The number of core routing suppliers continues to shrink, as Avici Systems Inc. (Nasdaq: AVCI; Frankfurt: BVC7) became the latest to announce it was exiting the business yesterday. (See Avici Abandons Routing, Targets PBT and Avici Quits Routing.)

The move marks the end of an era in which startups competed against major players like Cisco Systems Inc. (Nasdaq: CSCO), Juniper Networks Inc. (NYSE: JNPR), and Alcatel-Lucent (NYSE: ALU) for core routing market share.

The last five years have been filled with stories of failed businesses built on providing core routing products, and Avici joins a long and storied list of players that have either quit the business, gone under, or been swallowed up by larger competitors.

Give Avici some credit: It lasted much longer than other smaller players in the space. It made it to the public markets, and unlike a number of its competitors, it eventually turned a profit.

The same cannot be said for Caspian Networks, for instance, which swallowed up $300 million in VC funding and changed its business strategy twice before eventually closing its doors in 2006. (See Caspian Closes Its Doors.)

Other failed routing startups include Pluris, which ran through $215 million of funding before shutting down, Chiaro Networks ($210 million), Hyperchip ($140 million), Axiowave ($120 million), and IronBridge ($85 million), none of which provided much return on their VC's investments. That's more than $1 billion of funding washed down the drain over the course of just a few years. (See Pluris Shutdown Confirmed, No Tomorrow for Chiaro, Hyperchip Has a Hiccup, Axiowave Shuts Down, and Ironbridge's Last Ditch Efforts Fail.)

That list doesn't include Procket Networks, which had raised more than $300 million before being acquired by Cisco at the bargain price of $89 million in 2004. (See Procket Stuffing Its Pocket and Cisco to Pay $89M for Procket Assets.) Cisco quickly capped off the acquisition by discontinuing the Procket product line, saying it bought the company for its personnel and patents. (See Procket Reaches 'End of Life' and Valley Wonk: The Procket Puzzle.)

"I think there was all sorts of money chasing bad ideas back then," says Prudential Equity Group LLC analyst Inder Singh, remarking on how so many core routing startups got so much funding in between 1999 to 2002.

Ovum RHK Inc. analyst Mark Seery says the rush by startups to develop core routers was due in part to inflated expectations for the market in the mid-90s.

"The assumption in the mid-90s was that the big problem to solve was the bottleneck in the core," Seery says. However, "by 2001 it was no longer seen as a huge problem due to the way the telecom market imploded."

It also didn't help that so many vendors all sought to tackle one of the toughest sectors in the telecom equipment market.

"The core router is probably the hardest product to build in the telecom space," Seery says. "It is very hard to get a team together to build the products and very expensive to do, and very difficult to maintain financing over several generations of products."

On the carrier side, Singh says operators have been careful about the products they implement. "In choosing a core router, a carrier is making an architectural decision," and that "to back a startup, that can be a dicey proposition."

Carriers are looking for integrators and for vendors "that can provide an end-to-end solution," Singh says, which narrows the field of potential suppliers.

Cisco and Juniper are the clear leaders in the core routing market, what Singh calls a "duopoly," but there are other vendors in the sector. Also in the mix are Alcatel-Lucent and Huawei Technologies Co. Ltd. , which have core routing products to go with their end-to-end product portfolios. (See Alcatel Sticks to Core Plans and Huawei Goes Hard Core.)

And Ethernet players Extreme Networks Inc. (Nasdaq: EXTR), Force10 Networks Inc. , and Foundry Networks Inc. (Nasdaq: FDRY) have interesting stories to tell, as each has built core functionality into its Ethernet products. (See Foundry Claims Core Crown and Force10 Goes Terabit.)

The days of the pure-play core routing startup may be over, but that doesn't mean there isn't room for niche players with new technologies. Singh says there may still be opportunity for startups to innovate in the core routing market by working with the major players, but "they're not going to compete head-on with Cisco and Juniper."

— Ryan Lawler, Reporter, Light Reading