The Federal Communications Commission’s sixth annual “Measuring Broadband America” report wastes little time in reminding phone-based digital-subscriber-line (DSL) technology subscribers that they’re stuck in the slow lane. Still.
“Average DSL speeds have increased only slightly over these years,” the Dec. 1 report states. The result: In peak evening hours, DSL averaged 11.4 megabits per second, which is far below the FCC’s 25-Mbps threshold for broadband that President Obama has called not a luxury but “a necessity.” Cable and fiber-optic broadband, meanwhile, shot past DSL at 52.3 Mbps and 52.22 Mbps.
And that DSL figure overstates the speeds you’ll likely get from the largest providers. At AT&T (T), the service tops out at 6 Mbps, which the FCC study found yielded peak-time mean download speeds of 4.26 Mbps. Verizon (VZ) sells faster tiers, up to 15 Mbps, but so few get them that its 3 Mbps service was the fastest to show up in the FCC numbers—with peak-time mean downloads of 2.22 Mbps.
Why cable wins
Cable in particular has not just outstripped DSL but keeps getting faster. As the FCC study noted, the median cable download speed has increased by 47% annually over the six years the commission has been conducting these studies.
That reflects some of the built-in advantages of cable: The wires themselves don’t need to be upgraded to deliver faster speeds, and they already reach an overwhelming majority of homes.
Sure, the copper phone lines that carry DSL reach even more homes than cable, but the maximum speeds they can reach fade with distance. “The farther away the customer is from the source of the signal, the lower the bandwidth he gets,” Diffraction Analysis CEO Benoît Felten explained via email.
What’s more, DSL connections aren’t easy to upgrade — especially over long distances. As a result, many large telecom firms decided that wireless broadband made a better target for their investments.
“It’s not the physics, it’s the finance,” said Reed Hundt, who oversaw the FCC from 1993 to 1997, the same time DSL emerged as the first mass-market broadband. “The highest and best use of their capital was to put it into their wireless.”
Verizon was one exception — for a time. It launched its Fios fiber-optic service as a complete replacement for its copper phone lines but then halted its expansion into new markets in 2010, suggesting that securing video franchises from cities and counties to offer TV service had become too much of a burden. It’s only relented in one market: Boston, where years of lobbying by city officials (with an assist by New England Patriots owner Robert Kraft) persuaded the company to begin a buildout in Boston last spring that finally saw Fios go on sale there in early December.
That’s left huge swaths of Verizon territory stuck at painfully slow DSL speeds. And the FCC study’s numbers show that millions of people using other DSL providers are doing as bad or worse.
A new hope, maybe
That does not, however, mean that DSL is the rotary-dial phone of broadband. Already-deployed and upcoming advances to the technology can breathe more speed into it, at least for subscribers living close enough to the nearest “central office” network node.
AT&T’s U-verse provides the leading example of this, employing upgraded DSL to link homes to neighborhood nodes that in turn connect upstream via fiber-optic cable. Because the FCC study classified this hybrid service as a DSL product, U-verse is largely responsible for DSL speeds exceeding 10 Mbps.
U-verse mostly relies on VDSL, short for “Very high bit-rate digital subscriber line,” for those last-mile links. Further upgrades have allowed the company to push download speeds up to 75 Mbps in several dozen markets.
But a newer tweak to DSL, suitably named G.fast, can do even better in residences sufficiently close to the nearest node — up to 500 Mbps.
In September, CenturyLink (CTL) activated the biggest deployment of G.fast in the U.S., bringing 500-Mbps access to some 800 apartments in Platteville, Wisc. That choice of dwelling reflects the underlying physics here: G.fast has the shortest range of all DSL technologies.
As a result, CenturyLink spokeswoman Stephanie Meisse said the firm plans to focus on using G.fast to connect apartments, small office buildings and houses to existing fiber lines—saving itself the cost of extending fiber the last few hundred feet.
AT&T echoed that assessment, calling G.fast a way to bring some of the same speed as fiber to apartments with less upfront work.
“G.fast will be a real factor in 2017 but for a relatively small number of homes,” predicted Dave Burstein, publisher of the technical publication Fast Net News.
Felten, however, questioned how much sense this upgrade would ultimately make, given that it doesn’t scale to match demand as well as fiber. “Quite quickly the deployment costs become comparable to fiber to the home.”
None of the above
I suspect most readers would be content if their current phone provider would choose between upgrading to fiber or deploying souped-up DSL that still competed with cable speeds. But the risk is that they’ll do none of the above.
At Verizon, that seems to be the case. Spokesman Raymond McConville said the firm “is looking at” G.fast but has no plans to deploy it yet.
That would leave a lot of customers to stick it out with increasingly uncompetitive speeds or hope for some future salvation.
Maybe 4G LTE’s successor, 5G wireless, will emerge from its current cloud of hype to provide fast downloads without the traditional, restrictive data caps of mobile broadband. Maybe President-elect Donald Trump’s infrastructure plan will include a broadband buildout, something Hundt called a “holy cow” possibility.
Or maybe cable will continue to lock up the residential broadband market. You all are okay with that, right?
(Verizon has made an offer to acquire Yahoo Finance’s parent company, Yahoo.)
More from Rob: