Updated April, 2018
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: 16 May 2016 16 May 2016
Coming May 19 at Paris DSL Summit. I got this wrong when Calix announced G.fast over coax. I thought it a niche product because I didn't realize the DirecTV deal brought millions of lines of coax to AT&T. AT&T is not a niche.
CTO Rami Verbin is going to present unprecedented G.fast performance live Thursday. In the release below, the weblink also takes you to some video including from AT&T's Eddy Barker. Adtran's Ronan Kel will open the show with a live demo of a related DSLAM.
I doubt the 750 up, 750 down that AT&T is discussing is quite ready, but it's coming. Double speeds will require the next generation of chips that extend frequencies to 212 MHz. That's on everyone's roadmap and Sckipio is optimistic about early delivery.
AT&T and Telus provided quotes, a strong vote of confidence. "With dynamic bandwidth allocation, we believe AT&T can offer up to 750Mbps in both downstream and upstream performance over coax with today's chipsets," said Eddy Barker, Assistant VP of Technical Design & Architecture at AT&T. "In the next generation G.fast chipsets, we should be able to double that target â€“ achieving as much as 1.5Gbps in each direction.
- Published: 10 May 2016 10 May 2016
It works, finally. As always, the telcos were cautious and the vendors promised too much. The dam has broken now, with BT going at 3M homes passed each year. http://bit.ly/BThalffast CEO Patterson will keep going until “almost all” Brits can get what he’s calling ultrafast broadband.
AT&T looks to be the next big order, millions of lines with fiber to the basement and G.fast to each apartment. http://bit.ly/GfastATT. Chunghwa in Taiwan is covering almost the entire country. In Australia, Parliament is debating how much of the National Broadband network should switch to G.fast. Rio de Janeiro, Panama, Belgium, and Switzerland are on the way.
- Published: 09 May 2016 09 May 2016
Ready to upgrade 3M/year. Kevin Foster of BT minced no words. "G.fast must deliver speeds over longer distances ~ 350m (0.5mm Cu.)" BT's original plan was to support G.fast from 4M distribution points, for speeds of > 500 megabits. The finance guys decided to go only to the neighborhood cabinets, requiring 95% fewer installations and minimal new fiber. That raised the typical loop length from 35 meters to 350 meters.
To get even the lowered speeds (~300 megabits) at 350 meters is pushing the state of the art. The change in requirements has delayed most work from early this year to end of this year. The tests are going well to over 4,000 homes, at least at short distances.
Kevin thinks it can be done. He believes engineers can:
- Enable higher bits per tone
- Improve the receiver sensitivity (noise floor)
- Increase the transmit power
- Optimise the frequency usage with VDSL ? Increased vectoring group sizes ?
Rami Verbin of Sckipio saw the same possibilities in an interview last June. Some of them should be close to delivery by now. Foster also demanded "G.fast units must have a higher DPU port density (typically 48/96 ports.)" Fortunately, Huawei tells me they expect a 96 port unit late this year. Alcatel should be on par with them.
- Published: 09 May 2016 09 May 2016
DT is not delivering. They already are far behind schedule for the 24M homes they promised to pass with 100 megabit service. They also reduced the minimum speed to 50 megabits and do not even guarantee that speed. They hide behind almost meaningless "up to" figures; America now requires "typical speeds" and Britain wants to do likewise. DT has given us no reason to believe even a quarter of the vectored lines will deliver 100 megabits.
This is a financial, not a technical problem. The G.vector they are using provides 100 megabits to many in Belgium and 80 megabits to most. The technology is generally working fine, although DT has cut some corners. It's possible DT's engineers misestimated what the speeds would be. It's also possible that the original claims of 100 megabits were a "politician's truth." More likely, they've cut the numbers of cabinets and the fiber extensions. (DT has excellent engineers who are current with the state of the art. I doubt they made a mistake.)
DT's CEO Höttges' comments would make super-spinner Karl Rove proud.
- Published: 08 May 2016 08 May 2016
Budget guys refuse to service the 4 million distribution points across the UK. BT's CIO, Peter Bell, publicly confirmed what the industry has speculated since last spring: BT is not planning fiber to the distribution point. G.fast was designed for distribution points, up to 16 homes very close to the DP. That was BT's plan until the budget people got involved. Now, because of budget limits, BT intends to reduce the number of locations serving G.fast by 94-97%.
Bell's comments imply that France Telecom and Telefonica are facing severe financial distress. Fiber all the way home costs more than FTTdp. Telefonica has reached 15M homes with direct fiber. France Telecom has reached 5M fiber homes passed already, including most of Paris. They plan >2M more homes each year until 2022. Neither analysts nor investors predict a financial calamity. ORAN and TEF have a market cap of $94B.
BT hopes most customers will get 300 megabits rather than the 500-800 G.fast was designed for. Some customers will be very disappointed when they discover "superfast broadband" is slow to their home.
- Published: 07 May 2016 07 May 2016
Really G.halffast. A network where many will get 100-300 megabits is a major improvement, although it's not the 500+ megabits CEO Gavin Patterson envisioned with G.fast. Higher speeds would require running fiber to the millions of distribution points; the money men decided to install fewer than 10% as many boxes, going to some of the 90,000 field cabinets instead.
Going to the cabinets is likely to bring ~300 megabits to less than half and slower speeds to everyone else. The "typical" home is 350 meters, which if lucky can carry 300 meg. Some unknown percentage will get 3 megabits or less.