I’ve just spent the last few weeks trying to suck the cloud through a very thin and broken straw – my mistake, trying to use the Internet in the City of London.

The proposed telephone tax that came out of the Digital Britain report (and failed to find its way through parliament in the finance bill) was supposed to guarantee ‘fast’ broadband access across the country. What this meant in practice was a 2Mb/s (max) ADSL connection for rural areas that presently don’t get any broadband at all. I think the powers that be have been looking in the wrong place. If we want a strong digital economy then how about decent connectivity at a fair price for business users in our cities?

Right now there seem to be three choices out there:

1. xDSL

Digital Subscriber Line technology comes in two different forms:

ADSL

Asymmetric services usually have far greater download speeds than upload, which kind of presumes a web browsing workload (where requests are small in comparison to responses) rather than something where traffic is equal in both directions like email or voice over IP (VOIP). Using VOIP phones was part of my plan for the recent office move, so I knew that I’d need a decent upload capacity (say 1Mb/s) so an ADSL2 service that offered 20Mb/s down and 2.5Mb/s up looked like it would be barely adequate, and reasonably priced at £20/month. The best bit was that it could be installed quickly.

Sadly the install didn’t go so well, and we had a lot of hassle getting any connection at all. Worse still it turned out that our local exchange (the one that basically covers most of the City) couldn’t actually do ADSL2, so we got regular ADSL – 8Mb/s down and only 500Kb/s up (best case) – nothing like what I needed for the phones. It’s simply staggering that the exchange infrastructure in central London hasn’t been upgraded, but somehow that’s the case. I suspect two things come into play here – 1. very few residential properties in the area (and hence little pressure from Ofcom to meet consumer targets) and 2. a desire to get business users to buy more expensive services (of which more to follow).

In the end it took 3 weeks to get the service working properly, and I’m grateful to my friends at BT for pushing things along and providing the help needed to navigate retail, wholesale and OpenReach. It seems that the culprit was a dodgy (BT supplied) 2Wire router – I guess it’s always the last thing that you change that fixes things.

With the worst of the jitter and drop outs behind us ADSL is probably sufficient for a lot of the Web/SaaS stuff that we consume, but more than a handful of phone calls was going to remain a problem.

SDSL

Since I can’t get the upload speed that I need for VOIP with ADSL (and can’t get ADSL2) I’ve had to order an SDSL service. This is seriously not cheap at nearer £200/month for 2Mb/s (and the engineer I saw whilst troubleshooting the ADSL woes says I’ll be lucky to get more than 1Mb/s out of that since the exchange is so far away). There are cheaper services out there, but they all use SDSL-M, which is essentially the same tech as ADSL2, and thus not available to me in the middle of the City.

I’m told that SDSL is being phased out soon, which brings us to…

2. Ethernet First Mile

The use of Ethernet in the First Mile seems fairly new in the UK, though I understand it’s been commonplace in AsiaPac for the best part of the last decade (my brother in law used to get a 100Mb/s service to his flat in Sapporo for £12/month about 5-6 years ago).

I got some quotes for EFM, which started at £300/month for 2Mb/s! Even worse there was a 3 year minimum contract – a bit tough when your sub lease is only 2.5 years. Costs spiral upwards towards around £500/month for 5Mb/s.

I don’t think this situation will endure for too long, and I wouldn’t be surprised to see 5Mb/s for £99/month on one year contracts in the next year or two, which is one of the reasons I decided to suck up the cost of SDSL for the time being.

For the real speed freak there is of course…

3. Fibre

For as little as £550/month you can get a part share (say 25-50Mb/s) of a 100Mb/s pipe over fibre. Of course the fibre itself can carry a whole lot more than that, and the bandwidth that you can get for your £ goes up quickly. Whilst this seems good value against EFM it’s worth looking further afield. In Hong Kong they can get 1Gb/s for $200HK/month (which is less than £20), and I expect it’s much the same in S. Korea and Japan etc. Why are the economics of fixed line connections so different in that region?

I was very tempted to go for a low end fibre connection, but spending something North of £7k a year on data seemed like a lot for a small company, and there’s another problem – installation takes about 3 months, so even if I’d ordered the moment we’d chosen the new place (and before leases were sorted out etc.) I’d still be waiting another 6 weeks. Clearly the demand is out there – it must be a nice business to be in.

Getting by

Whilst the ADSL wasn’t working we ran the office on my MiFi (and one or two other 3G cards) and mobile phones. This wasn’t great, but it wasn’t awful either. 3G certainly works better than unreliable ADSL.

Getting better

What would have helped a lot here is if the building had some of the infrastructure already in place – say a fibre termination and router. Then new tenants could be flexibly given muxes on the fibre to deliver the bandwidth they need. This could potentially be a sweet deal between the landlord and his chosen telco, and would save a lot of delay and duplicated equipment.

Why isn’t this happening?

My guess would be that there’s still far too much profit in the traditional leased line business (and fixed line voice) and that prices are being kept artificiality high to avoid canibilisation.