Cable's next big network move: Expanding the upstream

While cable's migration to "10G" and symmetrical 10 Gbit/s speeds are years out on the horizon, operators are expected to accelerate plans this year to take aim at a more pressing matter – expanding and enhancing the upstream capabilities of their widely deployed hybrid fiber/coax (HFC) networks.

Cable's beefy downstream has held up and been expanded in recent years to handle a surge in video streaming, but cable's skinny upstream pipe has long been a weak link of the HFC network. Even as the ceiling of HFC networks have been raised to as high as 1.2GHz to accommodate more downstream traffic, upstream traffic has continued to be constrained to a small slice of spectrum – typically in the range of 5MHz-42MHz in North America and 5MHz-65MHz on many HFC networks Europe.

Tom Cloonan, CommScope
Cloonan says upstream data demands have climbed to the point that cable operators will soon be forced to enhance or upgrade this critical part of the HFC network.
Cloonan says upstream data demands have climbed to the point that cable operators will soon be forced to enhance or upgrade this critical part of the HFC network.

Data demands on the upstream aren't exploding off the chart, but they are growing at a high enough rate that cable operators will soon need to do something about it.

Aided by the accelerated use of connected cameras, upstream data demand grew at a rate of about 21% in the past year, up from the teens a few years ago, according to Tom Cloonan, chief technology officer, broadband networks, at CommScope, which acquired Arris nearly a year ago.

"Even though it's a moderate [growth rate], that does compound," Cloonan said.

That means cable operators are rapidly approaching a decision point with respect to the upstream and how they can enable speeds that extend well beyond today's marketed maximum upstream speeds in the range of 35Mbit/s to 50Mbit/s.

To deal with that pressure, some operators may try to improve the data efficiency with their existing blocks of upstream bandwidth or start to undertake more advanced and longer-term projects that will expand the amount of spectrum dedicated to the upstream.

"They [the cable operators] will have to make investments to be able to support that upstream spectrum growth," Cloonan said, noting that many operators will need to move the upstream beyond 42MHz or 65MHz within the next three to four years.

But even before cable operators make plant-touching moves to expand the upstream, Cloonan notes that there's work underway to help stretch out the lifespan of legacy upstream spectrum for perhaps as much as a year. This could be done by, for example, expanding the use of orthogonal frequency division multiplexing (OFDM) subcarriers that deliver more bits per hertz than legacy single-carrier QAM channels.

That's a "latent technology" already present in DOCSIS 3.1 upstream cards and cable mode termination systems (CMTSs) and there are some proprietary initiatives also underway, Cloonan said. "It buys them extra time and takes some of the stress out of making the transition."

Doing the splits
But the bigger move is upstream expansion, and operators have several options to pursue using DOCSIS 3.1 or emerging DOCSIS 4.0 technologies.

While a so-called "mid-split" would raise the upstream to 85MHz, cable ops will also be looking at even more extreme moves such as a "high-split" to 204MHz, or an "ultra high-split" that could elevate the upstream to one of several levels, including 300MHz, 396MHz 492MHz and 684MHz. If the upstream split is raised, it could likewise cause MSOs to extend the downstream to 1.8GHz or more (and perhaps on up to 3GHz) in the years ahead.

Cloonan expects the upstream augmentation options ultimately pursued will vary by operator and by the level of competition they are facing in given markets. A 204MHz high-split, for example, would give an operator about 1.4 Gbit/s of capacity – enough to build a super-speedy symmetrical broadband service.

Some welcome news for vendors
If cable operators do start to develop and deploy upstream-enhancing, it would of course serve as welcome news to CommScope and other cable access network vendors that have seen cable operator spending slow down as MSOs continue to chew through already-purchased capacity and as they mull their next-gen network moves.

Jeff Heynen, research director for the broadband access and home networking at Dell'Oro Group, believes operators are poised to spend more heavily on upstream initiatives, noting that these are near the top of many operator strategic plans for 2020.

Despite a challenging 2019 and 2020 showing only signs of modest spending improvements, "cable operators are going to invest significantly in upgrading their upstream bandwidth through mid- and high-split architectures, with the goal of shoring up one of the biggest drawbacks of their DOCSIS networks," Heynen noted in this blog post about the trend. "With a growing number of telcos and ISPs now offering symmetric 1Gbps services, cable operators are facing increasing pressure to expand their upstream capacity."

Heynen added that some "large operators" have already begun a move to a mid-split, while others are jumping directly to a high-split.

CommScope's Cloonan, meanwhile, expects early adopters to move ahead with mid-splits or high-splits within the next couple of years, and likely target those network upgrades to markets that are in the most desperate need of upstream capacity to handle actual customer demand or to deal with pressures from fiber-to-the-premises (FTTP) competition that's already delivering symmetrical 1-Gig service.

Heynen and Cloonan also acknowledge that any split change does require changes to the outside plant, including taps and amps, and that these hardware transitions will also be done in conjunction with new virtualized cable access network technologies that enable operators to more rapidly make changes and perform upgrades via software.

Low-latency DOCSIS also heating up
Expanding or enhancing the upstream is just one area where cable operators are ramping up HFC network activity. Cloonan also expects MSOs to be more aggressive with low-latency technologies built into both DOCSIS 3.1 and the coming D4.0 specs that can boost the performance of latency-sensitive apps and services like multiplayer gaming.

While latencies of 20 to 30 milliseconds is a range typically seen during normal busy hours and activity levels, these new capabilities are poised to help drop that average into the neighborhood of 5 to 10 milliseconds and put the industry on a path toward latencies of 1 millisecond.

Low-latency DOCSIS "is probably going to be a hot topic" in 2020 and some operators will likely start to turn on those capabilities within the next year, Cloonan predicted.

Related posts:

— Jeff Baumgartner, Senior Editor, Light Reading

HardwareSoftware 3/18/2020 | 7:48:14 AM
Upstream spectrum Of all the nonsense I have read on this forum, this one beats it! Mr. Noolan seems to forget that there are still legacy set-top boxes that still dwell in that midband spectrum he alluded to. he not only flew over the legacy video bands, but also the out of band frequencies that these legacy set-top boxes need for addressability. Mr. Noolan seems to forget that if you increase the maximum frequency of 800 MHz that is the average of most amps highest frequency that most MSOs deploy, then the spacing between node, and line extenders will have to be decrease to mitigate span loss that would result at higher frequencies.Bringing amps closer to do this may not be possible in some cases, and may entail putting additiona amp in. That would defeat the goal reducing line actives with the ultimate aim of N+0. Mr.Noolan did not take into account that ingress becomes a real issue with mid-split, and even worse with high-split given the spectrum that has been given cell phone companies, especially in the 700 MHz band. I think the days of HFC is coming to an end, and cablecos are delaying the inevitable. Once Verizon and other telcos who are deploying FTTH get their act together, the cablecos will be forced to abandon HFC.
Sign In