LESSONS FROM SPRINT'S ION: INTEGRATION IS INCREMENTAL

Four years ago Sprint trumpeted its Integrated On-Demand Network as a breakthrough offering that would revolutionize integrated voice and data services. But a series of cost and technology complications significantly hampered the success of the initiative. Other carriers can learn important lessions form ION's mistakes.

The fewer number of lines necessary to deliver service to customers, the cheaper your network, right? At least that was the idea behind Sprint's Integrated On-Demand Network, also known as ION, which was supposed to meld a variety of services over one connection. The problem was that cheap wasn't really an adjective that described ION, and the “integrated” connection proved problematic.

ONLINE EXCLUSIVE FIRST THINGS FIRST In order for an incumbent carrier to migrate to a packet-based next-gen network, it first needs to phase into time division multiplexing (TDM). Providers can implement packet switching over their TDM networks, according to Ali Kafel, vice president of marketing for Telica, a next-generation switch developer. But while packet telephony can be used inside the switch, it can't be used in the network itself. "You get a lot of benefit of reduced costs, operational costs, and it takes a lot less power," Kafel said of packet telephony. "Using packet switches on the TDM network is the next logical step. You don't need to build a packet network." According to Kafel, Sprint tried to do too much at once with its ION offering. "They tried to do step one and step two at the same time," he said.  By taking a phased approach, providers theoretically can firm up their offerings in test bed areas while they offer other services to existing customers. "The whole process of new customer acquisition is quite expensive, so having a direct route to increasing depth-of-services for existing customers is [key]," said Don Proctor, vice president and general manager of Cisco's multiservice switching business unit, referring to established carriers.  For example, AT&T's Integrated Network Connection is targeted at vertical markets that are particularly burdened by the need to get voice and data services to large numbers of sites. Large enterprises such as Home Depot are an example of that, which can save a lot of money by cutting down on the number of lines. But the major effort and expenditure of building out a next-gen network are only made with large clients sure to bring in perennial revenue. by Liane LaBarba

The notion of building a next-generation network may sound intriguing, but much to Sprint's dismay, the realities of implementing one were not as enticing. Now the question is whether the inadequacies and cost issues Sprint bumped up against have been remedied enough to convince other service providers to take the next-gen leap.

The answer may lie simply in the depth and breadth of a provider's network and size. Turning an older generation network into a next-generation network isn't a task that can be handled in one fell swoop; it must be partitioned gradually. And with the current economic conditions, it must be driven by customer demand.

Sprint launched ION in 1998 using a next-generation ATM-based network. The plan was to offer the one-connection service to commercial and consumer users. Since then, ION has encountered several technological and economic hiccups. In fact, as of press time, it was predicted that Sprint would shut down the project completely later this week.

“With ION we had some good successes, and we had some areas that were not to our expectations,” said Keith Paglusch, president of Sprint's E|Solutions Group. “We are taking a strategic look at it.”

The major rough spots for Sprint developed in the area of voice quality and the sometimes prohibitive costs of last-mile connectivity to customers. With those two factors, plodding onward wouldn't necessarily be wise for Sprint. “We had to ask if the economics of ION could be supported by ION — that's a big question,” said a Sprint spokesman, noting that the company is struggling with its consumer offering.

Now Sprint's objective is to wait until it can make more improvements in ION's voice quality. “The goal is for the packet-switched network to perform at the level of the public switched network,” the spokesman said. “If we want to provide voice, we need to provide a five-nines level of quality.”

According to one analyst, that problem stemmed from vendor equipment. “There were key aspects to the voice and data integration that they didn't work out, and Cisco was really at fault there,” said Jim Lawrence, program director for Stratecast. Cisco Systems' customer premises equipment was used in the ION buildout.

But Don Proctor, vice president and general manager of Cisco's multiservice switching business unit, said he wasn't aware of any problems, though he did note that quality issues with next-generation networks were much greater three to four years ago. “Codecs is getting better all the time,” Proctor said. “We have put a lot of engineering into controlling latency or delay.”

Those issues were similar to the ones that plagued ION. The amount of jitter, or the variation in delay, has been significantly lowered with next-generation equipment to a point that real-time services such as voice and video are unaffected. But without voice quality, Sprint has had a hard time realizing the benefits of its next-generation network, and that was only compounded by the questions left unanswered about how to access customers.

Essentially, ION's major flaws were at the network edge. It was too expensive, and its access methods were too convoluted. Sprint's initial plan was to use leased lines, but then the company partnered with data CLECs to use DSL. After those plans changed, Sprint turned to co-location. “Sprint must have changed its mind five times on what its access [method] was,” Lawrence said.

The challenge was transmitting voice over DSL, which worked much better in theory than in practice. But with ION, Sprint wanted to be different and offer customers something unlike AT&T and WorldCom's services.

“With WorldCom and AT&T pushing down long-distance prices, [Sprint] was trying to lock up the customers before the RBOCs were allowed into those customers,” said Russ McGuire, chief strategy officer for Telechoice. “Sprint said they were going to change the equation and offer a compelling and different [service].”

Despite its mistakes with ION, most agree Sprint was on the right track — it just may have been ahead of its time. The carrier also may have chosen the wrong execution plan.

“Part of ION was to build a unified core infrastructure,” Lawrence said. “And that still is the Holy Grail of convergence. Even before you get to the revenue benefits, the operational advantages [are there].”

But instead of Sprint's blanket approach to the next-generation network, a phased and more gradual migration may have been better. That way, the costs could have been amortized over a longer period of time, and each area could have been handled at its own pace, considering its own challenges individually. With such a phased approach, providers would be able to add more capacity to their current technology with incremental costs and keep that existing technology in place or gradually replace it.

Although the fate of Sprint's ION is in limbo, implementing a next-generation network may be more palatable for smaller providers. “We are seeing smaller companies be more willing to [implement next-generation] networks,” said Proctor, noting that smaller providers often add next-generation services to existing offerings for existing customers.

Not only does smaller providers' nimble size make next-generation networks easier to implement, but “it is part of their survival,” Proctor said.

Luckily for Sprint, its own survival has not depended on the success of its next-gen network.