Will BT actually deploy *any* fibre beyond the cabinet?

G.Fast has historically been described as a stepping stone technology to FTTH, with fibre running to the distribution point and then using the existing copper drop wire with G.Fast technology to deliver 300 - 500 Mbit/s and perhaps eventually 1Gbit/s to the end user. G.Fast to the distribution point would undoubtedly be much faster and cheaper to roll out than FTTH, as the fibre only has to be taken as far as the distribution point. The FTTH upgrade to replace the drop wire can then be done only as and when you need to...

BT has been talking a lot about G.Fast and in its field trials, it has demonstrated fibre to the distribution point (i.e. typically within 10s of metres of each property). But if you look carefully at its briefings in the last 6 months, BT seems to only be planning to actually deploy G.Fast from existing FTTC cabinets and says it is uneconomic to roll out fibre even as far as the 4 million distribution points it has, let alone all the way to the home.

Even the system vendors that work with BT have technolgoy that they claim would get much more performance out of the existing copper loop / cabinet architecture. Huawei talks about 400Meg at 300 metres and still delivering 100Meg at 800 metres using their SuperVector product and 35Mhz of bandwidth instead of 17Mhz whilst Alcatel / Lucent talks of 200Meg to 500 metres and 300Meg to 250 metres with their VPlus vectoring solution.

No doubt a quick change of line card in the cabinets to upgrade lines from VDSL2 to a long range version of G.Fast will be much cheaper for BT than actually going to the time and effort of rolling out fibre any further into the access network. But it will also remove the vital stepping stone the country needs to get to full FTTH, perhaps for the foreseeable future. After all, if BT can really get 100Mbit/s to most people by going down this route, where will the value be in ever rolling out fibre all the way?

No wonder Sky and TalkTalk seem to have gone quiet on their FTTH experiment in York. No other scale provider is going to be able to compete with BT given that BT's cabinets are all in place already and a cheap speed upgrade is already in the offing in the next few years. They would be unable to compete with BT on price in this scenario, whilst BT is allowed to keep milking its existing copper plant forever.

The losers will be those that are too far from their cabinet to benefit (i.e. more than 800 meters, which will particularly affect rural households) and everyone else who struggles to get a reliable broadband connection over their existing cruddy and outdated copper wiring which BT refuses to fix. I anticipate that BT won't care about any of these "uneconomic" customers who can't get a decent service from them, providing that most people can and more critically that BT does not lose too much market share to Virgin. If this is the route BT is going, we're going to need a very strong regulator and a Universal Service Obligation that does not fall behind where everyone else is to ensure BT is not permitted to make a large digital divide a permanent and enduring feature of the UK.

But perhaps the biggest loss will be removing the opportunity to actually transition the whole country from unreliable copper to a much more reliable and future proof fibre solution. Unfortunately, the public at large seems unlikely to put much value on having a more reliable and future proof connectivity solution. So we will all have to put up with BT giving us a cheap and sometimes dodgy service over our ancient copper lines.

BT have FTTPs in very limited amounts

but of course its very expensive to rollout

i think maybe in 20 years we may have 70% FTTP in the country and thats being optimistic.

Half the country already has access to 200 Mbps its not via FTTP but its still a fixed speed connection medium (1.65% of this 50% is actually FTTH)

In reply to a post by rgp:

... G.Fast will be much cheaper for BT than actually going to the time and effort of rolling out fibre any further into the access network. ... After all, if BT can really get 100Mbit/s to most people by going down this route, where will the value be in ever rolling out fibre all the way?

Think you have answered your own questions.

We can speculate that a 100/20 connection might be good enough for most residential purposes for the next 10-15 years maybe. Somewhere on TBB it was posted, a long time ago, that something like +80% of people are within 400-500m of a cabinet (somebody can correct my numbers I am sure). Expect BTO to offer faster options, FTTP On Demand, to those who are willing to pay for it.

So Yes, you are right, BTO will not be in an immediate rush to push out Fibre to the DP, those on long lines will be loosing out again while BTO wait for another handout from BDUK.

BT definitely have a choice to make: retain VDSL2, but add vectoring and profile 35b (which is the standardised name for the two proprietary ideas of Vplus and superVector); or use G.Fast.

The two aren't compatible, as they both use the spectrum above 17MHz in incompatible ways ... so it is one or the other.

