Optical networking strategies of service providers will continue to evolve, reports Communications Industry Researchers

Radical or revolutionary alterations in optical networks of service providers in the U.S. will be hard to find over the next several years. The historic trends of embracing incremental change and adopting interim solutions will continue over that of sudden or dramatic leaps to new technology, such as all-optical crossconnects or OC-768, according to a new 200-page study published by Communications Industry Researchers, Inc. (CIR).

This new report from CIR was drawn from in-depth interviews conducted with 25 U.S. service providers over the past three months by Mark Lutkowitz, CIR's Vice President of Optical Networking Research, who is considered by many to be the most unbiased demand-side analyst in the market. The discussions revealed where these companies were placing their technology dollars, which factors would lead them to purchase from specific vendors, and what they envision for themselves and the industry at large moving forward. And, since the methodology of the report relied solely on the comments offered by those companies who purchase optical networking systems the resulting conclusions were not vendor-biased and the report itself was not vendor-sponsored.

The report, "Service Provider Strategies for Optical Networks," determines that while the telecom service providers will continue to deploy technology at their own pace, the only exception to the gradual shift paradigm lies with the deployment of wavelength intelligence products, especially at Interexchange Carriers (IXCs). Optical switches with electronic fabrics produced by companies, such as Tellium, BrightLink Networks, Cisco Systems, Sycamore Networks and Ciena will be purchased in huge quantities. Ciena is the clear favorite to be the biggest winner in the first leg of the optical switching race, according to CIR, in a market that could easily reach $3 billion over the next three years. CIR also says that it is no exaggeration right now to state that for some carriers to be able to even survive, they will need to deploy these elements into their networks to lower costs, to offer much faster provisioning of services, and to provide more efficient network management.

CIR's report goes on to state that the long-term movement towards switches with optical fabrics will happen in a more step-by-step fashion and mainly by incorporating optical modules into the existing installed base of optical to electrical-to-optical (O-E-O) switches. CIR strongly believes that those companies selling O-E-O gear will not only continue to gain substantial market share over the "pure" optical suppliers, which lack the capabilities service providers are buying today, but furthermore, might very well limit the opportunities for selling to carriers in the future. CIR's research found that network operators favor working with vendors, which already have solutions installed in the network, and are far less inclined to consider suppliers promoting the infamous "forklift upgrade".

With respect to optical switching moving into the metro space, CIR's research did not foresee an expected rush of deployments by local carriers until metro DWDM becomes established.

While the deployment of brand new optical intelligence devices, particularly those with STS-1 granularity, will dramatically increase within the IXC networks, CIR believes that carriers will continue to address the issue of achieving higher transport capacity through greater utilization of existing technology. CIR's outlook for OC-768 is not as rosy and instead sees the carriers increasing the number of DWDM systems and channels in their networks and/or lighting up new fiber in their conduits as the primary ways of getting higher bandwidth in the long haul.

CIR's report supports the notion that it is improbable that the commercial availability of the next-generation TDM system, OC-768, will become real any time soon, regardless of whether or not carriers would find it useful. CIR also found that although there has been deployment of the current OC-192 products since at least 1995, many manufacturers are still struggling with this technology casting doubts by some contacts at the service providers as to the future of OC-768.

Meanwhile, in the case of the ultra-long haul DWDM market, CIR does believe in the technology but also concludes that it will probably represent no more than 15 percent of the total U.S. DWDM market. The nature of the traffic dynamics points to more adding and dropping of traffic in IXC networks between cities that are relatively close together than indicated by the hype many vendors are pushing regarding the new technology. In addition, although ultra-long haul DWDM is attractive in creating a simpler network with much fewer electronic regenerators, the initial cost advantages of the solution are not necessarily apparent to all of the experts we spoke with at the IXCs.

Regarding metro DWDM, the unfilled promises that have existed for close to four years continue to remain in carriers' memories. Many engineers feel that because of the prohibitive costs of the solution, the metro DWDM market is not going to take off in a big way for at least another two years.

In the access space, there will obviously be a heavy concentration on migrating from SONET to IP services including potentially using gear from such manufacturers as Mayan Networks, Ocular Networks, and Redback Networks. There will continue to be significant installations of SONET-based systems in this network segment because SONET is still considered the most sophisticated transport layer by carriers, and many planners expect the cost of such devices to continue to decline.

It is also important to note that the installed base of low-cost, coarse enterprise WDM solutions, such as from ADVA Optical Networking, will keep on climbing upward for delivery of data services.

CIR's report also points out that there will be some impressive penetration of the novel and much lower cost Ethernet over glass types of architecture in the access/metro portions of greenfield networks, which are not encumbered with legacy gear. Suppliers of these products include Extreme Networks, Atrica, and World Wide Packets. While an order of magnitude drop in cost in constructing an Ethernet network might look attractive, achieving suitable margins might turn out to be a different matter. Some of the service providers might wind up underestimating the level of penetration into cities and buildings that is necessary to make such a business sustainable.

What Does It Take To Win? Manufacturers of optical networking gear need to realize that it is no longer adequate to state that their core competency is just in transmission. Those suppliers that lack a sufficient number of engineers, probably in the double-digit range, with the relevant expertise, such as in MPLS, to design those products with a control plane capability will be at a significant disadvantage. Service providers will increasingly want to extend their visibility across their network end to end as well as to synchronize all of the functions and components of the optical network.

Regarding the optical access vendors, the best advice from CIR is to remember that the carriers are looking for low cost and simplicity. It is better to be more of a clean play - target and solve one particular problem. Carriers tend to want best of breed in their access architectures and they are not going to be inclined to believe that a box that supposedly accomplishes way too many tasks is going to be superior in the particular function needed in a specific network.

CIR believes that SERVICE PROVIDERISTRATEGIES FOR OPTICALANETWORKS will be an important tool for both determining and assessing the optical networking strategies of large as well as innovative carriers in both the long haul and metro/access spaces. The report is also a must read for the both the hardware suppliers looking to sell into the carrier market and the companies who supply their components. In addition, the report also provides an historical framework in order to explain present trends in specific networks as well as give an important perspective on expectations for future market dynamics in the optical space.

The report is available from CIR. For more information, visit www.cir-inc.com.