But this cost caught me by surprise. NewScientist (4 November 2017) reports on how much energy it takes for the “mining” that supports Bitcoin, and it’s non-trivial:
We have known for a while that bitcoin hogs energy. That is down to the way it works with the blockchain. Each transaction starts with a user broadcasting the details of that transaction to a network of linked computers, where it is duplicated in thousands of identical, unfalsifiable ledgers. “A blockchain, including bitcoin, has to operate on the assumption that no other computer can be trusted,” says Teunis Brosens, economic analyst at ING. So instead of trusting anything, each computer independently verifies part of the transaction, in a process called mining.
Mining prevents computers creating fake ledgers. They need to show “proof of work”, a gruelling cryptographic puzzle that takes so much processing power that generating false entries becomes prohibitive.
All that processing guzzles a lot of electricity. That’s still peanuts compared with the energy use of the internet, but one recent estimate put the annual electricity consumption of bitcoin mining at 23.07 terawatt hours, roughly the amount of electricity used by Ecuador each year.
I had no idea it was that high, and it really calls into question the scalability of this approach to electronic currencies. As someone who’s taken on scalability and performance issues a time or two, this doesn’t sound like it’s solvable simply because it’s designed to make false entries far too expensive.
That is, you solve the scalability problem and your trust level will deteriorate something approaching zero.
So, as NS reports, a new approach is being evaluated:
The latest solution is a radical one: change the way blockchain works altogether. Vitalik Buterin, the creator of cryptocurrency network Ethereum, announced last month that he would adopt a completely different way of doing transactions, known as “proof of stake”.
He adds his voice to a chorus who think that instead of proving a computer is trustworthy by taking out a “proof of work”, they could vet themselves by placing a small amount of money into a fund, which they get back if the validation turns out to be authentic, says Brosens. In a similar way to proof of work, it is difficult for fraudsters to replicate.
I’d hesitate to jump right on that bandwagon until it’s been fully evaluated by people smarter than me. Experts. By which I mean computer-savvy criminals.