The ANFP has already been re-written to allow G.Fast, with the appropriate change to allow G.Fast DPU nodes to be sited deeper than the current PCP. Right now, therefore, BT are following the G.Fast route.

But the location of the DPU is not a binary one of "at the PCP or DP" alone. They can be located anywhere between the two. If it is too uneconomic for DP placement, they might target distances of, say, 200-250m instead. Presentations by Sckipio suggest that BT is indeed chasing that target longer term, even if their first steps will be to put the DPU at the existing PCPs.

If BT were going to be happy getting 100mbps to half the premises, they would only have planned on vectoring for VDSL2. Instead, they seem to be taking this on to the next generation - where they have to be aiming at that 300Mbps target to 50%+.

In reply to a post by godsell4:

Somewhere on TBB it was posted, a long time ago, that something like +80% of people are within 400-500m of a cabinet (somebody can correct my numbers I am sure).

This is the spread of D-side lengths:
http://postimg.org/image/bp372fcnn/

17% within 200m of the PCP
49% within 400m
72% within 600m
84% within 800m
90% within 1km

The distance distribution from cabinets you quote is very interesting...

Skipio have produced an slide showing speed vs. distance for G.Fast (20 Mhz - 106Mhz) and VDSL2 35b which suggests that a likely G.Fast target is 300 Mbit/s at 300 metres. Many UK properties are much further than this from their serving cabinet, so if you wanted to pursue a pure cabinet deployment strategy (like they are in Germany) it would appear that you would be much better off deploying VDSL2 35b instead of G.Fast...

Based on the distribution of line lengths from the cabinet, it would appear that BT will have to push fibre further from the cabinet if they really are going to use G.Fast in the 20Mhz-106Mhz band as per the ANFP, as customers would otherwise be better off sticking with VDSL2 from the cabinet where their line lengths is 400 metres or more which seems likely to be a lot of customers.

But deploying fibre deeper into the distribution side would seem to be at odds with BT's apparent ambition to deploy G.Fast from existing FTTC cabinets. Deploying G.Fast only from the cabinet starting at 20Mhz doesn't really seem to make sense, as too many customers would be too far away to benefit...

So, on the assumption that BT can only use 20Mhz-106Mhz for G.Fast (as per the ANFP), BT must be planning to deploy at least some fibre into the distribution side as part of its rollout so that its target customers are no more than 300 metres away...

But how many actually take-up the 200Mbps Virgin option? It's an £8 per month premium to the 100Mbps service and a £13 / month premium to the 50Mbps service.

If Virgin's take-up is very low, this suggests that the public mostly isn't prepared to pay much of a premium for more speed. This doesn't bode well for BT's G.Fast rollout, but perhaps BT feel they have to do it to try and avoid losing business to Virgin...

Well so far the G.fast public trials have been fibre to the DP

Take-up of faster (i.e. price premium) products is low, even on Hyperoptic and Gigaclear the full Gig products are not the most popular option, i.e. speed above a certain point becomes bragging rights versus a higher utility bill

In 16 years Virgin I don't think has ever had a headline speed product slower than BT but they have not destroyed BT products in the areas they overlap.

Some of the up to campaigning could actually be good for BT and bad for Virgin Media i.e. median speeds or lower quartile type speeds some are calling for don't look great for Virgin Media.

In reply to a post by rgp:

Somewhere on TBB it was posted, a long time ago, that something like +80% of people are within 400-500m of a cabinet (somebody can correct my numbers I am sure). Expect BTO to offer faster options, FTTP On Demand, to those who are willing to pay for it.

Complete nonsense. The majority of lines are > 500M, and quite a significant percentage have aluminium runs which turn these 500M lines into 1000M runs in reality.

The local loop is in a terrible state currently.

Edited by dave2150 (Wed 04-May-16 03:37:39)

In reply to a post by MrSaffron:

Take-up of faster (i.e. price premium) products is low, even on Hyperoptic and Gigaclear the full Gig products are not the most popular option, i.e. speed above a certain point becomes bragging rights versus a higher utility bill

In 16 years Virgin I don't think has ever had a headline speed product slower than BT but they have not destroyed BT products in the areas they overlap.

Some of the up to campaigning could actually be good for BT and bad for Virgin Media i.e. median speeds or lower quartile type speeds some are calling for don't look great for Virgin Media.

Although all VM have to do is increase the most basic Broadband package to 100Mbit, to completely eclipse everything BT have to offer for the next 5 years.

