Networks
March 2019
The garage rang the other day to say that my car had told them it needed some attention; it had gone behind my back and sneaked on me. Odd as that may seem, it is all part of the ‘Internet of Things’ (IoT), a meaningless phrase which it is jargon for the way that all sorts of unlikely items are becoming connected so that information can pass between them.
IoT has considerable benefits; in manufacturing, it can link production lines with suppliers; in agriculture, machines in the fields can continuously report the conditions all over an estate, allowing much more accurate planning and use of farm machinery.
In Spain, the city of Santander is trying out a massive IoT experiment. They are encouraging all inhabitants with smartphones to download a widget which then links to over 12,000 sensors around the city. This combination delivers very accurate information about where people walk and shop, how they travel, even how they use water. The sensors can report when the waste bins are full and when the grass in the park needs watering; it should all make urban planning and management much more efficient. That’s the plan, anyway, and other towns are watching: Guildford is already testing ‘smart’ traffic lights that respond to traffic.
All this sounds fine, but I have two major concerns that I suspect we will only really sort out by trial and error.
The first is privacy. Whilst it’s true that the more information we have, the more informed our decisions might be, but who owns that data? Some of it may be about you; do you want to take part, however anonymously? It’s perfectly possible to argue that collecting mass data is irreconcilable with any notion of personal privacy. I worry that the current approach generally assumes that all data collection is a good thing, and that the time to worry about the moral and security aspects is later.
My second worry is more technical and concerns the sheer scale of the network that is being created, and hence the scale of what can go wrong. Remember the first rule of computing: what can go wrong, will go wrong.
Once everything from waste bins to jumbo jets are part of a network, there is the possibility of an unintended but spectacular domino effect. There was a good example of that last year; you probably read, or even experienced, the collapse of the O2 mobile phone network in December. However, it didn’t just affect mobile phones but many other things that use the network; that included payment terminals, displays at bus stops, satnav systems and much more.
Despite rumours at the time that evil foreign hackers were to blame, it was actually just a stupid mistake. A bit of software at the heart of the system was not renewed on schedule, so shut itself down as a safety precaution, just as it had been programmed to do. The whole mess was caused by good old human error, nothing to do with shadowy enemies of the State.
But the effect was just like the row of dominos; one bit of a network fails, which causes another to be blocked and shut down, which trips over a third, and so on. Remember those traffic lights in Guildford? It’s not hard to imagine them being caught up in the muddle.
The O2 failure demonstrated how dependent we have become on such networks and proved how important it is that systems that work in ‘the cloud’ must have a Plan B for when they fall over or when the humans in charge simply get themselves into another fine mess.
We do, after all, reap what we sow, even if we used Satnav and IoT to plant the seeds.
A few links:
