What If Cable Does the Full Duplex?
Cable's Full Duplex DOCSIS 3.1 technology may still be in the innovation stage, but if it eventually morphs into an industry specification, the practical effect on cable infrastructure will be dramatic.
Most notably, Full Duplex -- if it works as planned -- will offer cable operators a way to deliver symmetrical 10-Gig speeds over legacy hybrid fiber-coaxial (HFC) networks. Since upstream throughput has always been cable's Achilles heel, an upgrade to the 10-Gig threshold would be a major competitive breakthrough. In contrast, DOCSIS 3.1 today promises up to 10 Gbit/s in the downstream, but only up to 2 Gbit/s in the upstream. (See CableLabs Makes Symmetrical Multi-Gig Push .)
Based on advances in wireless technology, Full Duplex works by allowing downstream and upstream data signals to occupy the same frequency ranges. CableLabs is cautious about sharing too much detail about the Full Duplex roadmap because the standards organization doesn't know yet quite where its exploration of the technology will lead. But CableLabs Director of Network Technologies Belal Hamzeh is willing to explain the idea at a conceptual level.
Describing the Full Duplex process, Hamzeh says, "If the device is transmitting or talking on the upstream and it's receiving at the same time at the same frequency on the downstream, and the device knows what it's transmitting, it can subtract that from the signal it is receiving."
In other words, as long as the sending device can filter out the signal it's transmitting, the device can figure out what message is coming through at the same time in the downstream.
As a reference point, Hamzeh points out that old-style PSTN voice networks are also Full Duplex communications systems. People can both talk and listen over phone lines at the same time. However, PSTN systems have a distinct advantage over data-driven networks -- namely that the human brain is a very sophisticated processor, able to filter incoming and outgoing signals automatically. Machines typically have a harder time with the same task, although some are sophisticated enough now to make the attempt.
Hamzeh is also willing to speculate on how long it could take for DOCSIS 3.1 to get a Full Duplex upgrade.
"The way we're looking at Full Duplex is it's an extension of DOCSIS 3.1… so it will not take the same time span required for DOCSIS 3.1. It'll be a much shorter time frame," says Hamzeh.
Since DOCSIS 3.1 took roughly 18 months to develop, that would mean Full Duplex delivery might only take a year or so to work out once it entered the official research and development stage. Hamzeh says that the current innovation phase is likely to last three to six months. Thus, if Full Duplex progresses to R&D from there, then the industry could be looking at access to the technology by late 2017 or early 2018. Other senior cable technologists, however, estimate that it could take till the end of the decade and beyond for Full Duplex to materialize.
As an important addendum to that timeline, cable operators might do well to consider Full Duplex as they evaluate whether or not to extend the upstream frequency range on their cable plant to 85MHz and beyond. The so-called mid-split has begun to look more appealing under the DOCSIS 3.1 regime, but it may not be necessary if upstream and downstream signals can actually share the same frequency ranges. If the optimistic timeline holds, a year and a half may not seem too long to wait to make a decision on whether to do a costly mid-split or forego the option altogether in favor of a Full Duplex implementation.
— Mari Silbey, Senior Editor, Cable/Video, Light Reading