GIG-BE: Guinea Pig to the RBOCs

GIG-BE: Guinea Pig to the RBOCs – global information grid bandwidth expansion project

Byline: Ed Gubbins

After years of trudging laboriously through the 271 long-distance approval process, knocking down the LATA boundaries that force local carriers to surrender their traffic to interexchange networks, RBOCs have earned the right to control much of their own inter-LATA traffic. But for the most part, RBOCs with long-distance approval still pay wholesalers such as WilTel Communications and Level 3 Communications to do their LATA-hopping transport for them, which is a little like taking the bus even after you get your driver’s license. It’s cheap, sure, but you don’t want to do it forever if you don’t have to.

So RBOCs now find themselves in uncharted territory, with the opportunity to build networks the likes of which, from a functional standpoint, no one has seen since Ma Bell, but with technology Ma never dreamed of. In the process, they are confronting new questions about how to offer inter-LATA services in the hopes of stealing IXC customers.

As RBOCs debate the various ways to build these new networks, they will soon have the benefit of a test case to observe: The federal government is building an immense communications grid with many of the same properties RBOCs would want in their own networks.

The global information grid bandwidth expansion project (nicknamed GIG-BE) – proposed by the U.S. Defense Information Systems Agency and overseen by Science Applications International Corp. (SAIC) – is designed to serve as a resilient, big-bandwidth communications network for about 100 military installations around the world.

GIG-BE employs the kind of cost-efficient, all-optical technology that has been dangled above the U.S. telecom industry for years without wide-scale deployment. It includes the type of multiservice provisioning layer BellSouth is deploying now. And at the end points, it even boasts the same kind of passive optical networking for fiber to the premises the RBOCs are now promising for the public network.

In addition to serving as a convenient test bed for wide-scale deployment of these technologies, GIG-BE poses some interesting questions for RBOC network planners. For example, GIG-BE is said to deliver an entire wavelength to every military base in its network.

“If I can do a wavelength to every base in the military, I could do a wavelength to every LATA in an RBOC,” said Tom Nolle, president of consultancy CIMI Corp. “What are the consequences of that? With wavelengths so cheap today, it doesn’t make sense to put electrical service technology in place, betting on whether customers will buy Ethernet or Fibre Channel. What [RBOCs] really ought to do is just offer a wavelength between two points within a metro area and let the CPEs decide whether it’s Fibre Channel or Ethernet.”

Taking the lead among RBOCs, Verizon Communications unveiled a bold new architecture plan this spring called IFON, for IntelliLight flexible optical networking, that somewhat resembles GIG-BE. IFON would link all the regional networks Verizon has been quietly building in each state as it won long-distance approval, funneling traffic into an IP backbone through super switches that would preside over each LATA. As with GIG-BE, the wavelength is the basic unit of transport.

“Verizon’s IFON RFP was directed at creating light paths between LATAs, minimizing the service intelligence that has to take place in long-haul networks – technology that’s very similar to what GIG-BE proposes to do,” Nolle said. “One was influenced by the other, or they were both influenced by a common appreciation of technology considerations. Either way, it’s inconceivable that there isn’t some cross-pollenization taking place.”

If so, that cross-pollenization could increase dramatically as the networks develop further. Winning the GIG-BE bake-off that’s currently underway, for example, could help start-up equipment vendors gain legitimacy with RBOCs on many levels: They will have passed the scrutiny of rigorous government testing, and they’ll enjoy the financial stability of a nine-figure paycheck. As a result, vendors who win a piece of GIG-BE might more easily win a piece of the RBOCs’ business, too.

Bubble-fed vendors Corvis and Ciena both have innovative long-haul products that haven’t yet been invited into RBOC networks. That could change, however. Both start-ups are finalists in the long-haul portion of GIG-BE, with a good shot at winning the deal.

As SAIC narrows its choices over the next few months, Verizon could be mulling over many of the same candidates for its own project – a long-haul optical switching RFP issued in June. The vendors in the running for the Verizon RFP are “a healthy mix of traditional long-haul players and, surprisingly, some very serious start-ups with viable, innovative products,” according to Stuart Elby, Verizon’s vice president of network architecture and enterprise technology. CIBC World Markets analyst Rick Schafer added that the Verizon RFP looks “very favorable to Corvis.”

