100G backers seek to prevent repeat of 40G delays

By Stephen Lawson, IDG News Service
June 19, 2008 04:10 PM ET

. What's this?

Carrier backbone vendors working on 100G-bps (bit-per-second) network equipment are trying to head off a repeat of the rocky development cycle for 40G-bps gear.

On Wednesday, representatives of companies in the Road to 100G Alliance spoke on a panel at the NXTcomm trade show in Las Vegas. They said they are organizing early to make sure that implementation of various pieces of the technology is coordinated before 100G products hit the market, probably in 2010. The group was formed last year.

The alliance accelerated its mission on Monday, forming a technical committee to fill in gaps in interoperability among elements that are under development for 100G products. To start addressing seven key technical issues, the committee formed working groups on optical interfaces and electrical interfaces.

Some of the officials said the work is essential to prevent long delays like those that plagued 40G-bps technology. That gear, the fastest generally available today for optical wave-division multiplexing networks, hit mass adoption by service providers last year after development efforts that began in the late 1990s.

Related Content

Service-provider mergers, carriers' consolidation of multiple networks, growing mobile data subscriptions and IPTV (Internet Protocol television) are creating demand for fatter pipes across the core of carrier networks, according to Infonetics Research analyst Michael Howard. Specifically, service providers want equipment that can send more traffic over a single wavelength of light so they can make more efficient use of the optical fiber in their networks, he said.

Commercial shipments should begin in 2010, said Howard, who moderated the discussion on Wednesday. In the meantime, carriers will be aggressively rolling out 40G gear to meet growing demand, he said. By 2012 or 2013, spending on the faster products should surpass investment in 40G equipment, Howard said.

But developing this kind of technology is "rocket science," Howard said. "The technical requirement is not just two and a half times 40 to get to 100. It's not that easy," he said.

There is already a need for 100G-bps connections in data centers, but not yet on wide-area networks, according to Ovum analyst Mark Seery. Standards are needed but component costs must also come down first and the next two years will be dominated by the now mature 40G, he said.

"It's not too early, but people should temper their expectations a bit," Seery said.

Development of 40G technology began about 1998 amid the dot-com boom's heated predictions of demand for bandwidth. Networks ended up being overbuilt before 40G gear really hit the market, so early entrants to that market became victims of the telecommunications crash beginning in 2001. But that wasn't the only problem, according to Niall Robinson, vice president of marketing at Mintera an equipment maker in Acton, Massachusetts.

Early 40G development was so uncoordinated that few components worked together, Robinson said. In 2003, building its first product with 40G support, Mintera had to use 18 different chips, including many it had to develop to make incompatible chips from silicon vendors work together, Robinson said. More chips meant higher cost.

The IDG News Service is a Network World affiliate.