CS mentioned at end, sorry for length but a good read
Optical Networking: Drawn To The Light
By Joe McGarvey, Inter@ctive Week
January 7, 2000 3:08 PM ET
The heat generated by optical networking technologies last year is only going to intensify over the next few months, on both the product development and the service rollout fronts. On the product side, technology developers will be pressed to produce on several big promises, including the development of optical transmission gear that can send signals over long distances without regenerators, as well as continued progress in replacing today's optoelectronic network designs with all-optical components. Meanwhile, service providers in the U.S. and Europe have some critical deadlines looming for bringing up fiber-optic backbone service on domestic and international routes. It all adds up to a lot of light being promised at the end of the 2000 tunnel.
In for the long haul?
The latest thing in Dense Wavelength Division Multiplexing (DWDM) technology is the ability to stretch optical signals from one end of the continent to the other without having to regenerate those signals along the way. In today's networks, signals have to be regenerated about 30 times in a coast-to-coast connection.
So-called ultra-long-haul DWDM systems started attracting attention late last year. Start-up Qtera won rave reviews for its gear - and then cashed in to the tune of $3.25 billion in a December 1999 acquisition by Nortel Networks. In 2000, long-haul gear from Qtera, Ciena and Corvis will get a thorough workout in field tests and beta trials.
Manufacturers of long-haul optical equipment claim their products can reduce operational overhead by as much as 80 percent in next-generation networks, largely by cutting into the need for amplifiers or intermediate-range DWDM gear or line cards. But some market watchers aren't sold on the need for a 2,500-mile circuit.
"How much traffic needs to go from New York to San Francisco without jumping off somewhere along the way?" asks one analyst who requested anonymity.
Optical multiplicity
Call it Moore's Law traveling at light speed. Although network operators such as Qwest Communications International and Williams Communications are just starting to deploy OC-192 (10-gigabit-per-second) data links in their networks, chances are a few of the more bleeding-edge carriers will be testing OC-768 (40-Gbps) interfaces before the end of this year.
So far, the only network equipment maker pushing the OC-768 button is Lucent Technologies. Mukesh Chatter, the terabit router wunderkind who moved to Lucent with that company's 1999 purchase of Nexabit Networks, has predicted that Lucent's Bell Labs will be testing 40-Gbps products in the near future. If Lucent is ready to show off its biggest and fastest DWDM gear, can Nortel be far behind?
Cisco seizes the light
Next to predictions that Brad Pitt will break the heart of some Hollywood starlet, about the surest thing going is that Cisco Systems will raise its stake in the optical networking arena this year.
Actually, the acquisition-focused executives at Cisco got a head start in December 1999, when they announced a buyout of the optical systems business from Italian conglomerate Pirelli. Going from basically bupkes in the optical arena in the summer of 1999, Cisco has assembled an end-to-end portfolio with the Pirelli acquisition and the prior purchases of Cerent and Monterey Networks. With these three acquisitions, plus an earlier deal to nab PipeLinks, Cisco is suddenly fine-tuning its optical networking assets, rather than accumulating building blocks.
One of the things to look for this year will be Cisco's strategy to acquire ultra-long-haul DWDM expertise. It could follow Nortel's lead and spend a few billion dollars on a start-up - as in Corvis - or it could put its considerable capital to work accelerating the development of long-haul optical technology acquired in the Pirelli deal.
Photon finish: Not this year
If the eggheads at leading research facilities had their way, a What's In/What's Out technology list for 2000 would feature electrons in the Out column and photons on the In side of the ledger. That scenario isn't likely to happen this year, unless the pace of progress in the search for the all-optical circuit accelerates considerably.
Despite claims by some start-ups of major breakthroughs in developing integrated circuits that pass photons across a transistor instead of much slower, fatter electrons, most experts still think it will take photons another few years to catch up to electrons in terms of network sophistication. A photon-based circuit, which would presumably produce much smaller and more powerful semiconductors than are available today, is no less than the Holy Grail for service providers looking to cut costs.
