No winners in the router wars

There might be a router war starting, but what’s the battleground?

Earlier this month, start-up Axiowave released its XCR128 convergence router. Last month, Cisco announced its CRS-1, and Juniper announced its Infranet architecture. Earlier in the year, Avici announced a marketing deal with Nortel. Tellabs is rumored to have won a deal with Verizon on the product it acquired with Vivace. On the flip side, start-ups Tenor and Equipe folded. There might be a router war starting, but what’s the battleground?

Axiowave thinks the battleground is QoS and legacy services, and you could also say the CRS-1 is Cisco’s belated adoption of a more Multi-protocol Label Switching-centric architecture, one its competitors adopted earlier. Even Avici’s focus on high availability could be called a legacy-over-IP story, given the traditional five-nines telco requirement for reliability. But legacy and QoS aren’t the answer, because every major carrier acknowledges that legacy service revenues are declining at near double-digit rates.

Tellabs, Vivace, Avici, Nortel, Alcatel . . . all of the partners or buyers in the recent deals between start-ups and incumbents think the router wars are all about relationships. The old-line equipment vendors’ sales forces know whose hands to shake, so give them a chance to sell vital new gear and they’ll turn the market on its ear. Certainly partners are a necessity for start-ups, owing to the historical conservatism of major carrier buyers, but no carrier is going to buy a lousy box from an old sales friend if it compromises the buyer’s career. A strong sales conduit can move a good product into the carrier arena, but not a weak one.

Juniper’s Infranet announcement suggests that the key to the router wars is a different kind of relationship, one between networks and computing intelligence. But Juniper’s acquisition of security edge player NetScreen Technologies earlier this year might provide the real insights. Juniper is moving to the network edge, away from its traditional core incumbency – moving into corporations, broadening its markets beyond its familiar carrier focus. The simple reason for the move is that all hardware sales reduces to selling lots of boxes, and the best place to do that is where most boxes get installed. Look at any network diagram and you’ll see thick concentrations of devices at the edge and a sparsely populated core. The world market for terabit routers is in the hundreds of units, but one company easily could buy 1,000 edge routers.

But it’s not just the “more boxes at the edge” truth that makes the NetScreen deal interesting; it’s the security angle. Cisco’s deal with Linksys shows that the inevitable result of mass-market broadband is commoditization of the customer edge. The router that Cisco used to sell to branch offices for $10,000 now could be put on a chip and sold to a teenage gamer for $100. Does anybody really think Linksys will be just a cheap router in the future? It will be a digital rights manager for content, a VoIP portal, a bunch of higher-value things – because it has to be to sustain profits. Just fleeing to the edge doesn’t necessarily create profit growth. You have to flee routing as well.

The router market is maturing, the vendors along with the products. But it’s also dying as we know it. At the core, it will be pressured by optics and at the edge by increased competition and the need to lower the price of the boxes to drive a market expansion focused more on consumers and small and midsize businesses. Security and cost concerns will drive most of the traditional router features out of the customer edge devices. All the features of the network of the future will end up concentrated in a thin layer at the provider edge, connecting commodity customer premises equipment with fat-pipe optics.

The trends we see today aren’t a sign of a router explosion, but the beginning of an implosion. There will be no winners in the router wars; the survivors will morph into something else. The router giants of today will be giants tomorrow because they’re able to figure out what the successor product to routers will be.