"
x
10G

Cable techies see big network changes coming

Seeking to stay competitive as rival fiber and wireless providers furiously build new broadband platforms, cable technologists are working on a vast wave of upgrades to their legacy networks that they hope will carry the day.

Appearing on a Light Reading-hosted forum at the recent virtual SCTE Cable-Tec Expo show, five cable technologists spelled out many of the moves the industry is making to prepare its hybrid fiber/coax (HFC) networks for the next decade and beyond. In particular, they highlighted new technologies and techniques like DOCSIS 4.0, spectrum mid-splits and high-splits, the Distributed Access Architecture (DAA), XGS-PON and network virtualization as critical measures.

In a short keynote preceding the broader panel discussion, Ted Boone, senior director of solutions engineering for Cox Communications, laid out the MSO's vision of an all-IP, 10G architecture in the near-term future. That 10G vision calls for Cox to upgrade its HFC networks to the next-gen DOCSIS 4.0 spec, expand plant spectrum to 1.8GHz, roll out DAA and carry out spectrum mid-splits and high-splits to create both more downstream and upstream capacity.

ESD vs. FDX

Like most cable operators that have made a commitment so far (with the very notable exception of Comcast), Cox is embracing the Extended Spectrum DOCSIS (ESD) version of DOCSIS 4.0, which involves raising cable system plant spectrum from the current 860MHZ or 1.2GHZ ceiling to 1.8GHz. With that additional spectrum available, operators can expand both their upstream and downstream capacity by upgrading all their equipment on the network and in the customer premises.

Boone explained that Cox chose ESD over the other D4.0 option, Full Duplex DOCSIS (FDX), because ESD aligns better with the MSO's current plant architecture. Unlike ESD, FDX keeps the plant spectrum limit at 1.2 GHz, but creates more upstream capacity by enabling both downstream and upstream traffic to run over the same blocks of spectrum, rather than over separate, dedicated blocks, by using a technique called "echo cancellation." But, to make the technology work, operators must currently pull fiber much deeper in their hybrid networks and remove all or most of the signal amplifiers they currently use.

"The big advantage we see with ESD is the ability to operate a traditional node-plus-five (N+5) architecture," he said. "It fundamentally enables a faster, cheaper deployment." Another advantage of ESD, he noted, is that it creates "more downstream capacity" as well as additional upstream capacity.

Kjeld Balmer, head of network technology for Stofa in Denmark, said ESD makes more sense than FDX for his company's cable systems as well because of their legacy architecture, especially since the operator just recently completed a major rebuild of its network.

"We have a traditional N+5 architecture," Balmer said. "It would be quite problematic to go to an N+2 or N+3 [architecture]… So we're looking at a softer ESD [upgrade] now."

Market analyst Jeff Heynen, vice president of broadband access and home networking at Dell'Oro Group, said most other cable operators are leaning toward ESD for similar reasons as Cox and Stofa so far. But, he noted, that inclination could shift given the work that Comcast and its allied vendors are now conducting to develop FDX amplifiers that would obviate the need to pull fiber as deep in the access network.

"If FDX can function in an N+2 or N+3 environment, then that makes it more of an alternative than having to push it [fiber] all the way to the headend," Heynen said.

Remote PHY vs. Remote MACPHY

Just as it is committed to the ESD flavor of D4.0, Cox is committed to the remote PHY variant of DAA, which shifts the PHY layer of the DOCSIS network to a node or shelf at the edge of the network. Boone said Cox has now deployed "north of 12,000 remote PHY nodes" nationwide, covering more than one-third of its network, and plans to keep deploying remote PHY devices (RPDs) for at least the time being. "It [remote PHY] works remarkably well," he said.

Similarly, Balmer said Stofa intends to stick with its early commitment to remote PHY, which it has already deployed throughout much of its regions. When Stofa chose to go the RPD route several years ago, the only remote MACPHY option available came from the embattled Chinese vendor, Huawei.

"We didn't dare to go in that direction," Balmer said. "So we went with R-PHY."

Despite the cable industry's early embrace of R-PHY, though, many technologists expect the balance to shift in favor of remote MACPHY once DOCSIS 4.0 upgrades are carried out. That's because remote MACPHY, which places both the PHY layer and the MAC (Media Access Control) processing at the network edge, offers a simplified data plane and potentially lower latency for customers.

Cox, for instance, is looking to start rolling out remote MACPHY once it begins installing DOCSIS 4.0 active devices on its network in the next couple of years. "We expect to continue the deployment of RPDs until DOCSIS 4.0 actives become broadly available," Boone said. "That's kind of where we see the pivot."

Both Heynen and Randy Levensalor, principal architect at CableLabs, said they see other operators making the same calculation. That's especially true now that CableLabs has issued its Flexible MAC Architecture (FMA) specs, which support DAA with the MAC deployed either centrally in the headend or distributed out at the node.

Rob Wilmoth, chief architect on the North America service provider team at Red Hat, said the decision to go remote PHY or remote MACPHY depends on several factors, including capital costs, state of plant, familiarity with the technologies and vendor relationships. "It typically boils down to a cost-benefit analysis versus performance," he said.

Spectrum mid-splits and high-splits

Recognizing the growing need for more upstream capacity exacerbated by the COVID-19 pandemic, cable operators are increasingly turning to spectrum mid-splits and high-splits to carve out more room for upstream traffic. A high-split upgrade expands the amount of spectrum dedicated to the upstream in North American DOCSIS cable systems from the current range of 5MHz-42MHz to a much greater range of 5MHz-204MHz, while a mid-split expands the amount of spectrum dedicated to the upstream in North American DOCSIS networks to a new range of 5MHz-85MHz.

Boone said Cox, which has been rolling out mid-splits for several years and plans to start high-split trials in January, aims to expand such upgrades broadly over the next four years. With mid-splits now covering "north of 20%" of the operator's nationwide network, he said Cox hopes to expand a combination of mid-splits and high-splits to 85% of its network by the end of 2025.

Balmer said Stofa has already rolled out "a high-split solution" to 85% of its network in Denmark. "We've already expanded our spectrum and [spent] quite a lot of capex," he said.

Heynen said the metrics he tracks, which includes DOCSIS upstream licenses, show that numerous cable operators are moving forward with mid-splits and high-split upgrades, especially in Europe. "It's not just Cox," he said.

Related posts:

— Alan Breznick, Cable/Video Practice Leader, Light Reading

Be the first to post a comment regarding this story.
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE