In for the long haul: there's no lack of challenges��from the core to the customer��for optical equipment vendors - fiber optic communications

There are signs that optical technology is ready to take on challenges beyond simply providing plentiful capacity in long-haul networks. Across the board vendors are talking about the need to drive optical technology beyond the core to economically support new services such as Ethernet while continuing to reduce equipment costs and, more importantly, operational costs. Even so, the starting point for any discussion about optical networks is the price cariers pay to transport bits (see Figure 1), and the buzz today is all about 40-Gbps wavelengths. "We have to continue to drive down the per-bit cost of capacity, and given the overall capacity demands, 40 Gbps is a lower-cost way for carrying capacity than 10 Gbps," says Rodney Alferness, senior vice president of optical networking research at Lucent's Bell Labs.

Although the predictions are that the move to 40-Gbps channels is not going to be nearly as aggressive as was the transition from 2.5 Gbps to 10 Gbps wavelengths, delivering greater capacity over longer distances is an ongoing challenge. Lucent may have set a record here when in March it disclosed it had successfully tested a 2.56-Thps (64 channels at 40-Gbps each) DWDM system over a distance of 4000 km. Simultaneously, the vendor also announced Deutsche Telekom would begin field tests of Lucent's new LambdaXtreme's 40-Gbps capability. Though Nortel is as much a proponent of 40-Gbps technology as Lucent, Paul Haddad, Nortel's director of optical marketing, cautions that cost considerations might nudge carriers to first deploy 40-Gbps channels to move traffic between sites that typically are no more than 800 km apart. Among other things, this viewpoint is based on issues such as CD (chromatic dispersion) and PMD (polarization mode dispersion) that affect the integrity of optical signals at higher rates. At th e same time, Haddad notes that the muxing/demuxing by systems that support lower line rates but high channel counts also involves increased costs.

However, vendors agree that OADMs (optical add/drop multiplexers) that incorporate a flexible add/drop capability are essential for reducing carriers' operational costs. Full flexible add/drop is a step up from OADMs that might limit the number of wavelengths which can be dropped off at a location as well as whether wavelengths can be dropped at any location where an OADM might reside. Flexible OADMs give carriers the ability to remotely configure at what points channels are added/dropped. At the Optical Fiber Conference held earlier this year in Anaheim, Calif., Ciena unveiled the WavelengthDirector, an OADM that can add/drop any or all channels. The company says the device enables dramatic reductions in a network's capital and operation costs because it processes only the wavelength needed at a site, in contrast to back-to-back terminals that process all wavelengths.

While the debate on OEO (optical-electrical-optical) switches vs. OOO (optical-optical-optical) switches isn't about to end anytime soon, there is for the most part a consensus that there's a role for both types of switches. For the moment, OEO switches clearly have an edge because in addition to enabling end-to-end provisioning of services and restoration across a mesh backbone in the event of a fiber cut or equipment failure, these switches include a DCS (digital cross-connect) capability to groom traffic (typically down to the STS-1 level) so as to optimize network bandwidth. They enable a protected path across the network to be shared by multiple services, and this in turn gives carriers the option of offering protected and unprotected services. Additionally, the SONET ADM capability built into OEO switches makes them an alternative to deploying multiple ADMs in COs to terminate multiple metro SONET rings, or forcing carriers to install additional ADMs to terminate newly commissioned SONET rings (see Figu re 2).

By any calculation, the deployment of DWDM systems in the metro hasn't been as aggressive as was anticipated just two years ago. Still, Joe Padgett, Nortel's director of metro optical marketing, believes a turnaround will happen this year. "ILECs have a business proposition to where they now want to offer wavelength services to enterprise customers," he says. In fact there are indications that ILECs are coming around to lighting up the fiber in the metro with DWDM. In October 2001, SBC announced the availability of its MON (Multi-service Optical Networking) solutions for connectivity to mainframes, data centers and storage networks in the metro areas within SBC's footprint. Combined with optical CPE, the MON supports a variety of transport protocols such as ESCON, Fibre Channel and GigE. And Verizon recently selected Lucent's Metropolis platform under a three-year agreement to roll out ring-based DWDM regional inter-office networks. Despite the softening in the CLEC market, ILECs are not alone in wanting to l everage the capacity enabled by metro DWDM networks. One day after Lucent's Verizon contract was announced, start-up Quantum Bridge disclosed that Time Warner Cable was using its optical access solution based on PON (passive optical network) technology to offer business customers in Tampa and St. Petersburg high-speed IP and Ethernet services.

The higher capacities being created by optical networks are driving the demand for next-gen multi-service switches that can efficiently pack wavelengths, says Rob Koslowsky, Cisco's director of marketing for optical services. Cisco's two key products for this are the ONS 15454 (developed by Cerent) and the ONS 15327, a smaller version of the 15454 that Cisco developed for the network edge. Among other things, these boxes, sometimes referred to as next-gen switches, provide SONET transport (they also have integrated DWDM capabilities) for a number of services such as ATM, Ethernet and TDM. Although Cisco has shipped hundreds of ONS 15454s, it was undoubtedly a coup for the vendor when AT&T earlier announced that it had installed more than 100 ONS 15454s in its network and planned to deploy "significantly more" to aggregate lower-speed traffic onto OC-48/192 pipes. While the selection of the ONS 15454 clearly was a disappointment for Ciena, which in the MetroDirector K2 has a competing product, Ciena also came out a winner with AT&T's disclosure that it had deployed Ciena's CoreDirector optical switch in more than 40 cities and planned to install "significantly" more CoreDirectors to route traffic across its optical network, parts of which have been upgraded to 160 10-Gbps channels.

In the current soft market, a common theme from optical equipment vendors is that even more important than reducing the capital costs of building networks is lowering--by at least as much as 10x--the cost of operating those networks. To a large extent this means building more intelligent devices that share a common control plane, so various boxes know where they are and are able to request bandwidth across the network. Sycamore Networks, for example, makes much of the fact that it offers three optical switches (SN 3000, SN 16000 SC and SN 16000) that share a common network operating system. "The challenges today are reducing capex and opex, service differentiation so carriers can offer new services with better margins, interoperability with legacy equipment and the ability to scale as carriers see demand," says Kevin Oye, Sycamore's vice president of systems and technology.

[FIGURE 3 OMITTED]

Fig. 1 North America transport market.

        ($ billions)
Year  Metro  Long Haul

2001   4.8      5.6
2002   3.7      3.6
2003   3.9      3.8

Source: RHK

Note: Table made from bar graph

Sam Masud is senior technology editor at Telecommunications(R) (smasud@telecommagazine.com).

COPYRIGHT 2002 Horizon House Publications, Inc.
COPYRIGHT 2002 Gale Group