Nortel Unveils Passport 20000
The Passport 20000 quadruples the capacity of the vendor's existing Passport 15000 (from 40 Gbit/s to 160 Gbit/s), while adding gigabit Ethernet, OC12, OC48, and OC192 interfaces, plus the ability to manage VPNs (virtual private networks) in native Ethernet mode. The box also supports Multiprotocol Label Switching (MPLS), as does the 15000.
“Carriers are asking, 'What can I invest in today that will drive ATM and frame relay evolution to MPLS?' The Passport 20000 is a vehicle for doing that,” says Oscar Rodriguez, president of Nortel’s Intelligent Internet division.
Unfortunately, some of the box's more compelling features, including support of gigabit Ethernet and the ability to support Layer 2 VPNs via native Ethernet, won't be available until the second half of this year.
Also missing is a customer testimonial, which was generally expected to surface with today's news. Nortel claims it's been trialing this box for the last year or so and plans general shipment to a limited number of paying customers this quarter and next. But there's nothing to show for these claims.
How much do the delays matter? Perhaps not that much -- yet. Nortel appears to be in sync with the timeframes of Alcatel SA (NYSE: ALA; Paris: CGEP:PA), Équipe Communications Corp., Lucent Technologies Inc. (NYSE: LU), and Marconi PLC (Nasdaq/London: MONI). All of these vendors are just starting to issue "in trial" announcements of newly equipped Asynchronous Transfer Mode (ATM) and frame relay switches that also support Internet Protocol (IP) and MPLS (see The Great ATM Switch Blitz).
Nortel and its rivals say they're responding to renewed focus on equipment thought to be "legacy" just a few months back. "One and a half years ago, carriers wanted to converge all services. Now there is a pushback," says Todd Morris, director of the Passport line at Nortel.
Nortel appears to think there's plenty of time left to sell ATM and frame relay switching for its own sake -- at least in core networks. While the vendor is clear that MPLS support is important because carriers will need it to migrate to fully packetized networks over the next four years or so, spokespeople say MPLS isn't a panacea.
"Carriers don't think MPLS is carrier-ready," says Morris. Instead, carriers are focused on supporting voice, data, and video services the traditional way -- via ATM.
Still, Nortel says it's vital to stay abreast of MPLS in order to prove its feasibility and pave the way toward future protocols such as GMPLS (generalized MPLS), which Morris says many carriers are looking forward to adopting in the future.
That protocol, a subset of MPLS, adds the ability to manage quality of service over a variety of transports, including ATM and wavelengths. But the jury's out on when, if ever, it will be fully adopted (see Optical Signaling Systems).
In the meantime, Nortel says another opportunity may be just as urgent -- the need to create VPNs via Ethernet. Right now, Nortel and other vendors support the creation of VPNs via IP routing at Layer 3. But they say corporate users are increasingly demanding Layer 2 VPNs for increased security and reliability. Ethernet also adds speed.
Nortel recently touted Ethernet VPNs in announcing new metro products (see Nortel's Optical Ethernet Blitz). But the vendor won't have its Ethernet VPN functionality in Passport 20000 for another few months.
Only time will tell whether Nortel will suffer in the market as a result of the delay. Much depends on how quickly its rivals provide Ethernet VPNs in their gear. So far, announcements have been made, but shipment and adoption are still in staging.
All this raises many questions: Are Nortel and other vendors hedging their bets about the need for certain features? Are they waiting for MPLS -- or something else, such as the VPN alternatives being studied by the Internet Engineering Task Force (IETF) -- to emerge in force for VPN creation? And as carriers start to spend on next-generation equipment again, will the addition of functionality to ATM-based switches become passé?
Some analysts say they are concerned about the ultimate shelf-life of platforms like the Passport that are still cell-based. "I think the issue is whether people will continue betting on cell switches," says David Passmore, research director at the Burton Group consultancy. "Sure, you can give your ATM switch a lobotomy to support MPLS... But eventually, packet-based cores will emerge that require a packet-based architecture."
Nortel's ready to take the chance. Time and the pattern of carrier investment indicate that ATM is an optimum launching pad for new services, the vendor says. At least for now.
— Mary Jander, Senior Editor, Light Reading