DOCSIS 4.0 calculus: Could ESD + FDX = cable’s silver bullet?

Operator decisions around DOCSIS 4.0 are commonly framed as responses to an “or” question – that is, whether an operator will be using the extended spectrum (ESD) variant or full duplex (FDX) variant. But what if the “or” became an “and”? Could that slight tweak silence debate about whether DOCSIS has enough gas left in the tank to fend off fiber advances?

Follow me into the technical weeds if you will – I promise to clear the path as best I can.

Both ESD and FDX are set to deliver 10G capabilities for cable operators, but they go about doing so in very different ways. ESD involves adding more spectrum, jumping from a max of 1.2GHz in DOCSIS 3.1 to 1.8GHz. But it keeps the traditional approach of using dedicated chunks of spectrum for the upstream and downstream traffic flows. The extra spectrum gives operators bigger downstream and upstream pipes, though, and thus more capacity and faster speeds.

FDX, meanwhile, upends the old approach, sticking with a 1.2GHz-sized pipe but using noise cancellation to allow upstream and downstream traffic to be transmitted over the same spectrum.

Comcast is the most notable FDX proponent while Charter Communications and many others are going the ESD route. Some operators, like Mediacom, said it’s possible they’ll eventually use a combination of both versions, deploying the best option for a given market, even if they start with just one or the other.

But in a recent interview, Cox Communications SVP Guy McCormick asked an interesting question: what if you combined ESD and FDX? That is, what if you ran FDX over 1.8GHz of spectrum rather than stopping at 1.2GHz? Wouldn’t that open the door to even more capacity that plain old one-or-the-other DOCSIS 4.0 could?

Each variant of DOCSIS 4.0 is designed to deliver 10 Gbps downstream and 6 Gbps upstream, for an aggregate of roughly 16 gigs of throughput. Would combining them double those numbers?

I took these questions to the folks at CableLabs, who know a thing or two about DOCSIS. And here’s what they told me.

TL:DR

So, could the combination of FDX and ESD take operators to 20G and beyond? In short: No, but there is a way to get there.

Before we continue I have to stress that the following information was shared in the context of a conversation about what is theoretically possible. That is, we're not revealing some new spec that's under development here. Still, the information is valuable for cable companies who may be wondering how in the world they're going to compete with 25G or 50G fiber. But again, at least for now, this is hypothetical. Got it? Good. Here we go.

Doug Jones is a principal architect on the wired team at CableLabs. He told FierceTelecom by email a merged solution is indeed possible. This would operate on 1GHz, 1.2GHz or 1.8GHz taps and give an operator "flexibility on their plant upgrade, the amplifiers, etc."

But it would not increase overall speed. That's in part because FDX in DOCSIS 4.0 only operates on a subset of the aforementioned 1.2GHz of spectrum - or, to be precise, the chunk from 108MHz to 876MHz. The main function of adding FDX to ESD would - again - be flexibility in deployments and apportioning spectrum to the up and downstream channels. 

"The way to increase the speed is to extend the spectrum to 3GHz, which would increase to 25 Gbps aggregate speed," he wrote. "And still add that extra FDX sub band (FDX operates 108 to 876MHz) which does not impact the 25G aggregate speed, rather, this just provides more up / down speed flexibility."

The move to 3GHz would require a new specification - a DOCSIS 4.1 or 5.0, if you will. Thus it seems 25G isn't something that will be enabled with DOCSIS 4.0, even if FDX and ESD are combined.

One more note for those eyeing FDX: CableLabs said that using learnings from DOCSIS 4.0 and advances in silicon technology it is possible that future technology could support wider FDX bands. So, that's a bright spot to end on.