Broadband, NGA, FTTC and the laws of unforeseen consequences

19:06 in guest post by Barry Forde

Yesterday, Barry Forde, the brains behind CLEO, blogged on the “WiFiPie & CHIPS… With everything” Group. Lots of people have asked for Barry to expand on his thoughts. So, here – by popular demand – is an extended version.

Going back to basics what are we trying to do? I’d suggest two things:

  1. 1. Solve the problems of not-spots and grot-spots by getting broadband to them
  2. 2. Do that in a way that isn’t a short term solution, but a route to true NGA

If Cumbria does the first of these without the second, then it will end up with a 2Mbps service. This, we all agree, will be a woefully inadequate level of service within a very short period time. What is more, it will have gobbled up money better spent on other solutions.

So, the big challenge is how to procure something that delivers both requirements within the funding available. Is this even possible?

There seem to be two options on the table:

  1. ♦ BT’s FTTC product, within which some bits of the district would get FTTC, a subset would get FTTP, and another subset would get satellite connectivity with some BET connections
  2. ♦ FTTP via commercial or community initiatives

BT’s FTTC product

In the Eden valley – the part of Cumbria I have studied most closely – BT’s FTTC product would face a number of challenges.

First, the technology uses the existing copper phone lines from the nearest cabinet (PCP) to the property. It uses VDSL2 modulation which can, in theory, deliver speeds of up to 200Mbps, but as with all copper transmission systems the speed falls off rapidly with distance (Shannon’s Law for those of you with long memories who did physics back in the days when it was still taught).

For that reason, BT refers to an “up to” 40Mbs service, which reflects a realistic view of how far they can drive over their copper. But the actual service they offer to ISPs, which is then available to the consumer, is a rather more restrictive “up to” 40Mbps downstream/15Mbps upstream one.

Originally, BT would not offer an FTTC service unless the property could get a minimum of 15Mbps downstream, but recently it announced that it would lower this target to 5Mbps. This means that BT will be able to claim that any property that it can reach with a 5Mbps service has “superfast” “highspeed” “next generation” broadband.

VDSL2 can only be run over copper originating in a street side cabinet ( the PCP you keep hearing about). If your property is connected directly back to the local telephone exchange, then BT will have to continue to service you using today’s ADSL2 technology. But they will probably upgrade the exchange to 21CN standard that is ADSL2+, which they quote as “up to” 20Mbps service.

How would this measure up in Eden?

Looking at the Eden Valley, there are 24 exchanges serving 23,000 properties. There are 91 PCPs deployed across these exchanges, with around 8300 properties serviced from them. So, out of 23,000 properties only 8300 could get the “Super High Speed Broadband “up to” 40Mbps” service from BT.

The remaining 15,000 would have to stay on ADSL2+ with its up to 20Mbps service. In both cases, properties further away from the exchanges or PCPs would get service well below the headline speed. Those currently in not-spots or grot-spots who are served directly from BT exchanges would continue to get the same service, despite the exchange nominally offering much higher speeds.

If you are lucky and you are served via a PCP you might get an improved service. For instance, if your village were 6km from the exchange but there was a PCP located in the village, you would definitely get much better service than before.  But, if you were still several kilometres from the PCP, then you would probably get 2Mbps, and you could forget “NGA” “Superfast” “high speed” service – it’s not coming to you.

I’d like to be able to tell you what percentage of properties would get high speed broadband via an FTTC roll-out, but it’s too complex a process to calculate accurately. My personal view is that out of the 23,000 properties’ no more than 5000 would get a 30Mbps service.

The other 18,000 would get a mix of speeds progressively falling away from the headline speeds of 20Mbps for ADSL2+ and 40Mbps for FTTC service. At the margins, there would probably be around 2000-3000 properties which would still be below the 2Mbps level, and these would be serviced with a mix of BT’s BET (Broadband Enabling Technology) and satellite. This would be similar to what’s happening in the Cornwall and the Scilly Isles project.

If FTTC were selected, then the approximate cost would be around £80k per PCP. This would mean £7.2m for all 91 PCPs in the Eden valley. You would also have to add the cost of upgrading the 24 exchanges to 21CN ADSL2+ service, which would cost another £5m.

Finally, if BT did do 25% of the properties using FTTP, and we were to assume that they cherry-picked urban properties, which cannot be serviced via VDSL, then the cost per property would be around £500. So, 5750 properties at £500 each gives us £2.875m.

