COLOGNE -- ANGA COM 2019 -- Forget all the fuss about cable's 10G initiative, even though the industry hasn't even begun to roll out 10-Gig broadband speeds yet. Some top cable technologists are already shifting their focus to 25G, trying to figure out how to make symmetrical 25-Gig speeds work over a highly optimized HFC plant.
Speaking at this large annual German cable conference here earlier this week, two leading cable technologists said they are now looking at how to empower the next next-gen version of DOCSIS, even as the current next-gen version, Full Duplex DOCSIS, prepares to take its place on the runway with field trials early next year. The two engineers said their research teams are studying various ways in the lab to boost the capacity and efficiency of the HFC plant to support speeds that would be two-and-a-half times faster than the new FDX spec aims to enable.
"We're finishing up work on 10G." said Robert Ferreira, GM of Intel's Strategy, Planning and Technology Office, referring to the company's silicon lab team. "We're pretty confident in all the underlying technology … Most of the work in my lab is now moving on to 25G."
John Chapman, CTO of cable access and a fellow at Cisco Systems, agreed that the industry's lab work on 10G is just about done and that the technology behind it is solid. With his team's focus also now shifting to 25G, Chapman is co-authoring a technical paper on the subject with Ferreira for presentation at the SCTE Cable-tec Expo show in New Orleans this fall.
"We think we can take the HFC plant up to 3GHz," Chapman said, as long as the silicon chips for network equipment can support that. "So there's a long road map for HFC ahead."
While there are several technical paths that cable operators can take to reach the current 10G goal -- depending upon whether they add more fiber, adopt a distributed access architecture (DAA) approach, expand their RF spectrum and/or carve up that spectrum differently -- the options narrow for reaching 25G. Both Ferreira and Chapman agree that operators will need to expand their RF spectrum capacity substantially to make that next great leap, raising the bar from today's 1.2GHz ceiling to at least 1.8GHz and probably all the way up to 3GHz.
With cable vendors just starting to churn out taps and other network equipment to support an upgrade to 1.8GHz, that may seem a tall order. But the two technologists argue that making the big jump to 3GHz network gear shouldn't be that big a deal.
"There's no fundamental reason why 3GHz taps should cost more than 1.8GHz taps," Ferreira said. "HFC has been the gift that keeps on giving. Not only can we get it to 10G, we can get it beyond."
Chapman, one of the fathers of the DOCSIS family of specs, outlined plans to use the current 1.2GHz of available plant spectrum to support current cable operators and the 1.8GHz to 3GHz portion to support upgrades. He also noted that his research team is looking at ways to use more powerful amplifiers to support FDX-type bandwidth and speeds even without extending fiber all the way to the node, as the new spec now requires. He said the big technical challenge there is the additional power that would be needed to support those amplifiers. (We will have more on that in an upcoming story.)
In the meantime, cable technologists are working on how to distribute gigabit speeds throughout the home as well by deploying technologies like WiFi 6. The goal, Chapman said, is to offer customers a seamless IP-centric network that can support gigabit speeds over fiber, coax and WiFi.
- CableLabs CEO: 10G a 'collection of technologies'
- Full-Duplex DOCSIS Not One-Size-Fits-All
- Cisco's John Chapman: DOCSIS Will Never Die
- Arris CTO: What's next for DOCSIS?
- Here Comes DOCSIS 4.0
— Alan Breznick, Cable/Video Practice Leader, Light Reading