Skip Navigation


Does Google Fiber pausing give BT a told you so card?
Thursday 27 October 2016 11:25:22 by Andrew Ferguson

Google Fiber while not covering as many premises as many had hoped for looks set to pause its roll-out. Whether pause is actually talk for just completing what we are committed to and a switch to a wireless delivery method, or a real pause before they hit the ground running in another year or two is a big unknown. The Google Fiber official blog carries a post by Craig Barratt, CEO of Access who is now stepping down but remaining in advisory role.

The roll out has delivered to various fibrehoods in some nine US cities and another four will see deployment happen, and there will be some job losses suggesting this is not going to be just a short pause but is a change in direction.

The signals have been out there for some time that Google may be switching to a fibre/wireless hybrid roll-out and given the exciting technology opportunities that exist with 5G and other wireless tech this is hardly a surprise.

While the number of customers and exact footprint is harder to ascertain than what BT is up to in the UK, it is clear that in the cities where service has arrived competitors have stepped up their game, and the service is very much appreciated. What is not clear is the split between the $300 up front but then free basic service and the full Gigabit, some indicators suggest the basic service is popular with landlords but a little disappointing for renters as there is no simple way to unlock the Gigabit potential of the glass to the property while renting.

Those with long memories may recall that the BT Group originally was going to roll-out much more FTTP when it started its fibre roll-outs, but once the timescale for installed FTTP was translated into labour costs and VDSL2 was seen to be performing better than originally thought a seismic shift took place and many areas got VDSL2 instead of FTTP. Perhaps Google has hit some of the same issues, i.e. some areas are nice and simple to roll-out in, but others take significantly longer or lots of time and money is wasted on negotiations to get access to infrastructure and as alternate technologies that allow whole streets and apartment blocks to be passed much quicker and cheaper arrive they may be the preferred method for roll-out. In the case of BT this is the new G.fast technology and for Google it is believed to be wireless technology.

Perhaps the problem is that nothing in the fibre roll-out is particularly challenging for Google, just time consuming and they feel its time to move onto the next technology challenge.

Comments

Posted by TheEulerID about 1 month ago
Google's investment model for fibre is rather different to that of OR's. OR are a wholesale only operator, so the investment model is rather different (also, Google are awash with cash).

I always thought that Google Fiber was a bit of an unusual investment for them. Most of their investments have been on technology and plays for a disruptive technology and they've no compunction in dropping products if they don't perform. Fibre infrastructure is not like that. Returns are slow, modest and very long term.
Posted by RuralWire about 1 month ago
This news certainly raises some questions regarding the adoption of a FTTP only strategy in urban areas. Perhaps a more sophisticated and nuanced approach needs to be taken? Swisscom recently announced (18 October 2016) that since September only G.fast-compatible hardware has been used for the Switzerland-wide Fibre to the Street expansion. An interesting addition to the mixture of Fibre to the Curb with vectoring, FTTB and FTTH.

https://www.swisscom.ch/en/about/medien/press-releases/2016/10/20161018-MM-Gfast.html
Posted by RuralWire about 1 month ago
Regarding the decision by Google Fiber to press the pause button on FTTP, the recent acquisition of the ISP Webpass probably provides an explanation.

http://googlefiberblog.blogspot.co.uk/2016/10/welcome-webpass-to-google-fiber-family.html
Posted by TheEulerID about 1 month ago
@RW

Certainly the investment studies have emphasised FTTC and FTTB type approaches are the quickest and most cost effective in urban areas. OR are a long way down the road to enabling all the viable cabinet locations, so g.fast is a bit late and they now have their "pods". But I do wonder if there's scope for major cabinet uplift for hardware that can support both VDSL2 & g.fast standards. Expensive, but a fraction of the initial rollout costs.
Posted by andrew (Favicon staff member) about 1 month ago
Smaller 16 line units exist that do both VDSL2 and G.fast, one of the BT lab demos involves upgrading a connection without any change of hardware at each end - or just user fitting new G.fast capable hardware.
Posted by TheEulerID about 1 month ago
@Andrew

That's interesting. I wonder how upgradable the existing cabinets are to support multiple modulation schemes? Perhaps if they just have passive backplanes it might be feasible to reuse parts. I also imagine there are big technical problems with vectoring groups over large numbers of g.fast lines, but that's an issue that would have to be resolved anyway.
Posted by andrew (Favicon staff member) about 1 month ago
Based on the ongoing work (that already has borne fruit with G.fast starting at 19 MHz rather than 23MHz) the aim is to push lower and co-exist in the some of the VDSL2 band too.

Just need a decent volume of people so we can model a real world range model.
Posted by WWWombat about 1 month ago
The notion of "upgradability" from VDSL2 to G.Fast mostly applies if the telco has already taken the stance of locating the VDSL2 DSLAMs within 200-300m of the end-user. Swisscom have done this with their FTTS rollout, and using the smaller-scale hardware.

