In 2016, G.fast looked very promising.
But only BT & Australia's nbn remain
Dark Blue: Firm commitments from incumbent: BT (10M), Belgacom, Australian NBN, Swisscom, Austria, Bezeq Israel, Chunghwa Taiwan, Telus Canada, Telekom South Africa, SK Korea, (U.S.) AT&T, Century, Frontier, Windstream, Belgium, Omantel
Mid Blue: Smaller carriers in Germany, Norway, Finland, Japan
- Published: 10 January 2017 10 January 2017
Hundreds of megabits to apartments for a very low cost. For a decade, millions have been getting 100 megabits from fiber to the basement + VDSL on existing building wiring. Now, there's a surge of wireless to the roof, led by AT&T in Minneapolis and Webpass (bought by Google) in San Francisco. Speeds range from 200 megabits to nearly a gigabit and system latency is good. The very visible and ambitious Starry is starting to test WTTR in Boston. They are running mmWave to the rooftop for backhaul and using G.hn to reach apartments.
** Update: I had a chance to talk to CEO Chet Kanoja. They have not seen a G.hn congestion problem and he doesn't expect problems going forward. Everyone's watching for field results. They found an ODM supplier for the G.hn at the transmitter and design their own gateways. Even more interesting, they are going millimeter wave to houses and apartments in volume. He's designed a 5 gigabit mmWave transmitter built up from 802.11ac that is point to multipoint. He says his cost is $800-$1,000, which will transform the economics. It implies he will be delivering something similar to Verizon's 5G at a much lower cost. I'll remain skeptical until that's widely deployed in the field, but the details he provided were convincing. Dave **
The very visible startup has big plans to deploy more cities. For a decade, millions have been getting 100 megabits from fiber to the basement + VDSL on existing building wiring. Now, there's a surge of wireless to the roof, led by AT&T in Minneapolis and Webpass (bought by Google) in San Francisco. Speeds range from 200 megabits to nearly a gigabit and system latency is good.
Verizon is going direct to apartments with their 5G mmWave, soon to go to several hundred test homes in semi-rural Massachusetts. There is no planned fanout with DSL or G.hn. (Verizon is holding off on their original plans for 5G to parts of Boston rather than fiber.) Kanojia did a remarkably fast multi-city rollout for his previous company, Aereo. He's promising to go wide very quickly this time as well. Aereo had an innovative way to deliver free to air stations over the Internet. The stations shut it down by persuading the Supreme Court it illegally denied them royalties. That was a very anti-consumer decision.
VP Virginia Lam Abrams writes, "we can get a minimum of 500-700 Mbps, even on cat 3." That puts them far ahead of DSL except G.fast and able to compete with cable. Most of the older problems with microwave, like rain fade, have been pretty much solved. The service should be reliable with good latency.
G.hn is a simpler technology than G.fast. Korea Telecom has jumped on hard, deploying to millions and selling abroard. PLDT has picked it up in the Phillipines. Marvell has extended the G.hn home networking system to cover buildings and continues to improve the chips. The G.fast guys believe G.hn will have a problem when the networks get loaded because it doesn't do vectoring. Chano Gomez of Marvell strongly disagrees. We'll know more when we have results from the field.
Gigabit microwave for backhaul costs only $3,000 from Ubiquiti, although Starry is going with a more expensive phased array point to multipoint system. Two and five gig microwave is coming to market. That 1 gigabit of backhaul can generally deliver 500-700 megabit service to many customers surprises everyone except network engineers. The experience is that very few customers demand high speeds most of the time. Two HD TVs can be served with 10 megabits; not many homes need more than that very often. Sharing works very well.
With low backhaul costs and minimal installation, WTTR could bring down broadband costs significantly. Starry expects they will and hopes to very rapidly grow their customer count.
I hope he succeeds; the U.S. desperately needs more competition.
Important note: I have held back several points in the original article to more carefully factcheck with the company. I wanted to publish the G.hn news. There are questions to resolve about how many apartments can be reached directly with wireless, what is the network total capacity ... None of the open issues negates my conclusion "The service should be reliable with good latency."
Here's some of their earlier pr, now superseded by the current technology.
Introducing Starry - A New Technology Company Revolutionizing How We Connect to the Internet