Waging a switching war

The goal for Ipsilon Networks this year is simple: "Drive Cisco Systems crazy."

That, according to Larry Blair, the start-up's vice president of marketing. To do it, Blair and his cohorts say they are promoting a new kind of switching protocol calculated to annoy and harass networking's entrenched industry leaders by making them rethink some of their best-established products. Unfortunately, the fight over yet another new standard may end up driving networking customers crazy too.

The new protocol is actually an old protocol, the Internet Protocol or IP. Ipsilon and others are combining the Web's pervasive protocol with ATM (asynchronous transfer mode) backbones to let administrators run intranet applications over the wide-area networks that connect the far-flung outposts of corporate empires with regional headquarters.

Essentially, the IP switching concept works like this: Take existing networks based on TCP/IP and layer the protocol on top of an ATM "backbone," the pipe that interconnects multiple LANs. Proponents of the technology argue that it negates the need for a heavy-duty router--an expensive and complicated piece of networking equipment that functions as kind of networking traffic cop--in most parts of a network. It also means that administrators can come up with one set of management policies for an entire network.

Currently, the IP switching market is in its nascent stage, but analysts expect the concept could have a profound affect on how networks are upgraded in the future. "This is the next era of the switching wars," said Skip McCaskill, an analyst at market research firm Gartner Group.

Ipsilon debuted its version of IP switching last year, a concept radical enough to force an almost-immediate response from a competitor that should have been far out of Ipsilon's league: Cisco.

Not wanting to be left out of the publicity generated by the new IP switching idea, Cisco introduced its own take on the technology last fall with a protocol called Tag Switching. Predictably enough, both companies would like their versions to be hailed as Internet standards by the Internet Engineering Task Force (IETF), the main standards body for Net protocols. Expect a year of maneuvering by both camps to insure that both IP switching protocols have the weight of vendor support behind them.

Late in the year, Cisco convinced 50 vendors to proclaim their support for the Tag Switching concept. On the other hand, Digital Equipment has signed on to resell Ipsilon's switching products.

While the fight between Ipsilon and Cisco is the most dramatic example of the change sparked by IP switching, all the other networking stalwarts are scrambling to integrate the IP switching concept into their own gear.

Today, Madge Networks introduced a switching strategy for IP and Novell IPX networks. Digital Equipment also announced a packet-based IP switching implementation for Fiber Distributed Data Interface and Ethernet networks.

3Com has introduced Fast IP, a switching concept based on IP. IBM has proposed an IP switching concept called Aris that is similar to Cisco's Tag Switching proposal, and Bay Networks has talked of an IP switching-based offering as well. Cascade Communications recently released a software package for IP-based networks that supports Ipsilon's and Cisco's IP switching protocol.

But of everyone who stands to lose from the new IP implementations, Cisco may be the most vulnerable.

After all, Ipsilon's products are shipping already while Tag Switching implementations are not expected to hit the market until the middle of the year.

Cisco argues that its method allows IP-based networks to scale for large installations far better than Ipsilon's implementation, according to Tom Downey, Cisco's director of product marketing for its Core Business Unit. Ipsilon's version creates too much signaling between network devices and creates too many data flows for a typical network to handle.

He added that Cisco had been working on an IP switching concept before Ipsilon's debut.

But analysts note that the small upstart certainly caught the attention of the internetworking industry leader. "Cisco doesn't seem to be paying a whole lot of attention to what 3Com and Bay are doing," observed Don Miller, an analyst at market researcher Dataquest. "Cisco is directing all this attention toward this pipsqueak start-up."

"Cisco has to come out with something. They have an aging architecture and an old product line," Miller continued. "It's imperative for them, more than anybody else, because they have the lion's share of the market and they need to protect their customer base."

On the other hand, this year could be a make-or-break one for Ipsilon, as it tries to gain customer converts, attract third parties to resell their products, and fend off increased competition from a variety of players.

"They've got the ideas, they've got the technology, but since they're a start-up they don't have the market presence," noted Gartner's McCaskill.

Other observers echoed that thought.

"They're dancing around under the elephant and if they don't watch out, they're going to get stepped on and it's going to be ugly," added Dataquest's Miller.