Level 3 Rekindles Optical
Tom Issenhuth, principal architect of core network development at Level 3, described the project during his talk at Heavy Reading's Links 2005 summit last week. He also gave attendees some insight into his thinking about the state of optical technologies.
"We've been rebuilding our entire optical footprint in a six-month period -- that's U.S. and Europe," he said. The core of the new network is overbuilt by a factor of two or three, leaving space for future growth.
In a sense, Level 3 is declaring that it's finally time to spend on optical again.
"Reuse opportunities are getting tapped out," Issenhuth told Light Reading after the session. "We're trying to get ahead of the curve," considering the demand is growing and some new technologies have shown themselves to be deployable, he said.
Level 3 isn't alone. In its earnings call last week, components vendor JDS Uniphase Corp. (Nasdaq: JDSU; Toronto: JDU) noted increases in metro and long-haul activity. (See JDSU Sells More, Cuts More.)
Issenhuth said carrier Ethernet is great but requires the operations, administration, and maintenance (OAM) functionality that's been promised. "We realize there are hardware changes to go do it, so please -- go do it," he said.
He also said GMPLS is a must. "Anything that adds some intelligence is important."
Following Issenmuth's talk, Heavy Reading chief analyst Scott Clavenna pointed out some trends in today's optical networking plans. Carriers say their requirements for future optical networks include Gigabit Ethernet and 10-Gbit/s Ethernet support -- nodding to the growing presence of Ethernet in access and metro networks -- as well as an intelligent control plane to handle dynamic IP routing, he said. For the latter, carriers say GMPLS doesn't provide everything necessary but has value.
Clavenna also noted carriers want reconfigurable optical add/drop multiplexer (ROADM) functionality -- nice news for the session's panel, which included four ROADM vendors. But that's a metro requirement. Issenhuth noted -- and panelists agreed -- that his core network didn't have much use for ROADMs. It's the metro network that needs the dynamic switching of wavelengths, whereas core wavelengths, he said, sometimes "stay up for years."
— Craig Matsumoto, Senior Editor, Light Reading