In reply to a post by rgp:

Many UK properties are much further than this from their serving cabinet, so if you wanted to pursue a pure cabinet deployment strategy (like they are in Germany) it would appear that you would be much better off deploying VDSL2 35b instead of G.Fast...

I disagree. Even though pro-FTTP campaigners describe FTTC as a cul-de-sac, profile 35b is worse than that: a one-way cul-de-sac that locks you in, and prevents a graceful switch to G.fast.

As the graph shows, there is no real benefit to 35b ... unless you have no intention of ever needing to use G.Fast. That is - you intend to jump to FTTP as the next step.

In reply to a post by rgp:

But deploying fibre deeper into the distribution side would seem to be at odds with BT's apparent ambition to deploy G.Fast from existing FTTC cabinets.

You're making a common mistake here: that BT intend to *only* deploy G.Fast from the PCP site.

Of course BT will deploy G.Fast from those PCP sites. It is a no-brainer, with both fibre an power on-site, there is almost no civil engineering cost whatsoever. The only consideration left will be whether there are enough lines to create demand for a 16-port DPU.

The choice to deploy fibre deeper than the PCP is neither at odds with this, nor in step. It is an orthogonal choice for BT - totally independent.

So what will they choose? I have no idea.

In reply to a post by rgp:

So, on the assumption that BT can only use 20Mhz-106Mhz for G.Fast (as per the ANFP),

BT can only use that spectrum because it cannot overlay the VDSL2 17a spectrum with G.Fast signals: the crosstalk would ruin both systems. The only way for BT to jump to 2-106MHz would be to bar VDSL2 from cables that have G.Fast in them (and vice-versa).

That, of course, loses any graceful deployment characteristics.

In reply to a post by rgp:

BT must be planning to deploy at least some fibre into the distribution side as part of its rollout so that its target customers are no more than 300 metres away...

Right now, the only concrete target is for "10 million lines by 2020, with 300Mbps". That's around one-third of all lines - which, coincidentally, is about the target that could be reached by adding a DPU alongside every FTTC cabinet and going no deeper.

To reach more - the "most of the UK by 2025, with 500Mbps" target - does indeed need deeper fibre, and another couple of standardisations, plus a couple of generations of hardware. Plenty of time yet.

Who knows where BT plans to go yet? I'm not sure that they do, at least not in sufficient detail.

In reply to a post by rgp:

But how many actually take-up the 200Mbps Virgin option? It's an £8 per month premium to the 100Mbps service and a £13 / month premium to the 50Mbps service.

If Virgin's take-up is very low, this suggests that the public mostly isn't prepared to pay much of a premium for more speed. This doesn't bode well for BT's G.Fast rollout, but perhaps BT feel they have to do it to try and avoid losing business to Virgin...

There was a good graph in the Ofcom Infrastructure report in 2014, showing the actual "sync speed" of all broadband subscribers, but broken down by each technology.
http://postimg.org/image/bly6o16cl/

That shows that VM were in the process of gradually upgrading people from a product set of 30/60/120Mbps to a product set of 50/100/150Mbps.

However, even within those moves, we can see that only 10% of VM subscribers thought it worth paying for the top tier (120 or 150Mbps), while 63% of them stuck to the bottom tier (20, 30 or 50Mbps). With around 4.5m subscribers in VM, that 10% probably amounted to around 450,000 properties overall.

Does that mean that there is equivalent demand from the whole population? 10% of approximately 28 million lines? I don't think so ...

I think it is s fair to assume that *all* people who wanted a 120Mbps+ line, and could get a VM service, will have eschewed BT's offering of FTTC, and deliberately moved to VM.

As VM covers approx 50% of the UK (ie approx 14 million lines), I think it is fair to believe that the demand for top-tier service amounts to 450,000 from 14 million lines: about 3%. If there was similar demand from the other half of the country, then you would be looking at nearly 1 million lines in 28 million: still around 3%.

A 3% demand for ultrafast speeds is about one-tenth of the demand for superfast speeds. A G.Fast DPU probably only needs to be able to handle 16-32 lines, and would cope with all the demand in the PCP area, even if it could only reach half of the subscribers.

Yes it's a shame I'll never get cable here.