Add those numbers together and you get a total £15m – assuming BT were generous and didn’t make any money out of the project!. At a guess they would come in at £1m per exchange, or £24m for the valley.  Its unlikely they would do the FTTP bit for the rural properties as the distances are much longer and using the Digital Britain Report figure of £10,800 per property in the rural areas the 25% would cost £61.1m.

FTTP via commercial or community initiatives

FTTP is regarded as the end game by most experts as it offers both very high speeds and longevity. Fibre optic cable is laid the whole way from a central hub (possibly the telephone exchange but any building will do) through to the property.

Two models are used, PtP fibre where each property gets a private fibre(s) between it and the hub, or GPON technology where groups of 32 or 64 properties share a fibre pair and the bandwidth available over it

PtP FTTP usually runs at either 100Mbs or 1Gbs and this bandwidth is available exclusively to the single property served by the fibre. It is also symmetrical. This is necessary for interactive services such as telemedicine and video conferencing.

With GPON it is usual to share a 2.5Gbs channel between the 32 or 64 properties which gives each one significantly less than the PtP technology, but it’s still a lot compared with copper delivery. There are versions of GPON that use a 10Gbs channel. Where independent telcos are rolling out FTTP they tend to go with PtP technology, incumbent telcos prefer GPON which has a much greater market share as a result.

Fibre is also a long term solution because modern fibre has a lifespan of at least 25 years, probably much longer, and can deliver bandwidths up to and beyond 10Gbs. It is also distance independent: out to 10Km uses very cheap optics and as the distance grows beyond that the optics are changed to higher powered ones to reach 40Km, 80Km or further. Clearly these distances are not needed for parish or district deployments. With fibre, the headline speed and the delivered speed are the same – none of this “up to” malarkey.

The issue with fibre is that it has to be laid from scratch, unlike FTTC where the existing investment in the copper is preserved.  The fibre itself is cheap, much cheaper than copper. The cost is all down to digging a trench and putting a duct in it, this is around 80% of the total cost.

The Digital Britain Report estimated the cost of running FTTP across the whole UK to be £29bn. This was broken down as around £380 per property in urban areas and £10,800 per property in rural ones. This reflects the much longer distances that would need to be dug in sparsely populated countryside compared to the terraces or streets of semis in the towns.

The RDPE pilot FTTP project around Lancaster

I have been working on an RDPE pilot FTTP project around Lancaster which is a community led build out by the farmers and others in the parishes served. Based on this work we have calculated (following a detailed design effort) that rolling fibre out to 100% of the properties across the parishes costs around £1250 per property in materials and labour costs.

The labour costs are based on farmers doing the digging across their land wherever possible, and, when this is impossible, using local agricultural labour (man plus mini digger plus mate) to do it.

The parishes of the Eden Valley are no different to those near Lancaster, so the numbers could be used for costing purposes. Assuming all 23,000 properties were fibre connected, and given that significant numbers of them are in built up areas like Penrith and Kirkby Stephen, I would expect the cost to be less than £1k per property, giving a total of £23m to do the Eden Valley.

In Lancaster, we are installing an FTTP hub in each parish (well not every but most) and connecting properties at 1Gbs to the hub. The hubs then link back to a central node with 10Gbs circuits and the central node has a dark fibre link (multiple 10Gbs wavelengths) to Telecity in Manchester where we peer and take IP transit. The cost of all that is included in the per property costs so again the overall cost for the Eden Valley is likely to be lower due to being able to spread the central costs over 23000 rather than 700 properties. Factoring all that in a cost around £20m looks feasible.

Conclusion

My guess is that FTTC would come in at around £24m up depending on how much profit BT seeks from the project. BT’s retail arm offers the BT Infinity FTTC product to consumers and charges £19.95/month for the “up to” 40Mbps/2Mbps service and £24.99 /month for the “up to” 40Mbps/10Mbps service. These charges assume you also take a BT phone line and pay the usual line rental on top of the broadband charges and take an 18 or 24 month contract.

The 15,000 users outside of the FTTC service would continue to take and pay for standard broadband as now. I imagine that users connected via satellite would have some subsidy applied to pull their charges down to match copper delivery solutions. Those using BET to get 1 or 2Mbps would again have the capital cost (£600 for 1Mbs service, £900 for 2Mbs service) met from the project and would pay standard ADSL broadband costs thereafter.

An FTTP rollout done via a Community Interest Company, based on the community offering wayleaves free of charge and doing substantial amounts of the digging themselves, would cost around £20m. For a 1Gbps service subscribers would need to pay around £100 connection fee and £30/m (£25+VAT @20%) for service. This covers running the operation, providing plenty of very high quality Internet transit etc.

