Networked societies and networked economies rather depend on there being a network.

We only tend to notice them when they are not there or not working. On my wired connection at home, there is decent bandwidth with very occasional catastrophic failure. On my wired connection at work, there is what feels like very limited bandwidth shared among a very large number of people – the electrons occasionally approach the speed of sound but rarely, it feels, the speed of light.

Then there is my mobile connection.  My phone often tells me it has a 3G connection, but in central London, it only does that when there is no actual data being moved about. The second I try to use the connection, it degrades to something dramatically slower. Not infrequently, it degrades to nothing at all.

The internet, of course, was famously to interpret censorship as damage and route around it. The trouble is that, in some very important ways, it can’t even treat damage as damage. If the final connection is missing, the health of the rest of the network is slightly beside the point (the main backbone is not as resilient and redundant as it looks either, but that’s another story). This has been a long running issue for the so-called final third: the mostly rural areas where there is not yet any semblance of any kind of broadband, with an active campaign to put the final third first. But it is also increasingly acute in the health of mobile networks, where the sudden growth in data dense traffic has left the operators floundering. There is an interesting article in the current New Scientist (but sadly behind a subscription wall) which argues that it is pretty much impossible for supply to meet rising demand in the short to medium term

Data gobbling smart phones are of course the source of the problem, as they overload networks with requests for web pages, email and video streaming 24/7.  If the use of these devices grows as expected, cellphone networks across the world could grind to a halt by 2013 – and since many core services depend on wireless communication, the results could be devastating. The only solution will be an overhaul of the way mobile communications are delivered.

Of course malthusian predictions of the doom laden consequences of growth have been with us since, well, Malthus.  His main claim to fame these days is to have spent two hundred years being wrong – though the past is no more a guide to the future in that than in anything else.  But this does serve to underline that the service visible to customers at the top of the stack is the result of the complex interaction of many factors, not all of which are susceptible to rapid change. The shape of the 3G network in the UK was set by the process which culminated in the spectrum auction in early 2000, a time when nobody had the faintest idea what all this bandwidth might be for, beyond some some vague and misplaced thoughts about videophones and football highlights. The technology due to be implemented over the next few years to squeeze more capacity into 3G is the result of concerted development which itself was well underway before smartphones were anything but a rarity.

There is already some very visible frustration around. As I write this paragraph, Paul Clarke is tweeting:
[blackbirdpie id=”29604310211″]
With some trepidation, I wonder whether he is slightly missing the bigger picture. Perhaps between London Bridge and East Croydon there are other operators with better signals, but I suspect everybody on every network has had the experience of sitting on a train looking glumly at their phone with no signal while else is chatting merrily away. Of course, often it is the very fact that everybody else is chatting merrily away which blocks the next connection. I have frequently marvelled that Vodafone does not seem to have spotted that the presence of steel rails stretching for long distances away from London indicates the predictable location of large numbers of people who are somewhere between keen and desperate to make contact with the world outside their train, but in my less frustrated moments I am ready to recognise that keeping somewhere between dozens and hundreds of people connected in a small dense clump through what is presumably at least a partial Faraday cage moving at over a hundred miles an hour is probably not the easiest engineering challenge to have to solve.

The selling point of online services – from Amazon to tax discs – is that they are better, faster and cheaper than the online alternative. That’s a very powerful selling point, because they are all three of those things. The customer experience, though, is a product of the infrastructure as much as it is the service which runs on top of that infrastructure. For an increasing number of people, the quality of mobile connections is one of the critical drivers of their service experience. Ofcom research shows that:

In the early stages of mobile broadband take-up, most people used it as a complement to an existing fixed-broadband service. However, by Q1 2010 there are some indications that more households are using mobile broadband as their only internet connection. Ofcom research finds that 60% of mobile users also had a fixed-line connection in Q1 2010, compared to 75% a year previously (see Figure 5.15), and our research suggests that the number of households which only had a mobile broadband connection doubled from 3% of all households in Q1 2009 to 6% of all households in Q1 2010 (note, however, that this should be treated as indicative only, as there is a margin of error associated with this consumer survey research) . With fixed-line broadband levelling off at around 65%, it appears that the growth in overall household broadband take-up (up 68% to 71% in Q1 2010) is now being driven by households getting online for the first time via mobile broadband, mainly by purchasing lower-priced contract plans or pre-pay offerings, but also potentially by purchasing a computer for the first time, with a mobile broadband tariff that includes the price of a laptop or netbook PC within the monthly contract. [my emphasis]

The power of the network depends on the power of the network. Perhaps we need a bit less attention on the first and a bit more attention on the second.

And as a footnote to all that, the same edition of the New Scientist suggests that we may not need to worry too much about the network congestion caused by the proliferation of devices, if we run out of the raw materials necessary to make them in the first place, though their article concludes in a splendid piece of technocratic optimism, quoting Kazuhiro Hono of Japan’s National Institute for Materials Science, “The important thing is to recognise the importance worldwide,” he says. With efforts focused on innovation, he adds, “the solution to this problem will come out in the future”.