There is this big issue you allude to in that the nominal "link speed" which gets so heavily advertised can be very different to what the real throughput is in a contended environment. All network technologies short of fully dedicated ones have contention points. They just occur in different places (or combinations of places). It's often extremely difficult to find out where this contention is occurring and as it often varies from location to location depending on local conditions it's extremely difficult to know what is actually going on. I'm sure the network operators have (or at least should have) data about the contention occurring at different points in their network architecture as part of their capacity planning processes. However, I don't see that being made publicly visible on any comprehensive basis (I know some smaller service-orientated ISPs have published some aspects of this).

In any event, this is a hugely complex area and a lifetime in IT (much of it involving performance and service issues on very large systems) tells me that quantifying service is a tricky thing and doing it well requires a lot of expertise and investment in management systems. There are two main ways of tackling system performance. One is to use "primary" performance measures. So that would be direct things like latency and throughput as actually seen by real users using real systems. That is very, very difficult, especially when not all elements of the journey are under the control of the operator (as is the case with servers). Even measuring at the boundary of your service domain is very tricky, and the volume of data is immense.

The second approach (which is easier and adopted by many service providers) is to give up on actual primary service measurement and resort purely to capacity planning measures. Things like the peak utilisation of network segments, loadings on router nodes and so on. Often it uses things like KPIs (key performance indicators) on some sort of dashboard arrangement. The principle is if utilisations are within parameters then this correlates with decent services.

Broadband speed tests occupy something of a middle ground in all this. They are OK at benchmarking and eliminating some sort of problems, but they don't directly measure the service seen by real users. That middle ground is often used in IT systems through the process of "dummy" transactions run automatically. You hope that these have some correlation with real world experience of end users, but they are still somewhat artificial.

All this is a bit of stream-of-consciousness stuff, but reflecting a whole career of pain dealing with the characterisation, measurement, monitoring and resolution of IT performance issues...

Care to expand on what a significant percentage of aluminium means in terms of number of premises affected?

VM can easily do that now if they want, but it might not be a good idea due to the capacity levels of the network and the small percentage who see it as obligatory to fill every bit of every second with some data.

Not if they can help it, which longer term they can't.

There may be more FTTP in there now they appear to have decided to stopped deploying it in such an asinine way and are testing cheaper and faster deployment methods, however there will be the bare minimum of FTTP required to satisfy the politicians and any perceived commercial needs.

Outside of the FTTP areas where they basically set fire to piles of cash they spent the bare minimum which is exactly what you'd expect a privately owned company to do.

G.fast will be deployed in the cheapest way while reaching the highest level of coverage which will eliminate even some urban areas from the deployment as they are too far from an FTTC node and no simple way of powering G.fast node, getting fibre to it and intercepting copper closely enough to make a difference.

What this does mean, though, is that there will be some areas where FTTP makes more sense than G.fast. Whether these will receive FTTP or simply be labelled as commercially unviable I've no idea, but given I'm probably in one of them, DP is in block paving, I'm interested to find out.

Of course they VM will not do that because they do not have the capacity to support that as any speedtest results would quickly show and they would fall foul of the ASA.

In reply to a post by WWWombat:

A 3% demand for ultrafast speeds is about one-tenth of the demand for superfast speeds. A G.Fast DPU probably only needs to be able to handle 16-32 lines, and would cope with all the demand in the PCP area, even if it could only reach half of the subscribers.

I think BT would be laughed at if they only roll out G.Fast to existing cabinets such that only 50% of the possible users on a cabinet could actually get a better service than they already have with VDSL2.

With VDSL2, BT can get away with selling "up to" speeds, as almost everyone gets much better speeds on VDSL2 than they would have got with their existing ADSL2 service. For instance, I am 1.5Km from my fibre cabinet and I get 20Mbit/s download and only 800Kbit upload. Yes, this is way less than the 40Mbit download, 10Mbit upload service that I am paying for, but it is still way better than the 1Mbit download, 800Kbit upload ADSL2 service I had previously.

You're paying for the speed you were quoted at the point of sale. Not the maximum potential of the product.

In reply to a post by rgp:

After all, if BT can really get 100Mbit/s to most people by going down this route, where will the value be in ever rolling out fibre all the way?

Removing so much complexity and operational cost. FTTP cuts all the BS - no expensive, power hungry, weather sensitive equipment on poles or in cabinets. No need to source power supplies for them all. Absolute futureproofing compared to any copper technology.

