While cable’s DOCSIS 3.1 platform looks to extend the life of HFC networks, an emerging technology called G.fast aims to perform a similar trick for DSL for the telcos as they look for ways to squeeze more speed out of their widely deployed copper-based networks.
While interest in G.fast and its 1-Gig potential reached frothy levels last week at the Broadband World Forum in Amsterdam as chipmakers and network equipment suppliers trotted out their latest wares based on the technology, analysts agreed that the G.fast has the potential to pump more life into DSL, but also questioned whether it will be a big hit, let alone a home run, for the telcos as they look for G.fast to turn copper into broadband gold.
But the potential for G.fast, which brings a new, speedy twist to the twisted pair, is indeed sizable, as telcos would certainly be eager to squeeze more speed out of their DSL lines and deliver services that can more closely compete with cable’s widely-deployed DOCSIS platform before having to take the plunge with pricey fiber-to-the-home upgrades.
But tying G.fast to any 1-Gig speed claims is somewhat of a misnomer, as that represents the aggregate data capacity (upstream plus downstream) it can achieve. And there’s another key caveat – G.fast, which requires a noise-cancellation technology called vectoring, requires very loop lengths (250 meters or less) for the platform to achieve a sizable data boost.
As would be expected, suppliers and chipmakers are bullish on the prospects of G.fast, even before the standard is fully baked. Most of the heavy lifting on the standards front was complete by last December, giving chipmakers enough confidence to move ahead with product designs, knowing that any small ticket changes can be handled in software.
“We still see a ton of value that we can unlock in the copper plant with advanced DSL technologies.” Jim McKeon, Broadcom’s senior director, product marketing, Broadband Carrier Access, said. Broadcom is a big enough believer in the technology that it has developed and introduced G.fast-based silicon for home-side gateway and network-facing equipment (please look further below for a snapshot of last week’s G.fast action). He expects Broadcom’s products to be ready for full production by the first half of 2015.
McKeon said Broadcom’s approach is backwards compatible with VDSL, allowing telcos to perform incremental upgrades. “Introducing G.fast will be a very straightforward exercise,” he said, predicting that there will be significant field trials underway next year alongside some small deployments before rollouts kick into high gear in 2016. “There’s a surprising amount of urgency coming out of the telcos for G.fast.”
And that interest has been global, he said, noting that there are about 400 million lines of DSL installed. “We’re still seeing a lot of potential in wireline broadband. We wouldn’t be investing in it if we didn’t see that,” McKeon said.
G.fast: Big Potential, Big Questions
Analysts who track the broadband access market and are keeping tabs on G.fast aren’t ready to call it a home run. But they understand the value proposition.
“G.fast is going to have legs, but the question is how long are those legs going to be,” Teresa Mastrangelo, founder of marketing analysis and consulting firm Broadbandtrends, said. “The telcos are excited about the potential for G.fast and the speed that it can provide.”
Jeff Heynen, principal analyst, broadband access and pay TV, at Infonetics, expects to issue his first G.fast forecast early next year, but said the total addressable market for the technology won’t be a straightforward exercise.
While telcos that are doing vectoring now, such as KPN of the Netherlands, Swisscom, BT, Belgacom, Deutsche Telekom, A1 Telekom Austria, and even AT&T and CenturyLink Communications, are among the candidates for G.fast, “I don’t think it’s a one-to-one relationship,” he said.
“Will G.fast be massively adopted? I’m still not sure,” Heynen said.
Both analysts also wonder if some telcos will decide that it makes more financial sense to just pull fiber all the way to the premises once they start looking at the short loop lengths required by G.fast.
Mastrangelo said operators will need loop lengths in the range of 25 meters to 40 meters to get the biggest bang out of G.fast, but notes that the initial trials are using more realistic loop lengths that enable speeds in the vicinity of 100 Megabits per second to 200 Mbps.
And she expects that wide deployments of G.fast will require some significant network engineering, which could steer telcos toward more surgical deployments in area where they are feeling the most competitive heat, perhaps look at driving deployment on a demand-based preregistration model, such as the one Google Fiber uses for FTTP.
G.fast’s Big Week
Last week’s Broadband World Forum in Amsterdam produced a wave of G.fast activity. Here’s a snapshot:
-Alcatel-Lucent unveiled a G.fast optical network terminal that will launch in the first quarter of 2015 following 12 trials with operators, including A1 Telekom Austria, BT and Orange.
-Adtran said it has G.fast trials underway with unnamed service providers in Europe, North America and Asia Pacific.
-The Broadband Forum announced its G.fast certification program, and identified the University of New Hampshire InterOperability Laboratory as the first testing lab for that effort. UNH-IOL expects to begin initial G.fast product testing in the first half of 2015, and to announce its first wave of certifications next fall.
-Broadcom added G.fast capabilities to its BCM63138 gateway system-on-chip, and a new family of G.fast-facing chipsets for DSL central offices – the BCM65200 DSP and BCM65900 analog front end.
-Sckipio Technologies demonstrated a “commercial” G.fast chipset (the DP3016-EVM DPU) with built-in vectoring connecting at rates of up to 700 Mbps on 16 ports simultaneously. Sckipio partners Lantiq, VTech, XAVi, and Zinwell were among those that showed off gateways and bridges that used the new silicon.
