Gfast map

gfast map copy

 

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

Light Blue: Incumbent likely:  France, Germany, Italy

Cologne-Cathedral-from-UNESCOKick Deutsche Telekom out of the buildings and everyone gets 100 megabits more.  Sharing copper lines became obsolete when vectored noise reduction became practical, first in VDSL and now in G.fast. As processors became fast enough to cancel noise across the lines in the binder, speeds double if one operator is in control. Deutsche Telekom is demanding a monopoly where they upgrade to 30-100 megabit vectored VDSL; City-owned Netcologne is instead offering 500 megabit+ G.fast.

Technical officer Horst Schmitz tells Golem.de they will connect 250,000 homes in 2017 and another 100,000 after that.

NetCologne has an established program to deliver fiber working with other cities. If DT doesn't increase speeds, others may be tempted. 

G.fast can run about 100 megabits faster (or go 100 meters further) if the first 20 MHz are not used for some other carrier's VDSL. John Cioffi of ASSIA makes that point strongly. I confirmed the engineering details with several in Paris. NC wants to kick DT's VDSL off the lines, just as DT is demanding to kick everyone off the lines in other parts of the country. I can confirm that several very senior people in England and Germany are listening. It's very hard for governments and giant companies to change to other forms of unbundling but 100 megabits more to every home is a powerful incentive. 

What no one is saying publicly is that traditional VDSL will fairly rapidly fail in competition with G.fast and gigabit cable. 17 MHz VDSL delivers 20-70 megabits a modest distance. BT sells it as "up to 72 megabits" but that's deceptive; I've seen no evidence that even 10% of lines can get that speed. Vectoring can double that but you can't vector VDSL and G.fast in the same binder. 

Over time (?3-7 years), so many of the VDSL unbundled lines will switch to higher speeds they will generally become too expensive to service. G.fast (and gigabit cable) cost very little more than standard VDSL, now a 10 year old and soon obsolete technology. There are many marketing scenarios possible but ultimately the 10x faster service will win most customers. 

That doesn't have to mean broadband competition dies. Sharing the line from the exchange to the home becomes untenable as chips get fast enough to do noise cancellation (vectoring). Vectoring increases speeds from 2-10X. That's worth changing the competition rules.

DT is considering using the 300 meg speeds of LTE combined with slow DSL to compete with cable. They already have DSL + LTE as an offering. They are rapidly upgrading to 3 carrier LTE and have enough spectrum to raise that to gigabit speeds in the next few years.

LTE at 19 megabits today is faster than Deutsche Telekom's average DSL speeds.