Telco, MSO networks adopt similar long-haul strategies

by Meghan Fuller Hanna

Overview
In the past, cable operators relied on third-party backbone providers to transport traffic between their regional networks. But today's cable operators are finding that it is less expensive to build and deploy their own long-haul fiber networks. As a result, their backbone networks are beginning to mirror the telcos'.


Today's telcos and MSOs share a common set of values and objectives. Reliability is obviously a key factor, as is time to market, simplified operations, and a lower overall operating cost. But they are also embracing a similar set of network technologies. As a result, many of the MSOs have begun to wrestle with the same decisions as their telco competitors regarding their long-haul networks, including the migration to 40 and 100 Gbps and the need to deploy reconfigurable optical add/drop multiplexers (ROADMs) not only in the metro but also in the long haul.

While telco and MSO networks may look similar now, it wasn't always that way. The cable industry initially offered just one service�broadcast video�and its networks were architected solely to support that service. They were uni-directional and almost universally fed via satellite. As such, there was little requirement for fiber.

In the mid-1990s, when cable operators began to offer digital video services, they also saw the need to consolidate their headends. At that time, it was not uncommon for each cable operator to have at least one headend, if not several, in a given metro area. All of those resources�the satellite receivers, modulators, etc.�were therefore duplicated. From an economic perspective, it made more sense to have one primary and one backup headend to support a much larger geographic area, and then connect communities in that area to the headend via a regional fiber network.

And because most cable operators are now major ISPs, thanks to their cable modem service, they need access to long-haul fiber. Traditionally, they had been handing off that traffic to a third-party backbone provider, but such a model has proven to be expensive.

Over the last few years, reports Gaylord Hart, director of MSO market segments at Infinera (www.infinera.com), “as MSOs have evaluated the â��build versus lease' model, more and more have come out on the side that building their own networks makes sense. For most cable operators, that is a combination of building some of their own fiber and then leasing dark fiber from other vendors, but then owning and operating the equipment that is attached to that fiber.”

Hart sites Cox Communications (www.cox.com) as the perfect example; the operator recently finished a 12,000-mi national network build that integrates all of its regional networks as well. And in this case, says Hart, Cox “was able to justify the cost to build this network on cable modem data services alone. So the ability to carry other services, like digital video, video on demand, and voice services is, in a sense, an extra benefit to [building their own networks].”

So while telco and MSO networks may have been widely dissimilar in the past, they are now migrating toward the same basic architecture to deliver voice, video, and data services. And one of the critical elements of both is the ROADM.

The MSOs were the first and earliest adopters of ROADM technology�in the metro. But according to Randy Eisenach, market development director at Fujitsu Network Communications (www.fujitsu.com/us/), the cable operators now see a need for reconfigurability in the long-haul network too.

“Based on some of the contracts that we've seen on these long-haul networks, they are moving the ROADM function into their long-haul nodes,” he confirms. “Previously, where they had back-to-back, point-to-point long-haul nodes, now they want full OADM and single-channel granularity at their intermediate points.”

Moreover, adds Mitch Auster, senior product marketing director at Ciena (www.ciena.com), the MSOs are beginning to look at the next-generation ROADM for these applications. Beginning last year, he says, “a large portionâ��I won't say the majorityâ��but a large portion have been of the colorless variety that can drop any wavelength at any port.”

Like the telcos, the MSOs are also exploring the use of directionless or direction-independent ROADMs to provide a more resilient underpinning to their core routers and core optical switches, says Auster. However, such devices are not yet commercially available.

Colorless, directionless ROADMs are more applicable in a meshed environment versus a ring-based metro network, and here again Auster says the MSOs and telcos are showing a similar preference for an emerging trend: meshed networking. Telco carriers like Verizon Business are deploying meshed topologies to provide increased resiliency and shared protection capacity for their high-speed links.

“We are now starting to get similar interest from some of the larger MSOs,” he confirms. “It's not quite there yet, but we do imagine that within a couple of years, they too will want a meshed optical infrastructure below their service infrastructure.”

Many cable operators still rely on 10-Gigabit Ethernet as their core transport stream, but like the telcos, they are migrating to 40 Gbps and have their sights on 100 Gbps as well.

Fujitsu's Eisenach reports that today's cable operators are interested in 40 Gbps for two primary applications. The initial application materialized when cable companies began to deploy core routers that happened to feature 40-Gbps interfaces. If they deployed two or three routers in a given metro area, for example, they needed to interconnect those, but, says Eisenach, “you're talking a relatively short distance over a metro locationâ��and a client-side to client-side connection.”

In the backbone or long-haul network, the driver for 40 Gbps is related to capacity per fiber. Because many of the cable operators are leasing dark fiberâ��a relatively expensive exerciseâ��they only like to lease one fiber pair. In most cases, it is actually less expensive to upgrade to 40 Gbps versus leasing an additional fiber pair. “When it comes to this choice,” says Eisenach, “this is where they are looking at 40G for longer haul distances.”

By all accounts, Comcast (www.comcast.com) is leading the 100-Gbps charge among the cable MSOs, having trialed Nortel's (www.nortel.com) version of the technology last March. The live trial demonstrated the viability of transporting a 100-Gbps wavelength over the same network that is currently carrying both 10- and 40-Gbps links.

Sources interviewed for this story report that, like their telco counterparts, other cable operators are also kicking the proverbial tires on 100 Gbps. “The real growth in terms of bandwidthâ��and this gets into the need for 40 and 100 gigâ��is with packet services, and [the telcos and MSOs] are pretty much on equal footing there,” maintains Auster.

Meghan Fuller Hanna is senior editor at Lightwave.

Lightwave: MSOs' metro, core test needs evolve
http://lw.pennnet.com/display_article/347352/13/ARTCL/none/none/1/MSOs'-metro,-core-test-needs-evolve/

Lightwave: Comcast tries 100G
http://lw.pennnet.com/display_article/323040/13/ARCHI/none/NNEWS/1/Comcast-tries-100G/

NPRG: Conversations with Communications Innovators: Mike Rouleau, Senior Vice President Business Development and Strategy at tw telecom
http://nprg.com/RESOURCECENTER/NPRGsPodcastSeries/tabid/209/ctl/ViewContent/mid/1606/ArticleID/536/Default.aspx

More in Network Design