It looks to me as if the two schemes would cost much the same but the outcomes are wildly different.

How could FTTP be funded?

If BDUK are putting up £5m, with the possibility of another £5m from EU funding, Eden is about half-way there for either scheme. (With the BT FTTC option, I suspect they would trim the rollout to fit the budget). For the community-led FTTP scheme, communities would have to raise the difference.

In the Lancaster model, we are working on the CIC being a company limited by shares. We will issue a prospectus and invite those living in the parishes to subscribe for shares suggesting a £1k subscription would be appropriate. The shares would receive dividends like all other shares and would have a market value equal to their face value of £1 each.

The Eden valley would need to raise £10m from a population of 50,000, so an average of £200 per citizen. If 25% of the homes in Eden were to subscribe for £1k of shares that would produce a sum of £6m. The remaining £4m could be borrowed via corporate bonds carrying say 5% interest with a 7 year redemption date.

How would the numbers then stack up? Well, assuming 25% of properties were to take up the service and pay their £25/month (ex VAT), we would have a revenue of £1.725m/pa. Against that, we have to fund the fibre to Manchester, the peering, the company overheads, the IP transit etc. This would come to around £500k/pa, leaving £1.225m of free cash flow.  A 5% dividend to shareholders would cost a further £300k (on £6m of shares). Finally, the loan of £4m would accrue repayment and interest charges of £680k/pa,  allowing for a 7-year payback with interest at 5%. So, total annual costs are £1.480m against income of £1.725m, leaving a tidy margin.

If take up were 50%, then the income would be £3.45m, but the cost base would not climb pro-rata. The loan and dividends remain the same as does the fibre to Manchester and company costs. The only change would be the IP transit, which might add another £100k/pa. So,  the net effect of higher take up would be to generate a large increase in free cash flow. As can be seen, the operation is very viable.

So the obvious question is: why wouldn’t a commercial entity come in and do it if there is money to be made? The answer of course is that the numbers only work for a community based project where the citizens put in the work and allow digging on their land.

A commercial company could not dig out using highways without pushing the cost up by a factor of twenty. Funding that sort of investment would require a much greater return from charges than is achievable.

How about a commercial company working with the community? Well I suspect that the willingness of farmers and landowners to grant free wayleaves and do plenty of work themselves would evaporate if it was a straight commercial company doing the project.

The law of unforeseen consequences

If the Eden Valley is to get FTTC, there are a number of damaging side effects which may not have been forseen by those running the procurement. First, PCPs tend to be located where clusters of properties exist. If the cabinet gets FTTC, the properties close to it will be offered the “up to” 40Mbps service at BT Infinity rates, which they will leap at given their current lack of service.

The result will be that those community-based networks currently operating or planned will have their heart cut out as their main customer base is lost. They will not be able to make a living with the residue,  and those that currently get a reasonable service from them will have to fall back upon on an inferior service from BT.

Communities interested in developing an FTTP project for their parish or village will also find that the economics have changed. The lower cost, higher density areas will have FTTC cabinets offering service, leaving the community to focus on the more remote properties that can’t get service via FTTC. But the cost per property connected will climb steeply from around £1k to something in the region of £3k, reflecting the fact that only long dig properties are in the mix.

Given a higher cost base and a reduced subscriber base the sums will not add up and community FTTP projects will be impossible. In effect the procurement will have distorted the market against community groups.

Finally, by rolling out FTTC, BT will have locked in their monopoly on the physical infrastructure. Because they refuse to supply dark fibre to third parties, anyone wishing to offer service would suffer the double whammy of a high backhaul cost, plus a subscriber base already locked in by BT. There would be no way for anyone to compete.

In a couple of years when it’s obvious to everyone that 40Mbps/10Mbps is inadequate and that what they need is a symmetrical 100Mbps service to support emerging NGA applications, where will the residents of Eden go?

BT will say: “OK we have a product, it’s called FTTP, but it’s expensive as we have to dig out to properties.” They will have no incentive to fund the digging of the last mile, since their monopoly is safe from competition and customers are locked in without the choice of an alternative supplier.

BT will then repeat the exercise of taking its begging bowl to government, claiming that it’s too expensive for them to dig out without public subsidy. Nobody else will be able to compete with the entrenched BT position following the FTTC rollout.

So in the short term, Cumbria would get a modest speed jump of, say, 30Mbps for a small proportion of the population, but many would still be on slow speed connections and the long term challenge of NGA (100Mbps or higher in my opinion) would remain unresolved, despite having blown the millions.

Tags: , , , , , , ,