Though fiber technologies and deployments continue to capture the most attention when it comes to broadband, two technology demonstrations at CableLabs 10 G Showcase provided clear evidence that DOCSIS 4.0 will keep cable operators competitive with their current outside plants. Furthermore, the advancements shown during each of the demonstrations could potentially accelerate the deployment of DOCSIS 4.0 technologies across a wider network footprint than previously thought. Assuming supply chain issues are resolved later this or early next year and lead times for next-generation silicon improve, DOCSIS 4.0 lab testing could move beyond the prototype stage earlier than expected.
Charter and Comcast both took to the stage during the CableLabs event to demonstrate their respective approaches to DOCSIS 4.0. Charter is the biggest proponent of the extended spectrum (ESD) variant of DOCSIS 4.0, whereby the outside plant spectrum is increased to 1.8GHz from today’s 850 MHz to 1 GHz, while Comcast is behind the full-duplex (FDX) variant, which uses 1.2 GHz of the spectrum but allows for the shared use of that spectrum for both downstream and upstream bandwidth.
The deployment of both technologies is predicated on the rollout of distributed access architectures, either remote PHY (R-PHY) or remote-MACPHY (R-MACPHY). Those products are available today, with Comcast already having deployed over 30k R-PHY devices (RPDs) in their network. Charter favors remote MACPHY and demonstrated an R-MACPHY node from Vecima Networks using first-generation silicon from Broadcom, which gives them the ability to provide service to a single service group today, growing to 2 service groups with the second generation chipset that should be ready later this year.
Regardless of the difference in technological and architectural approaches between the two major operators, both demonstrations focused on how the dramatic speed improvements expected with DOCSIS 4.0 could be realized without significant impact on the existing outside plant. In both cases—whether ESD or FDX—operators are going to need to upgrade their outside plant, specifically taps and amplifiers. There is no getting around that. However, if operators can make changes to the individual outside plant elements without having to disrupt the location of those elements and the power required to drive them, then that is a win-win all around.
First, Charter demonstrated speeds of nearly 9 Gbps down and 6 Gbps up using a cascade of 4 amplifiers. The amplifiers had been upgraded with 1.8 GHz modules from Teleste that can be installed in the existing amplifier stations without having to change any of the spacing between those stations. Charter also demonstrated signal levels and MERs (Modulation Error Ratios) all the way to a 2 GHz tap from ATX Networks and a prototype cable modem that are within similar ranges of today’s DOCSIS 3.1 networks. These results are critical in demonstrating that Charter could conceivably achieve near-10Gig speeds without having to go beyond the network changes they are planning to make. Those changes include simple faceplate changes to the taps (which don’t disrupt service) and module swaps at each existing amplifier station.
Eliminating the requirement to pull fiber deeper and alter their existing amplifier cascades further means Charter will reduce its capex spend and increase the speed with which it rolls out DOCSIS 4.0 to subscribers. It also means the operator’s current efforts to move to high-split DOCSIS 3.1 architectures will not be a stranded investment. Instead, those upgrades will simply be a stepping stone to what was shown in their demonstration.
Finally, another interesting element of Charter’s demo was its use of a GAP (Generic Access Platform) node housing for its R-MACPHY device. Charter has been a major proponent of GAP because it helps Charter solve its problem of having dozens of different node housings from several different vendors in their networks. It also gives Charter a modular platform that can be used to deliver DOCSIS 4.0, FTTH, and wireless services. It also incorporates the ability to add edge compute functions. Although the GAP node in the demonstration did include a compute module from Intel, that module wasn’t used in today’s demonstration. Nevertheless, with CableLabs focused on providing “optionality” for its cable operator constituents, the combination of GAP nodes plus the ability to harness edge compute functions so deep in access networks opens a world of service options for Charter.
Comcast followed Charter and demonstrated multiple access network configurations, all supporting production traffic and all being driven by a vCMTS core located in a Comcast headend not too far from CableLabs headquarters in Louisville, CO. Comcast first demonstrated a DOCSIS 3.1 high-split architecture using an RPD and 1.2GHz amps. From the same node, Comcast also demonstrated a 10G EPON OLT. This particular demonstration has direct applicability today, as Comcast is in the midst of an edge-out strategy, expanding its networks from an existing node base or hub site location. In some cases, these edge-outs will be done with fiber; in other cases, they will use coax. In both scenarios, Comcast can leverage the existing vCMTS core for the MAC layer and subscriber management control of both physical media, making it easier to turn up subscribers and achieve ROI.
Comcast’s other demonstration included prototype amplifiers with built-in echo cancellation, allowing them to deliver FDX in a node plus two environments. Up to this point, any FDX deployments at Comcast were in node plus zero environments, where Comcast had pulled fiber into neighborhoods and eliminated amplifier cascades altogether. For the vast majority of cable operators, node + 0 based FDX just hasn’t been justifiable from a capex perspective. The high cost of deep fiber rollouts is what led to the development of extended spectrum DOCSIS and its ability to be delivered over existing amplifier cascades.
But with Comcast proving that FDX can work in a node + 2 environments, using amplifiers with built-in echo cancellation, suddenly the deployment cost comes down considerably for operators. Is it enough of a cost reduction to garner interest among a large enough contingent of MSOs? That remains to be seen.
Potentially the most meaningful development of today’s showcase wasn’t even part of a technical demonstration. Instead, it came when Comcast’s Elad Nafshi said that the operator’s goal is to prove FDX can work in a node + 6 architecture. Comcast has promised to provide more information about the FDX amplifier concept and its performance and use cases later this year. But if Comcast can deliver an FDX amplifier and product set that delivers FDX across existing amplifier cascades, then, similar to Charter’s demonstration of ESD across existing amplifier cascades, the rollout of DOCSIS 4.0 can occur more quickly and less expensively than previously thought.
The net result of both technology demonstrations is that any concerns around the value of upgrading to DOCSIS 4.0 relative to the cost are now taken off the table. Instead, the question now shifts to one of timing. Operators around the world were firmly in the camp of one technology or the other, with all agreeing that some level of fiber would be a major part of any future network upgrade. Now, operators have two similarly-performing technology options in front of them, each requiring upgrading of existing amplifiers, but with two very large questions around timing and availability of key components.
Nevertheless, after today’s demonstrations, the value and relevance of DOCSIS 4.0 just improved significantly. With theoretical speeds matching those offered by today’s XGS-PON technologies, cable operators can easily buy themselves more time and continue to maximize their DOCSIS networks for many years to come.