Vendors plot varied ROADM ­migration paths

Th 211919

Though the largest incumbent carriers, Verizon and AT&T, are bullish on the use of multidegree reconfigurable optical add/drop multiplexers (ROADMs) in the near term, they aren’t the only customers in the market for ROADMs. Today, ROADMs almost always are deployed in single-ring configurations by small carriers and enterprise customers, for whom wavelength-selective switch (WSS)-based multidegree ROADMs may not prove cost-effective for certain applications. As such, the ROADM system vendors have adopted migration strategies to satisfy the disparate needs of their customer base. While most agree that WSS-based multidegree ROADMs will be required in the future, the vendors are taking different paths to get there. In some cases, they are relying on tried and true electrical switching to get the job done today.

Meriton Networks ( today supports multidegree ROADM functionality in its 6400 Optical Transport Platform (OTP). The vendor does have plans to incorporate the technology in its 7200 Optical Switch Platform (OSP) as well, but Meriton says its customers continue to weigh the operational costs associated with multidegree ROADMs as well as the additional planning and operational complexity the functionality introduces versus a more traditional approach.

Planning a ring network is straightforward because the number of nodes, the distances between nodes, and the type of fiber are all known, explains Meriton’s chief operations officer Bill Gartner. “As soon you start to say, ‘Okay, any of these nodes could become multidegree and could connect to another ring, for instance, or to several rings, then you have to answer the question, ‘What is the implication on the engineering rules for a service traveling from one domain or ring to another domain?’ What if you want to extend that out to another domain on the far side of the ring you just added? At some point,” says Gartner, “you find that you just can’t get the signal from point A to point Z.”

Meriton has done some network modeling with its customers and has discovered that four-degree ROADM functionality covers the vast majority of applications-upwards of 90%, says Gartner. “The question then is, ‘Do you go beyond four to six or eight?’” he muses. “The technology allows you to do that, but I think the ultimate limitation will be how comfortable the operations and planning folks get in terms of the constraints that might be introduced as you increase the number of degrees and all-optical multidegree configurations without doing OEO.”

Moreover, he says, as the price of 10-Gbit SFP and XFP transceivers has decreased, the argument for all optical becomes less compelling. There is a tradeoff between operational (opex) and capital expenditures (capex), and the opex costs have not been fully appreciated. Thanks to the ever-declining price of 2.5- and 10-Gbit/sec SFPs and XFPs, “you’re not talking about huge costs for 10-Gigabit handoff,” notes Gartner. “The entire argument for multidegree is to eliminate that handoff cost.”

There is an alternative for carriers who would like to introduce multidegree functionality without the additional engineering constraints: electrical switching. Meriton’s 7200 OSP, recently deployed by Tier 1 carrier Korea Telecom, performs an OEO conversion at each multidegree point; thus, the carrier only has to worry about the optical layer within any one ring or domain. The 7200 supports up to 128 degrees, adds Mike Pascoe, Meriton’s president and chief executive officer, who notes that the 7200 eliminates that need for back-to-back ADMs. “You’re really switching through much like you would with an optical switch,” he says. “They key is if the operational cost associated with trying to do multidegree ROADMs is fully weighed in, the [carrier’s] decision may take a different path.”

Though it too has WSS-based multidegree ROADM plans in the works, Ciena ( also touts the economic benefits of electrical switching or grooming. The vendor has adopted a hybrid electrical/optical approach, dubbed dynamic wavelength routing (DWR), to provide carriers with a cost-effective migration path from electrical switching to optical switching. An all-optical system results in a high initial cost, notes Vinay Rathore, director of service provider marketing at Ciena. While the cost of optical switching would decrease as traffic increased over time, Ciena’s DWR approach, by contrast, enables less expensive electrical grooming from day one. As capacity increases and electrical grooming becomes less efficient, all-optical switching or grooming will prove in.

Moreover, Ciena says its hybrid electrical/optical approach supports a flexible assignment of low- and high-granularity bandwidth. “The way we groom electrically, we treat every service as if it’s an actual wavelength, down to the granularity of 155 Mbits/sec,” explains Rathore. “Even a 100-Mbit Ethernet would be stuck into a wavelength of 155 Mbits/sec. That lets you treat it as a wavelength service, and then you can route that wavelength anywhere you want. Because you have the grooming at the electrical level and the optical level, you can basically treat a wavelength from 155 Mbits all the way up to 10 Gbits of capacity on a wavelength and route it from any point to any point on the network,” he adds.Th 211919

Cisco Systems has developed two-degree ROADM technology-in the form of a mesh multiring upgrade card for the ONS 15454 platform-that proves in from day one and provides an easy migration path to the more expensive optoelectronics of the higher-degree ROADMs

Ciena already has two-degree ROADM functionality embedded in its CoreStream products and is working on a next-generation WSS-enabled ROADM that will support up to six degrees. However, the vendor is leery about getting too far ahead of the market, admits Rathore. “If you look at the cost of components for wavelength-selective switching and the limited availability of them, the last thing you want to do is start deploying thousands of units when you’re not even sure if manufacturing can keep up with deploying it everywhere,” he says. Moreover, he adds, the market may not want to pay a premium for multidegree ROADMs in locations that may require only two degrees-and those two degrees not for some time. Any ROADM strategy “has to be somewhat practical,” he contends.

