Usually, the solution is to categorise different sources of power according to the time they take to respond. For example, coal and nuclear stations can crank up more power if necessary, but they need hours to days’ notice to do so. Hydroelectric and natural gas stations, however, can react in minutes. So the first category — baseload generators — provide bulk power and are paid a bulk price. The second category are kept on stand-by until there’s a surge in demand, but then they earn a premium price. And unfortunately, you can’t turn on the wind in the event that Drax is doing urgent maintenance or a million kettles just went on line.

But there’s no reason in terms of economics or physics that you have to respond to a shortage of power by adding more power. It would be just as good to cut demand. Of course, this is always the ultimate reserve, since when there is no more capacity, some people will just have to put up with a power cut. But those are highly inconvenient, especially because they are total, unannounced, and unpredictable.

What if it was possible to shift demand from the peak to the trough? Not all power-consuming activities have to happen now. Anything, for example, that uses a lot of electricity to make some sort of storable product could schedule its production to avoid the peak; in fact, if electricity is a big chunk of costs, it would make sense to do just that in order to benefit from lower prices. But the grid can’t count on this. Further, the benefits of “shaving the peak” don’t just accrue to the people who do it - the grid itself saves on buying ‘peaker’ power, and society benefits from avoiding power cuts. Why change your processes to help the electricity company? Similarly, it would be great if ISPs’ users wouldn’t leave their P2P clients running during the spikes in network traffic, instead pulling all that video in the middle of the night when the network is quiet. But who will do this just as a favour to the phone company?

Also, there are serious transaction costs - only the very biggest customers can take enough load off the grid by themselves to make it worthwhile for the power company to negotiate a contract with them and make the necessary management arrangements. That’s OK for the steelworks, but it’s not OK at all for, say, the supermarket lighting system…or the cold store. Cold store? In the Netherlands, they’re addressing the problem of using really large-scale wind power just like this. The refrigerated warehouses on the docks at Rotterdam are full of things that need to be kept cold - but don’t necessarily suffer from too much cold.

So when the wind blows, and there is so much electricity the grid needs to find a way of getting rid of it, the refrigeration systems are signalled to crank up, soaking up the surplus. When the wind drops, and the spot price of power rises, they know that it will take many hours for the temperature to rise far enough to be a problem - so the refrigeration goes off line, taking an equivalent quantity of demand off the grid. Electricity has effectively been stored for a windless day — with 100% efficiency, as the saving of power on the windless day includes all the power that would have been wasted, were the system running.

This could be done with every fridge in the country. But mutually advantageous trade isn’t happening, because the costs of transactions are too high and information is too dispersed. The more load an actor in this market could take off the grid, the more valuable this would be to all concerned. So, as with most of these situations, there are increasing returns to scale. If some sort of third party could aggregate large numbers of power users, they could trade with the grid just as the owner of a power station can; not only would the users benefit from lower average power prices, but the third party would also be able to get paid for the ‘negative watts’ they supplied, some of which they can pass on to make it worth the users’ while. It’s a classic two-sided market.

The analogy with Telco 2.0 is pretty clear. But here’s where it gets really interesting; to make “demand response” work, you need the ability to monitor thousands of customer premises’ electricity usage in real time, and tell them to cut back according to rules based on the price of power, the time of day, and the customer. The signals must get through. And you’ve got to be certain of the identity and location of the loads - you definitely don’t want free-riders claiming to cut back and then hogging the juice anyway, and God help you if you send a CUT USAGE 50% message to the liquid hydrogen factory rather than the casino.

You’re going to need something with a reliable database of users, location-awareness, secure and reliable messaging (not like these muppets, eh. Haven’t they heard of SS7?), and a billing capability…isn’t this starting to sound like a job for Telco 2.0? As we’ve often pointed out:

• it’s about deploying all your assets, including field engineers,
• it’s about assets at the network edge, not just fancy switches in the core,
• it’s about enterprises, not just consumers, and preferably both joined together,
• it’s about signalling, rather than bearer traffic, and
• it’s about messages that carry a very high social or economic meaning.

EnerNOC’s data traffic is essentially SMS, but they can currently supply 1.5GW of electricity to the Californian power grid in a pinch, without fuel, without emissions, at a moment’s notice, whilst simultaneously helping downstream customers to cut their electricity bills.

PS - Today’s oil price is just under $130/barrel and predicted to rise over the next year. Not only that, but things happen - suddenly.