“On its own, [winning the GIG-BE contract] wouldn’t be enough to get a vendor on our short list,” Elby said. “But if one of those vendors ends up on our list anyway, [the win] gives them an advantage. It separates them from their competitors.”

Of course, the federal government doesn’t necessarily share the same buying habits as the RBOCs. For example, RBOC business models require them to double-source each equipment contract and avoid vendors that depend too much on a single customer. The government, on the other hand, isn’t burdened with such qualms.

And there are important differences to note between the GIG-BE project and RBOC networks. First, GIG-BE will include new next-generation fiber that is capable of transmitting 10 Gb/s, 40 Gb/s or possibly even higher speeds. RBOCs, however, can’t justify replacing all the old glass they have in the ground, so they have to make do with slower speeds: 2.5 Gb/s isn’t a problem, and 10 Gb/s is attainable in most interoffice fiber, but according to Elby, “That’s where the penalty starts to hit us with the older fiber.” So instead of serving a LATA with a 40 Gb/s wavelength, Verizon is more likely to use four 10 Gb/s pipes or 10 2.5 Gb/s pipes.

According to Nolle, the RBOCs could even diverge from the government’s decision to include a multiservice provisioning platform element for interoperating with TDM-based service technologies. “The RBOCs are seriously looking at the question of whether they need to have such a layer,” he said. “The RBOCs don’t have any legacy long-distance business to support. They don’t have the same demands on their networks to support legacy [traffic] as the government does.”

Verizon’s Elby doesn’t quite see it the same way. “Any new gear I pick still has to interconnect to all the old stuff, even some of the Sonet stuff from the late 80s,” he said. “Operationally, that’s much harder than building this isolated GIG-BE network. Also, GIG-BE is being built for very specific services and applications. Our transport network has to carry everything, from the lowly 56-kb/s voice channel up to a 10-Gb/s Ethernet channel.”

As Verizon vicariously watches GIG-BE move forward, the other RBOCs are likely to vicariously watch Verizon, said Nolle, letting it test the waters both with the technology and with Wall Street. But the myriad other floating objects in the mix may prevent much decisive forward motion – chief among them, the uncertain fate of the major IXCs. If MCI were to emerge from bankruptcy with a vengeance, slashing long-distance prices, the RBOCs might opt to forego their grander network plans in favor of smaller steps.

“They would probably focus on adjacent-LATA services,” Nolle said. “Instead of building pervasive networks, simply try to extend service selectively from LATA to LATA. That’s not a bad way to test out the concept. I think they can leg into this.”

In one architectural sense, Verizon could even be leapfrogged by the carrier often regarded as the last in line for this evolution. In a currently active RFP, Qwest Communications is evaluating gear from start-up Mahi Networks that would allow it to do away with lots of expensive electronics and plumbing in its central offices by substituting stacks of legacy add/drop multiplexers, or ADMs, with single devices that incorporate both ADM functionality and that of the broadband digital cross-connects some carriers use to interconnect them. Verizon has been interconnecting its ADMs with broadband digital cross-connects for more than a year, which reduces the cost of stacked ADMs by interconnecting them optically rather than electronically.

“Verizon has taken an intermediate step that relieves some of the pain,” said Adam Shelton, director of product marketing for Mahi (where the CEO, Chris Rust, is a former broadband network architect for US West). “Qwest is behind right now in the sense of physical deployment, but they’ll be leaps ahead of everybody by the time they deploy. They’ll have the next-generation product deployed in the next six months when the other RBOCs might be six to ten months out.”

If economic forces and industry uncertainty delay RBOC network evolution, it would give the Bells more time to watch the development of the “fast-tracked” GIG-BE, which wouldn’t be deterred by those fluctuations in climate. Otherwise, the RBOCs might have to move forward on their own.

“I certainly think they’d want to use GIG-BE as guidance,” Nolle said. “But I don’t think they’d wait that long.”


To stay up-to-date on all the optical innovations coming out of this week’s NFOEC event in Orlando, click on the news tab of our Web site. WWW.TELEPHONYONLINE.COM

COPYRIGHT 2003 PRIMEDIA Business Magazines & Media Inc. All rights reserved.

COPYRIGHT 2003 Gale Group