For Cisco Systems (, which released a two-degree planar lightwave circuit (PLC)-based ROADM in 2004, pragmatism has been the key component of its ROADM strategy. According to Russ Esmacher, product manager of multiservice provisioning platforms (MSPPs), Cisco’s decision to begin with a two-degree ROADM primarily was based on the availability and manufacturability of components at that time. Cisco also felt its carrier customers needed to learn about the operational aspects of the technology before volume deployment. “This is a totally new thing for them,” notes Esmacher. “They just learned to operationalize SONET. As they bite off something new, we have to be very measured in our step.”

Cisco has a program in place for a multidegree ROADM but, says Esmacher, “You just can’t go from zero degree to eight degree in one fell swoop.” Instead, Cisco has developed a migration path to help carriers adjust to the operational challenges of the technology while also deferring the cost of expensive optics until they need it for a given site. “We looked at ‘How do we go from a two degree to more than two degrees in the future without fork-lifting all of our two degree?’” recalls Esmacher. “We introduced a mesh multiring upgrade card you can install day one inside a two-degree ROADM location. In the future, you could say, ‘Okay, I’m ready to take this site to three degrees or five degrees.’ Then you put in the complex, multidegree technology, but you do it in service and you utilize all your two-degree equipment.”

A graceful migration path from a two-degree ROADM to a multidegree ROADM also is more palatable to carriers for whom “reconfigurability is already on the union payroll,” notes Greg Smith, senior marketing manager of optical networking at Cisco. “Do they spend the extra money for the capex for the ROADM, or do they literally say, ‘It’s cheaper for me to go send a guy out in a truck and reconfigure the network in the middle of the night?’” muses Smith. “Until the technology matured and it was cheap enough to go the capex direction, that guy in the truck was actually a cheaper solution for a lot of them.”

While its competitors are adopting OEO strategies or more conservative all-optical migration paths from two-degree to multidegree ROADMs, Tellabs ( already has announced the availability of WSS-enabled four- and eight-degree ROADM functionality in its 7100 Optical Transport System. Group marketing manager Mike O’Malley scoffs at the idea that multidegree ROADMs place additional burdens on network planners.

“There is operational complexity when moving to ROADM,” he admits, “but it doesn’t necessarily have to do with the number of degrees, two versus four versus six versus eight. Once you’ve made that conceptual step to say, ‘I’m going to operationalize ROADM,’ there isn’t any additional effort necessary to take you to multiple degrees. We thought it was important to offer the customer the maximum amount of flexibility in terms of the number of rings they would then interconnect.”

That said, Tellabs appreciates that not every carrier will require an eight- or even a four-degree ROADM from immediately. Its four-degree module, for example, can be configured in two, three, or four degrees, enabling the customer to select the number of degrees most appropriate for each node or application.

Tellabs has seen interest from its customers across the board, says O’Malley, noting that incumbent carriers planning large-scale video deployments tend toward optimism with the technology. “What you’ll see is that those carriers who are bullish on video and bullish on IPTV are also then very bullish on multidegree ROADMs,” he says. “When you add ROADM to a DWDM system, it gives you the flexibility you need to provide high-bandwidth transport across an entire network and switch that video end-to-end wherever it’s needed.”

O’Malley also cites enterprise-based high-bandwidth wavelength services as a driver for multidegree ROADMs. Carriers targeting large financial and other Fortune 500 companies are deploying them for their ability to ease the provisioning of high-bandwidth services across the network.

Overall, carriers appreciate the value that reconfigurability brings to a DWDM platform, says O’Malley. “There’s both a capex and an opex savings associated with taking four discrete network elements and converging those all into a single platform. That’s what really has our customers excited about our implementation of ROADM,” he contends.

The vendors interviewed for this story all agree that four-, six-, and eight-degree ROADM functionality will be required for at least a portion of the incumbent carriers’ networks and likely some of the Tier 2 and 3 carriers and enterprises as well. That said, they have adopted different strategies and migration paths for introducing this functionality-which begs the question, “Which business strategy will prove the most successful?”

It may be too early to call, says Jason Marcheck, principal analyst of optical infrastructure at Current Analysis ( “With an eight-degree ROADM, you’re talking about a complex piece of equipment that is going to be pretty high in cost,” he says. “Some vendors might tend not to focus on real high capacity because a) there’s not really much need for it right now, and b) it’s very expensive. They don’t want to be positioning a product that essentially is going to price itself out of the market when there’s really not a need to necessarily talk that way.’”

There are others in the multidegree ROADM market, including Fujitsu (, which has added WSS-enabled eight-degree reconfigurability to its FLASHWAVE 7500, and NEC America (, which recently announced the addition of multidegree support for its WSS-enabled SpectralWave 4200.

Siemens (, the current market share leader in long-haul ROADMs, also has a WSS-based ROADM in the works but, like Cisco, is reticent to announce product before the market is ready. The capital expense of wavelength-selective switching is still too high for most vendors, reports Hans-Juergen Schmidtke, vice president of product management in Siemens’ Access and Transport Division. He too cites the operational aspects of multidegree ROADMs as an additional hurdle to their deployment, believing that over the next year, ROADM deployments will continue to be blocker- and PLC-based. “But wavelength-selective switching will get there,” he says.

At the end of the day, says Marcheck, “The way [vendors] are able to use technology to drive down costs and simultaneously create a solution that’s robust and it works, that’s going to be a big point of differentiation going forward.”

More in Network Design