their investments, especially in the early stages of technology deployment, such as IMS, so proprietary implementations will persist."

Van der Made also said that KPN wanted more time to develop IP-based services, because emulating the PSTN over IP has proved more difficult than was at first envisioned and is taking far longer than planned.

He said that KPN would not be switching off the old network until it could offer identical functionality and QoS on the IP network. "We're not switching networks until we can do it properly," he said. "Of course, we don't want the two networks sitting side-by-side for too long, but we're prepared to wait and see how others resolve the problem."