In reply to a post by rgp:

The losers will be those that are too far from their cabinet to benefit (i.e. more than 800 meters, which will particularly affect rural households) and everyone else who struggles to get a reliable broadband connection over their existing cruddy and outdated copper wiring which BT refuses to fix. I anticipate that BT won't care about any of these "uneconomic" customers who can't get a decent service from them, providing that most people can and more critically that BT does not lose too much market share to Virgin.

Depending on where you live, those customers may well be enjoying FTTP today. Such is the BT technology lottery, where some rural dwellers (and I mean really rural) get the very best, some are on the end of FTTC (whether they're on great lines or poor ones) and others get relegated onto satellite.

Edited by deleted (Wed 04-May-16 21:12:07)

Just because you have an aluminium line which is more than 500 metres long, you are wrong to think that the majority of lines are >500metres.

Also, I suspect the amount of aluminium in the network isn't that great in % terms - certainly not a significant % (how much do you think there is?)

In reply to a post by rgp:

In reply to a post by WWWombat:

A 3% demand for ultrafast speeds is about one-tenth of the demand for superfast speeds. A G.Fast DPU probably only needs to be able to handle 16-32 lines, and would cope with all the demand in the PCP area, even if it could only reach half of the subscribers.

I think BT would be laughed at if they only roll out G.Fast to existing cabinets such that only 50% of the possible users on a cabinet could actually get a better service than they already have with VDSL2.

I think your maths is rusty.

A capacity that is 50% of subscribers on a PCP is about what an Huawei 288 cabinet manages in some locations. Adding a 32-port DPU is going to be nearer 5%.

What is the problem in supplying limited capacity when there is only limited demand? It is the principle that telecoms has worked on for over a century.

. Yes, this is way less than the 40Mbit download, 10Mbit upload service that I am paying for

Yawn. There went any signs of sensible discussion.

In reply to a post by WWWombat:

BT definitely have a choice to make: retain VDSL2, but add vectoring and profile 35b (which is the standardised name for the two proprietary ideas of Vplus and superVector); or use G.Fast.

I suspect that profile 35b was never in any real prospect of being rolled out in the UK.

If it hadn't been for the decision to concentrate on g.fast I believe we would have seen profile 30a here instead.

Edited by gt94sss2 (Thu 05-May-16 00:40:45)

I don't think it was about capacity but about the number of people that would actually get an up lift. If you only have 30% of users signing up to a service and only 50% of those could see an up lift from G.Fast then essentially you only have 15% of users in a cabinet that you can sell the service to (and not all of those would bother to upgrade because they will already be getting decent FTTC speeds).

The real benefit on G.Fast is moving the node closer to the property so that those with lower FTTC speeds actually get a quantifiable increase that they would be willing to pay extra for - that is much less likely if you only do G.Fast at the cab.

Just about any infrastucture build in the late 1960's and 1970's will most likely be aluminium based because of the global copper shortage at that time.This would also include cable repairs of the existing infrastucture during that period.

It may not be a significant percentage overall but in towns that had a lot of development during that period there will be a fairly high percentage of properties served by all or part aluminium cabling.

In reply to a post by simon194:

Just about any infrastucture build in the late 1960's and 1970's will most likely be aluminium based because of the global copper shortage at that time.This would also include cable repairs of the existing infrastucture during that period.

It may not be a significant percentage overall but in towns that had a lot of development during that period there will be a fairly high percentage of properties served by all or part aluminium cabling.

Both my lines are partially aluminium, this cuts me from an 80/20 sync (as was the estimate at the time of order) to a ~43/6 sync on both lines (as the BT Wholesale estimate now says after engineer visits etc).

The average age of houses on my road must be about 300 years, mine is about 500 years for the main house, extended in the Victorian era, so unless the lines were not installed until ~1960/70 for some reason, BT just decided they liked aluminium here.

You have your answer in that the news is reporting a BT objective of installing FTTP to 2 million more premises (in addition to what is being done under BDUK).

However, I think the areas chosen will probably be relatively high density with suitable passive infrastructure. If you look at the north Swindon trial using connector-based technology, then this might well be the roll-out model that's chosen. In areas where g.fast is a significantly more cost-effective method of delivering ultrafast, I think that will take precedence. There will also be a certain amount of politics in this in deal. Partly BT will want to be able to claim they are the UK's largest FTTP operator, and partly to mollify Ofcom who, for good or ill, will see the FTTP figure as important in international comparisons.

