Most commercial IT applications execute in places connected by optical fibre, and so a sufficiency of data communication capacity and reliability can usually be assumed. For that reason, the location of information or of processing seems not to matter any more: it may as well be in the clouds. Cloud computing, where users give up on knowing where their software is executing, is both very fashionable and mildly successful.
Some people think that the approach will work for smart energy information systems. For example, Greensmith in the US offers distributed control of storage batteries via the cloud. Up to a point, that’s fine: but the idea begins to break down when one considers that even the cloud, vague as it is, has a definite geographical extent. The utility company offices, from which people want to operate the controls, are probably always located within easy reach of the cloud. But it seems likely that many smart energy management installations will be located off-cloud, as it were; in places where the reliable and fast data links don’t reach.
Where the systems are located at consumer premises, the likely available communications channels will be the smart metering infrastructure or the domestic broadband connection. In Great Britain, it’s already becoming apparent that if DECC is to achieve the government’s targets for population coverage with smart metering, the metering infrastructure’s data rate per home may be surprisingly small; and that it will probably be kept busy by the substantial messaging demands of the supply industry. The consumer’s DSL connection will almost certainly still be what it is now: the consumer’s. Consumers use their DSL connections for whatever they like, and some will not take kindly to the idea of a smart energy business using their network capacity. Consumers also have the habit of reconfiguring their DSL, breaking it, or simply being disconnected for non-payment.
If on the other hand consumer smart-energy systems have their own communications, that won’t be cheap: either it will be wired, with a high cost of installation, or it will be wireless, with a much higher cost per bit than we’re used to now: the age of almost-free surplus wireless capacity won’t last much longer.
So while the regional energy control hubs may be in data centres with uninterruptible power and multiple fibre connections, the great majority of smart energy devices won’t be. Resilient and always-on communications capability won’t always be available where the equipment is. The information systems that look after smart energy devices, therefore, are going to need enough autonomy to operate through communications disconnects. And the localization of every bit of processing will have to be deliberately designed to fit around the capacity and availability limitations of the local communications infrastructure.