The 40G transmission flap: AT&T’s story

While MCI and Sprint split hairs over who was first with their recent 40G bit/sec trials, rival AT&T points to its testing of high-speed Internet services that took place in the summer of 2002.

The 40G flap started when MCI announced on May 24 that it had successfully sent IP packets from San Francisco to San Jose at 40G bit/sec in what it claims was the fastest IP transmission ever. MCI achieved this milestone by using the latest routers from Cisco and transponders from StrataLight. MCI’s field trial lasted two weeks, and involved simulated traffic including HDTV-quality video conferencing, gaming and music downloads.

Nine days after MCI made its 40G announcement, Sprint issued a press release announcing a similar 40G trial. Sprint claimed it was the first ISP to carry live, production Internet traffic from its business customers over a 40G link between Stockton, Calif., and San Jose.  Sprint put its recent trial together using two of Cisco’s latest routers, transponders from StrataLight and DWDM equipment from Ciena.

We covered MCI’s 40G announcement in last week’s issue of the ISP News Report and Sprint’s response to MCI’s claims in Monday’s issue.

Meanwhile, AT&T has yet to respond with a press release of its own on its 40G plans. Nonetheless, an AT&T spokesman asserts that AT&T was the first ISP to test 40G bit/sec transmission over DWDM with a 1,000 kilometer loop between New York and Philadelphia. AT&T says it completed this trial in July 2002, and announced it in a presentation paper at the Optical Fiber Communications Conference in March 2003.

Last November, AT&T followed up on its 40G trials with the announcement that it would be testing next-generation optical transport gear from Siemens for use on its high-capacity routes. Siemens says its SURPASS hiT 7500 system is ultimately capable of supporting transmission speeds up to 40G bit/sec. AT&T said in November that it planned to deploy the Siemens gear in its network during the first half of 2004.

So which ISP is ahead on 40G service deployment? Telecom analysts say the race is too close to call.

“There is not a clear indication of who is first in 40G service,” says David Willis, vice president of technology research services at Meta Group. “Linking a couple nodes is very different from having production services across the entire core. Nobody has completely done that yet, although they are all on the path.”

Willis says the more important question is whether enterprise customers even want 40G service. He says ISPs may deploy 40G service on their backbones to improve their traffic engineering, but he says the technology is a long way from being a selling point to business customers.

“There are few customer edge requirements for 40G service today,” Willis says. “We’re just getting to the point where data center interfaces and trunking interfaces need to be at 10G. I estimate it will be about six years before 40G service is necessary for enterprises.”

Enterprises that are likely to be first adopters include financial services firms and the federal government, particularly the Department of Defense, Willis says. “Storage interconnections or mainframe channel connections or disaster recovery are all potential applications for that level of capacity,” he adds.

Indeed, rather than boasting about their 40G deployments, Willis says ISPs should be careful about building too much capacity before business customers need it.

“There’s a hazard of jumping too early into this technology, especially if they don’t have a need for it,” Willis says. “If you can’t fill those pipes, you’re better off waiting.”