What I don't expect to happen is these 2 million premises being largely rural infill.

From what I read there is not going to be much overbuild in residential areas even if the passive infrastructure is there.

New build, business parks and high streets. Brownfield sites get G.fast.

I'm sure there will be exceptions but this seems to be the way they are going, even though there are going to be millions of residential premises with the infrastructure in place.

In reply to a post by nemeth782:

In reply to a post by simon194:

Just about any infrastucture build in the late 1960's and 1970's will most likely be aluminium based because of the global copper shortage at that time.

Both my lines are partially aluminium...

The average age of houses on my road must be about 300 years, mine is about 500 years for the main house, extended in the Victorian era, so unless the lines were not installed until ~1960/70 for some reason, BT just decided they liked aluminium here.

There are a couple of complicating factors to both of these statements...

It isn't correct to consider "the 60's" and "the 70's" as a long period of copper shortage (or, more importantly, high copper prices).

Prices were indeed slightly higher from 1968 through to 1975, but there were two significant spikes in 1970 and 1974. I suspect use of aluminium peaked in those two years, rather than being prolonged over a decade+.
Copper price graph here

However, the years in question are also quite significant in the growth in telephone ownership.

In 1966, there were only 4.3 million homes with a phone line, out of (I think) around 17 million - so only 25% of households had a telephone. In 1970, it was 35% of households; in 1975, it had reached 52%, and in 1981, it had reached 75% (of around 20 million).

50% of homes connected over a 15 year period.

That means connecting roughly 3.5% of households per year, or perhaps 700,000 lines: far higher than the rate of building homes.

Even if aluminium was heavily used in only 2 years, it could affect 7% of properties, and amount to about 1.4 million lines.

Looking back, I agree. The 50% could be read in relation to uplift, rather than capacity.

I'm not sure it matters to BT though ... VM have, for a long time covered less than 50% of the country, and don't get laughed at for it. If BT are only aiming at 10-12 million, should it matter to BT *which* 10-12 million they target?

Absolutely right , does my cabinet have FTTC or not, how far am I from my cabinet, what is the quality of my lines etc etc, your internet speed is the proverbial postcode lottery. G.Fast etc etc is not solving the issue, if you already have an FTTC cabinet and you get a G.Fast upgrade, great, your speed will go up, if on the other hand you are like me, on an FTTC cabinet with approx. 2km of copper cable run, it will do diddly. It does not solve the current Swiss chees effect with speeds, it does not fill in the holes, it just creates and ever widening gap between the haves and have nots. It also does not take the lines out of the equation for limiting speeds which means as speed requirements continue to grow we will be back doing this again and again and again until it is finally done properly. We will end up in the same situation we had with trains where we were still trundling out every more sophisticated steam trains decades after the rest of europe had switched to electric trains. We never seem to be able to bite the bullet and move to the inevitable until we have tried every heath robinson contraption we can think of.

It is the 21st century and it is time for a 21st century system, FTTP. Broadband is now an essential utility like electricity, but we seem to accept rubbish from BT that we would never accept from our Electricity supplier. What would you say if your electricity supplier said you can only have a lower current because you are further from the substation, your lights will dim when your neighbours closer to the substation start hovering.

Of course Fibre on demand is massively expensive when each installation is done on a one off adhoc basis, instead of systematically dropping fibre off at all premises along the way.

It is time for the government to take this in hand, and if they want to know what to do the best similar example out there is New Zealand.
No rushing to the next must have speed threshold when they realised if they gave the current monopoly money to upgrade to FTTC (think BDUK) they would be back doing it again and again in the near future. So they sat down and had a good think about it, with the current monopoly shouting about how it would be too expensive and take too long to do FTTP. Their answer was, too expensive? we are not so sure and the benefits to the economy seem to be large multiples of the investment. Take too long? How long is too long, we have plenty of time. They decided the only sensible thing was to do it properly and do it once which can only mean FTTP, and set about getting it done with the following features;
- All lines to deliver minimum 100mbps to each and every subscriber and be gigabit capable, ends any debate about smoke and mirrors over copper
- 10 Year rollout - reduces pressure on funding and engineering resources for such an infrastructure project
- Create a holding company (Crown Fibre Holdings) as the overseer, facilitator and funder of the project
- Create local fibre company in each geographical location and have an open tender process for the right to partner with crown fibre holdings to build and own the network in each area
- The objective of the tender is for Crown fibre holdings to have to contribute as little money as possible
- Crown fibre holdings is a secondary shareholder in the Local fibre companies meaning if the companies make excess profits in the future the Govt gets its money back, so not just a bung to BT like BDUK.
- A local fibre company partner is not allowed to be involved in retailing the services, ended the equivalent of the current debate about BT Openreach and BT retail.
- No more advertising up to speeds, has to be minimum speed only

