Kestrel Emerges from RBOC Trenches with Certified Next-Generation Gear
Like other optical networking start-ups, Kestrel Solutions (Mountain View, CA) has secured a huge pot of money to execute its business plan, actually turning away investors in what it deems a record $105 million fourth round of financing. But unlike other start-ups, Kestrel's target customers are the stalwart regional Bell operating companies (RBOCs). By subjecting its next-generation technology to the pain of Bell certification procedures, Kestrel hopes to open the door to the huge RBOC equipment market and subsequently win over competitive local exchange carriers with the technology.
"We've been aiming our engineering at an RBOC-quality product," says Kestrel chairman Meldon Gafner. RBOCs comprise 2/3 of the addressable market for the Kestrel product, he notes, reasoning that the RBOC market was a good place for Kestrel to start. "If you get major carriers to certify your product, CLECs will be right there," he says. Kestrel's Talon product has been in carrier labs undergoing testing since December 3, 1999. The customer list includes three major carriers, two of which are Bell companies.
RBOCs, CLECs, and start-up vendors
Kestrel's RBOC-targeted approach differs from conventional start-up wisdom. RBOCs are notoriously slow to adopt new technology because of laborious, and some would say bureaucratic, compliance requirements and tests. The RBOC-rival competitive local exchange carriers (CLECs) may deploy new technology more aggressively than RBOCs for competitive advantage. The administrative and cultural barriers put forth by RBOCs compared with the prospect of rapid deployment by CLECs prompts many start-ups to target new carriers as first customers for new optical networking technology.
Kestrel has raised a total of $185 million for its cause, which has yielded the TalonMX, a PMD-immune, 10 Gb/s optical metro add/drop multiplexer based on optical frequency division multiplexing (see Metro Optical Transport Start-up Proposes Non-WDM Transmission). "It takes that amount of money to deliver all the OSMINE stuff," says vice president of marketing Dawn Hogh, wryly referring to the both venerable and vilified Telcordia certification procedures: Operations Systems Integration and Modification of Network Elements (OSMINE), and Network Power, Protection, and Network Equipment Building Systems (NEBS). Having recruited employees with a history of working with RBOCs at all levels of the technology procurement chain, Kestrel considers itself well positioned to succeed in the RBOC space.
Broadband access adaptations
One benefit to working closely with RBOCs early in product development cycles is the ability to adapt the product to unforeseen, yet significant, changes in RBOC strategy. Kestrel didn't see digital subscriber line (DSL) technology as the initial, primary driver for its optical transport equipment when it started product development, Meldon recalls, but now touts the capability as critical (see figure 1).
"The single most important thing in the RBOC world is DSL rollout," Gafner declares. The RBOCs are at war with former parent AT&T for broadband access, Gafner surmises, rolling out DSL equipment to compete with AT&T's cable properties.
"DSL rollout meant we had to make a couple of modifications to work in controlled environmental vaults [CEVs]," Gafner says. The vaults are remote transport nodes that house local loop equipment and between four and 20 fiber connections in a 10x16-foot space. The CEVs are typically served by 180 Mb/s asynchronous transmission equipment, OC-3 (155 Mb/s) synchronous optical network (SONET) multiplexers, or, in some cases, OC-12 (622 Mb/s) equipment. "We've been in lots of them, let me tell you," Gafner reports.
Beyond rolling out few megabits of DSL per customer, the Holy Grail of broadband access for RBOCs is to send three separate high-definition television signals into a home, Gafner says. Some 50 Mb/s of bandwidth will do the trick, with signal compression. Working backwards from that, a CEV requires 30 Gb/s or 40 Gb/s of capacity, Gafner calculates.
In addition to accommodating DSL rollouts by designing Talon equipment to operate in CEVs, Kestrel yielded to the preponderance of asynch equipment by including asynch interfaces in the first release of the product, instead of a later release as originally planned. A cross section of circuits going down RBOC fiber in metro environments today yields 80% asynchronous or OC-3 signals, and 20% OC-12 and OC-48 (2.5 Gb/s), Gafner reports.
Transport alternatives
Old and new metro DWDM systems also handle a variety of interfaces, but Kestrel maintains that the different variety of DWDM approaches don't hit the metro access sweet spot like the Talon, which appears to the customer like an OC-192 with OC-3 granularity. But like the DWDM crowd, Kestrel has had to convince win carriers over from just pulling more fiber and plain old SONET equipment to scale metro networks.
"We have not had head-to-head competition with DWDM because of cost," Gafner says. "[RBOC's] don't see DWDM as cost-effective yet with all their add/drop and OAMP needs," Gafner says, noting the exception of inter-office transport applications. New metro access platforms that address these needs will not be commercially available as soon as the Talon, Hogh reasons, extending similar reasoning to non-DWDM, next-generation access products. "[RBOCs] won't deploy anything that's not OSMINE certified—that leaves a lot of the next-gen guys out," Hogh says.
Interoffice transport DWDM systems do not prove to be economically attractive with per-channel bit rates lower than OC-12. Introducing per-channel submultiplexing on these systems to share channels among several lower-rate services incurs cost and management disadvantages, Hogh maintains. Plus, DWDM filtering, wavelength configuration, and other high-performance optical attributes do not lend well to simple engineering and operations. "There are not a lot of optics PhDs out there working on outside plant," Hogh points out.
"We get compared to existing products," such as asynch and SONET, Gafner says. The Talon is competitive with OC-48 SONET multiplexers, where more than an OC-12's worth of capacity is required, Hogh explains. The Talon sports the same first cost as the OC-48, scales to an OC-192's worth of capacity, yet costs 30% to 50% less than an OC-192 multiplexer over the lifetime of the deployment, she maintains.
Despite these disadvantages, OC-192 SONET is being deployed on new fiber routes. But on old fiber, a major consideration with RBOCs, polarization mode dispersion (PMD) often makes the use of OC-192 prohibitive. PMD immunity is one of the benefits of Kestrel's novel multiplexing approach and forward error correction. One Kestrel customer couldn't get OC-192 equipment to work on a 64 km route—PMD made the bit error rate too high. The carrier fired up the Talon on the route, which worked flawlessly, Hogh reports.
By: Erik Kreifeldt