"Superfast" Surrey

Last week Surrey County Council sent out an update on their plans for spending their BDUK grant. I was struck by how unambitious their goals are, despite previous promises that we were to get the best broadband in Europe.

I would suggest you read the newsletter on their website, but it hasn't been uploaded yet - not a good sign that they are in tune with the Internet age.




One of the highlights is the result of a survey which asked the public what broadband speed they wanted. The result was unsurprising, most people said that 20Mbps would be right for them. This is a perfect demonstration of why you shouldn't ask the question in this way - responses are limited by the public's perception of what's possible and what they've heard of in the media. Without any knowledge of the technology or the possibilities, you aren't going to get a meaningful answer to an open question like this.

1Gbps SM Fibre SFP £23 retail

I suspect those questioned were assuming the price of the service would scale linearly with the bandwidth; the economies of Moore's Law are not commonly experienced to those who don't work in the network industry, they don't know that a 1000Mbps fibre transceiver nowadays costs just £23 retail (less each day by the time you read this; 10Gbps will be this cheap soon) and the cost of the cable is the same irrespective of the speed it runs at.

Perhaps they would have got a different answer if they'd asked "do you want your broadband to run as fast as is technically feasible for a reasonable price?", "do you think Internet applications in the future will be more sophisticated and more demanding on the network?" or even "do you have more than one TV in your house?" since 20Mbps is barely enough to stream one 3D HD TV channel.

The problem with planning for new technology is we don't know what applications are coming, and we don't know to what extent the limitations of today's access technology (and the way it is priced) is constricting the development of these new applications.

This is a solution that needs to be technology led, i.e. build it as fast as possible, not demand led.

What Surrey County Council should be doing

Surrey won't share with the public their criteria for selecting a provider to spend our money with, so here are some points which I hope they haven't overlooked.

1. BDUK promised "community hubs" to allow local solutions to last mile access - are these in Surrey's plan?

2. ask the providers to justify any plan that doesn't run access connections at the fastest feasible speed supported by the technology, which for fibre is 1Gbps symmetric

3. be pro-FTTH. There is a false impression that the council needs to be "technology agnostic", presumably from a civil service culture of being "politically agnostic" or "commercially agnostic" (not favouring any one company over another) but this doesn't apply to technology.  Surrey CC should immediately discount satellite based solutions as unfit for purpose, and favour suppliers who find a way of laying fibre rather than those who would spend the money on improvements to the existing copper network.
4. exclude "shared media" solutions such as PON since they can't be unbundled on a per-subscriber basis, preferring "home run" style roll-outs (exactly how the telephone lines run to your house today but using fibre)


5. look to more that "download speed" as a metric for service quality.  Broadband access networks should have symmetric download and upload speeds, the Internet isn't about "passive consumers" downloading paid-for content, despite the big telco and media companies' dreams.

6. demand a solution which doesn't transport the packets out of the local area for routing. The current BT model for broadband routes all the traffic via London which is hugely inefficient and is the reason why there is usage-based charging as these trunks and interconnects are congested.  Without local routing there is no incentive to run services locally, but local services would be better (faster, more capacity, cheaper to run.)  For example, it isn't feasible to run CCTV on broadband (camera on an ADSL line at a remote location) as the usage cost would be prohibitive - the ISP charges because they route all the traffic via London even though the user only wants the camera to be viewed from another ADSL line in the same BT exchange area.

I await to see the outcome of the Surrey process but I fear I will be totally underwhelmed by the result.

Sadly I predict the money will be given to BT on a promise they will upgrade their network with their VDSL and GPON solutions - the worst of both worlds - and they will take two years to do it.

Two years on we will lament the missed opportunity and people will say "we did the best we could at the time and no one said there was a better way".  I'm saying it now - there is a better way!