The rollout is now about halfway through, with the following effects;
- The current monopoly took about five minutes to ditch it's retail operation
- The current monopoly suddenly found huge quantities of investment capital down the back of the sofa and has won approx. two thirds of the tenders as it scrambles to maintain it's market share
- It is far cheaper than the existing monopoly claimed it would be, especially for them as they have the massive advantage of existing conduits and local knowledge
- Without the current monopoly distorting the retail scene it has suddenly become much more competitive with around 90 retailers fighting tooth and nail for your business

It is a fascinating case study in how to do this properly, and it contrasts with here where we seem to do BT's bidding, and Australia where they decided to do FTTP and deliberately lock out the current monopoly and try and do it too quickly among other issues with the way it was structured and as a result have reverted to FTTC. You may have heard BT point to Australia's faltering FTTP rollout when justifying not doing FTTP, you will never hear them mention NZ.

Looking at the NZ experience the UK govt doesn't have to do much, all it needs to do is make it inevitable that a FTTP network is going to happen, and then stand back and watch as BT suddenly embraces FTTP and trips over itself to stop itself from becoming irrelevant.

Only read the first few lines as that is a very long post.

However, the idea of G.Fast when fully deployed would be to move smaller aggregation nodes closer to the premises and therefore shorten the distance of the copper without having to build thousands of new cabinets. G.Fast may in the first instance be deployed at existing cabs for "quick wins" but is not the expected end game. And if they do get to the end game then you could find you have a G.Fast node very close to your house and therefore you could then get much faster speeds.

G.Fast ultimately is not about how far you are from the cabinet but how far you are from the node.

Is there not a major problem with the New Zealand model?

Its looking at a target of 75% availability of ultrafast at the end of the 10 year roll-out i.e. 2019 and there is a plan to extend to 80%.

Can't see that going down very well in the UK, where we have over 50% ultrafast coverage already, and if commercial plans deliver could be in the 70% region in 2018-2019

NZ$2 billion (£1 billion) Government investment to cover around 1.28 million premises. Which if scaled to the size of the UK and same target of 80%, is £18bn. For 100% suggests £22bn which is very oddly in the ball park estimates BSG and others produced back in 2009-2010.

Lets ignore the guff about what G.Fast cannot do - which is all based on assumptions that cannot yet be made - and focus on this part ...

In reply to a post by ufb200:

Looking at the NZ experience the UK govt doesn't have to do much,

Your rose-tinted prose makes it look incredibly easy.

Unfortunately, it missed two incredibly difficult hurdles that the UK government will face: VM and the EU. And even Ofcom might throw itself into the problem.

The EU loves competition, but it has to be the right sort. Competing to tender on a single national infrastructure (even if done regionally) isn't enough - it loves infrastructure competition. Real, competing, separate physical networks. Try to do something that establishes one single network, and you might find a fight with the EU on your hands.

The EU doesn't like state aid either. Try to do something that establishes one single network, using any amount of state funds, and you might find an incredibly tough fight on your hands.

Leave the EU? Hmmm. We still want access to the EU market - and that will leave us tied to the same market-driven rules that apply now. Use of state aid in an unauthorised way will be the fastest way to disqualify an independent UK from trading.

These issues of competition and state aid take on an interesting turn when VM is already competitive with ultrafast speeds in over 50% of the market, and are currently investing significantly (for the first time in yonks) to extend that into the 60-65% area.

VM, in competition, won't like the prospect of a nationally-organised (and possibly state-funded) competitor, and will appeal through every court in the land and onwards into the EU. Hands will be tied. Progress will not be made.

And that, incidentally, is one key difference between the NZ and Aus models: NZ didn't have much of a functioning cable industry, while Australia did. NZ could convert from one monopoly infrastructure to another; Australia couldn't, and would need to replace all of them.