Considering the attention and money being thrown at optical networking start-ups, the year of the photon may not be as far off as once thought.
Incredible shrinking network
The infrastructures of next-generation carriers and service providers will shrink significantly in the coming year. Breakthroughs in optical networking will enable carriers to construct bigger and faster networks that no longer require electronic gear to provision new services or regenerate laser signals. Carriers such as Qwest and Williams project savings related to reducing operational overhead in the 75 percent range.
Most of the streamlining will come with the commercial introduction of ultra-long-haul DWDM gear and so-called optical switches. While long-haul DWDM equipment will reduce costs associated with boosting laser signals, optical switches will make it possible to reroute wavelengths of traffic from one junction point in the network to another. In the past, this process required Synchronous Optical Network (SONET) equipment, which converts optical signals to electronic form, routes those signals to their proper destinations and then converts them back to optical form.
Although the SONET equipment market will continue to thrive in 2000, look for most carriers to curtail the use of SONET multiplexers and add/drop modules in the core of their networks.
Things could get meshy
Replacing SONET gear in the core of data networks will mean the rise of a new form of network protection. Say goodbye rings, hello mesh.
SONET's ring topology has served the networking industry well.
Countless fiber cuts or equipment failures go virtually unnoticed because of the restorative powers of SONET rings. But with data traffic taking over the lion's share of network bandwidth, SONET rings, which were designed primarily for the voice network, will become a thing of the past.
From SONET's ashes will rise a new architecture for constructing long-haul networks. A mesh approach to network construction, the model supported by optical switches, is based on an abundance of point-to-point connections between major network points. By maintaining three or four alternative routes into a metropolitan area, service providers can ensure continued service if links are interrupted.
The major advantage of a mesh design over SONET rings, experts say, is that a mesh architecture does not force carriers to reserve 50 percent of capacity as a backup mechanism, as with most ring-based architectures.
A few billion points of light
One question nagging at executives at giant networking companies is: Just how high will the valuation of optical networking companies go?
As 1999 came to a close, Nortel shelled out $3.25 billion for Qtera, an optical start-up with great potential, great management and no revenue. Did Nortel get robbed or did it scoop up a bargain? If you hop into the way-back machine for a short trip to 1998, you'd have to surmise that Nortel got taken to the proverbial cleaners and Chief Executive John Roth should be taken to the woodshed. But if you consider Cisco's $7 billion bid to acquire one-product Cerent and the $19 billion valuation of Sycamore Networks after a spectacular initial public offering last fall, Nortel may have pulled off the shrewdest deal since Manhattan went for a bag of shells.
The valuation of optical start-ups should continue rising through the stratosphere in 2000. But don't be surprised to see a little old-fashioned ingenuity figure into the portfolio-building strategies of networking's richest and most aggressive companies. Lucent, for one, last month worked an extensive reseller agreement with Tellium, an up-and-coming optical networking giant. The deal gives Lucent access to Tellium's portfolio without ponying up a few billion dollars to bring the company into the fold.
Squeezed out in the middle
Consolidation will continue to be the defining theme of the network equipment industry. Actually, "consolidation squared" might be a more fitting way to describe the process of growth-through-contraction that will pervade the networking market.
Start with the number of companies that will become divisions of larger companies by year's end. While the ever-increasing flow of start-ups in the optical networking segment will keep the overall count of independents relatively constant, the networking arena is turning into a handful of giants surrounded by newcomers. Rich McGinn, Lucent's CEO, says middle-tier players will find the going tough in an environment best suited for comprehensive Goliaths or quick-moving Davids.
This year also may bring some big changes for veteran data networking companies such as Cabletron Systems and Newbridge Networks. Such stalwarts could end up filling out the enterprise product lines of companies that already have world-class optical networking products in their portfolios. |