In the UK, BT haven't rolled out VDSL hardware in a way that envisaged this future. Allowing FTTC DSLAMs to be 50-100m away from the PCP is evidence enough ... a distance that would lose most of the highest speeds. Hence the pods that are co-located with the PCP.
Posted by WWWombat about 1 month ago
Within the existing cabinets, I'm pretty sure that the DSLAM hardware (including the backplane) is a generation too old for ultrafast speeds. Huawei, at least, have a newer generation of hardware that is being used as a base for the G.Fast DSLAMs.
Posted by WWWombat about 1 month ago
As for Google fiber... Here is the take from across the pond:
https://www.washingtonpost.com/news/the-switch/wp/2016/10/26/why-google-fiber-is-no-longer-rolling-out-to-new-cities/

- Financial pressures
- Insufficient demand
- Pesky incumbents
- Expensive to offer content over the top
- Fear of wireless
Posted by Kebabselector about 1 month ago
Another Google announcement that initially pleased the IT press and shareholders - but like many of their services/products they ditch it once the gloss wears off.
Posted by TheEulerID about 1 month ago
The reference to a new very fast wireless standard is very interesting. However, it looks like that will only work if there are lots of poles on which to mount the wireless transceivers as it surely can't reach very far. Given in urban areas much of our cable goes underground, that might not translate well to a UK environment. I would have also thought that power suppliers were an issue.
Posted by RuralWire about 1 month ago
@TheEulerID - There has been speculation that Google are interested in evaluating, amongst other things, pCell developed by Artemis Networks. Webpass have been involved.

www.artemis.com
Posted by WWWombat about 1 month ago
pCell looks rather interesting.

In essence, a system of multiple antenna/remote-radio-heads, that takes advantage of cloud-RAN techniques and adds an element of beam-forming. A bit like a TV soundbar made of 20+ small speakers.

Of course, it requires a lot of radio heads to be beneficial to multiple users: lots of poles to be mounted on, plus power, of course. And lots of fibre, or more wireless, for the fronthaul duty.
...
Posted by WWWombat about 1 month ago
The capacity available to users appears to depend on having lots and lots of transceivers. You don't need many radio heads while your user base is low, but you will need to keep adding more to keep up with demand for that capacity.

In a suburban setting, I suspect this would turn out to be somewhat equivalent to a G.Fast node placed at each DP, needing power and fibre backhaul.

Initially, though, you may only place a radio at the equivalent of a cabinet.

The biggest advantage is that it has no need for a final drop into the home - copper, coax or fibre.
...
Posted by WWWombat about 1 month ago
The advantage of pCell is that it works (invisibly) with 4G devices.

Here's some equivalent research from Ericsson as part of their work towards 5G, where you'd expect beamforming to be a major component.
https://www.ericsson.com/research-blog/5g/massive-beamforming-in-5g-radio-access/
Posted by amiga_dude about 1 month ago
The problem Google Fiber has been having is AT&T an co and these court cases.

http://arstechnica.com/tech-policy/2016/09/att-sues-nashville-in-bid-to-stall-google-fiber/
Posted by michaels_perry about 1 month ago
Why are we talking about what Google is doing in the USA when it does not have any current bearing on the market in the UK?
Posted by WWWombat about 1 month ago
Why?

Because the Google Fiber project is held up as a poster child by anyone who wants to proclaim how marvelous fibre is, and how it just takes one company with vision to change the entire market.

It is used to show that all we need in this country is to regulate BT some more, and allow in a similar visionary pioneer.

Turns out that filthy lucre gets in the way, after all.
Posted by andrew (Favicon staff member) about 1 month ago
And I thought that was clear enough from the headline I used, and in article I did highlight similarities with the early Openreach roll-out and changes in priorities they made too.
Posted by gerarda about 1 month ago
@wwwombat

Could the pcell technology be mounted subscribers premises instead of poles?
Posted by pkaulf about 1 month ago
Told you so?

Told you so that incumbent operators will do everything in their power to stall and stifle competition and avoid having to get the wallet out?
Posted by DrMikeHuntHurtz about 1 month ago
Thatcher is the reason we don't have FTTP.
Posted by WWWombat about 1 month ago
@gerarda
I guess it could, but to what aim? You'd lose a number of benefits, so I doubt you'd do it.

The main gain comes from beam-forming of multiple antenna heads pointing at a subscriber simultaneously. That needs multiple antenna heads covering each subscriber ... which would be lost if you put them inside homes.

There is a question of how much power you'd accept coming out of an antenna inside the home ... limiting its use for other subscribers.
...
Posted by WWWombat about 1 month ago
In the end, if you swap to a model of a radio head inside each home, unusable by others, with fibre backhaul to each one ... haven't you just recreated FTTH with a router and a private WiFi AP?

This is radio, with shared coverage. Tens of pole, hundreds of subscribers, and fairly high transmit power. As takeup grows, it becomes hundreds of poles, thousands of subscribers, and lower power ... but still enough for a few hundred metres.
...
Posted by WWWombat about 1 month ago
There *are* in-building, in-shopping-centre, in-stadium applications. But probably not in-home ones.
Posted by WWWombat about 1 month ago
Thinking about it, UK Broadband could make use of pCell, in the kind of deployment they're doing around Swindon.
You must be logged in to post comments. Click here to login.