Unplugging the space Internet

Mark Handley of University College London has revealed Elon Musk's Starlink to be a leading contended for the title 'world's worst technology project' (New Scientist, 10/11/18 p 5).  The project will consume a lot of skilled effort. And it will add to climate change from the manufacture and launching of 900 comsats per year - a vast number.

And for what?

Starlink will allow some banks to gain a milisecond level advantage in that form of automated high-stakes gambling called high-frequency trading (HFT). HFT is a singularly useless activity that supports no businesses and feeds no children. And though it does make some bankers richer few of us are likely to see that as an advantage.

There are better uses for the money and effort that Starlink will consume. Musk should pull the plug now.

Good News 1: Green clouds!

In Information Technology 'cloud' is the coming thing. With cloud services an individual or business can use IT services without having to buy, manage and support the necessary hardware and software (other than your PC or smartphone). The hardware and software is the responsibility of a service company - users just use the service (and pay a fee). Just like Hotmail or Facebook in fact.

I'm mentioning this because it looks like a win for the climate according to Ian Bitterlin, whose company runs large computer centres, blogging for the British Computer Society here. Bitterlin starts from the fact that many existing corporate computer centres are rather inefficient in energy terms; it's not unusual for them to consume two kilowatt-hours of electricity for every one kilowatt-hour used by the computers, etc., in the facility. Higher ratios are not unknown.

New data centres run by specialist companies offering cloud services tend to be more efficient in both their use of electricity, wasting only 10-20%, and in getting useful work from that power. Bitterlin puts the likely savings at 74% which I find over-optimistic. But 60% seems plausible.

But that's 60% of the energy use for IT services that transfer to the cloud. That's an entirely speculative fraction of the 7.5 gigawatts used by IT. It might be 2%, or 40%, or some other number.

But whatever the proportion we can expect energy savings as it happens. 

But not, let's note, for wholly new cloud services such as Google and Facebook were not so many years ago. Qualified good news then.