Ofcom regulates in the middle of this. It has recently swapped from the faux-competition offered by LLU, towards seeking a serious competitor to deliver fibre - without disturbing the commercial choices being made by both BT and VM. It didn't come close to suggesting either a NZ or Australian model ... and there's probably a reason for that.

The government, in the meantime, is skint. It doesn't want to pay for anything, and would far rather leave it to commercial entities to take the responsibility.

This particular government, running to 2020, can see 3 main strands to growing the infrastructure that are all also going to carry on running to 2020ish. Even if they had the will and the money, they likely couldn't achieve any more coverage growth than is already going to happen.

all it needs to do is make it inevitable that a FTTP network is going to happen, and then stand back and watch as BT suddenly embraces FTTP and trips over itself to stop itself from becoming irrelevant.

Isn't that what Ofcom is trying to do? Make an FTTP network inevitable through PIA?

Of course, Ofcom would be happiest if they brought in a third (quasi-nationwide) operator of fibre using PIA, but that BT stayed loyal to G.Fast. There might be room for three operators that way, each with separate USPs ... but BT tripping over itself to selectively deploy FTTP would quash that stone dead.

In reply to a post by ufb200:

if on the other hand you are like me, on an FTTC cabinet with approx. 2km of copper cable run

Apologies to your sensitivities, but telecoms is a game of statistics (even in NZ), and you fall on the wrong side of them.

With a statement of "If you are like me ... with approx 2km of copper" you are talking to approximately 2% of the country. You are in the middle of the "hardest to reach".

In the NBN, that would leave you covered by satellite.

In New Zealand, you wouldn't be covered by their current UFB rollout (75% by 2019), or even the extended rollout (80% by 2023). In fact, you're right on the cusp of not even fitting into the "rural broadband initiative" for 5Mbps speed - which targets 90% of those not getting UFB. NZ's overall aim is 97.8% of the country ... and you'd be just outside that.

Do you think a UK state-sanctioned model would serve you better than these role models? Better than the model they've already chosen (90% target, then 95%, then not really any further)?

If your plans paid off, and the UK converted to a state-organised monopoly FTTP network, and chose to go beyond a 95% target, you could likely expect to be upgraded in 15-20 years - after the 95% have been reached.

I've never understood why rural fibre champions always seem to demand an FTTP-only solution for the entire country. It would solve their problem, sure, but not for a long, long time.

I wouldn't describe the network as a "terrible state" personally. Bits are, overall though it's not. What they do have in their favour also is the most advanced copper test system in the world too.

In reply to a post by MrSaffron:

Care to expand on what a significant percentage of aluminium means in terms of number of premises affected?

I don't think anyone could give you exact figures. I wonder if even OR could. In this area I would estimate roughly half of all lines have a degree of aluminium in their length somewhere. So as for a significant percentage, yes I agree with that.

Edited by deleted (Thu 19-May-16 13:28:05)

I found this on Kitz, GPO Press Notice (dated 1969), which indicates that aluminium was used as an overall cost saving exercise and not just a response to high copper prices. This would imply that aluminium was used for a significant amount of time, possibly 10+ years.

Also bearing in mind it wasn't just new infrastructure that would be aluminium based but most likely all repairs to damaged cables requiring replacement as well, which would also bump up the number of affected lines.

Yes that sounds correct to me. I would estimate aluminium was used for 10-15 years. There's a heck of a lot of it about.

Seens some people are on g.fast

found this posted over on car forum

http://s33.postimg.org/lvdz55man/speedtst2.png

What makes you think this is G.Fast ? Were there more details ?
Looks to me like lots of the companies I visit in the city who have a good infrastructure. (And use BT Global Services or others for that matter)

G.fast is 300/50 *rumours?*

and yes you could be completely right

but even my fttc says its bt global services.

as further info goes nothing just a random post i'd seen while looking for information on my cars clutch.

Edited by epyon (Wed 25-May-16 18:40:03)

If I knew where it was could probably say yes or no as to whether G.fast or a FTTP connection (but upload if correct in the test) suggests it would be a FTTP trial rather than plain old native FTTP.

That could be g.fast. It would be my guess that it is.

Think about it, if it was really a BT Global services leased line then that's a very odd speed for a leased line. You can pretty much rule it out on that basis. As far as the IP address goes, well they do move IP addresses round the BT Group and you can get a few days or weeks where it may show Global Services when it's a BT Consumer connection.

Most companies in the city would mostly have a symmetric service rather than asymmetric.