[Cryptography] One Bitcoin Transaction Now Uses as Much Energy as Your House in a Week

Karl gmkarl at gmail.com
Tue Nov 7 21:51:09 EST 2017


On 11/8/17, Benjamin Kreuter <brk7bx at virginia.edu> wrote:
> On Wed, 2017-11-08 at 00:25 +0000, Karl wrote:
>>
>> I’m sorry, but this report of energy use is not the metric you are
>> referring to.  Mining energy is used to secure the entire bitcoin
>> network. It would be exactly the same if there were no transactions
>> at all as if there were a quintillion transactions per second.
>
> That is also wasteful.  Compare that to the energy needed to secure a
> bank.  If the volume of transactions being processed grows, the energy
> that is needed to secure the transactions will also grow; if the volume
> shrinks, the energy needs shrink.

That is a failure of the banking system, not of Bitcoin.  Bitcoin can
secure an unlimited number of transactions for a community-determined
energy cost.  This scales far better than having to provide more
security infrastructure as the number of transactions grows.  It's
efficient to have a constant solution, not wasteful.

>> Rather, this quantity should be compared to the combined energy usage
>> of all the financial security systems of the world combined.  All the
>> security cameras.  All the guards, bank tellers, and PoS
>> machines.  All the armored trucks moving cash around: manufacturing,
>> driving, fueling, and maintaining them.  All the 24/7 ATMs
>> running.  All the constant financial infrastructure that you can
>> think of, and all the energy involved in sourcing and processing the
>> materials for it, running it, maintaining it, and repairing it.  Not
>> just the transactions.
>
> OK, but then for Bitcoin you would have to include all the Bitcoin
> ATMs, exchanges and their energy needs, and the materials, energy, and
> physical security surrounding the supply chain for mining hardware.
> Even then, you would still not have an apples-to-apples comparison if
> you did not account for the scale at which Bitcoin operates versus that
> of Visa or ACH or Swift.

Yes, exactly.  And because the decreased infrastructure needed to
operate it, this will be much less.  And, unlike Visa or ACH or Swift,
the cost of securing Bitcoin is completely unrelated to the number of
transactions being made.  It doesn't change with scale, and it handles
transactions for the entire world.  So the comparison is actually
valid!

>> The energy used by bitcoin mining is directly related to the price of
>> bitcoin divided by the price of electricity.
>
> And the price of Bitcoin will approximately track the number of
> transactions being processed, at least as Bitcoin currently works, as
> demand will increase when transaction volume increases (maybe some
> other technology will change things and we can revisit all this if and
> when that happens).  So we should expect to see, and in practice have
> seen, the mining effort increase as the transaction volume grows.

No.  The price of Bitcoin is not related to the number of
transactions; in fact, most people are using Bitcoin as an investment
rather than an exchange medium.  A quick glance shows that over the
past year the blocksize has gone up by around 60%.  Meanwhile, the
price has risen by 800% !  The price and mining energy have
exponentially risen, but the transaction volume is rising only
linearly, and slowly.

>> As electricity becomes scarce, the energy requirements of the bitcoin
>> network fall in synchrony.
>
> Not entirely; the energy requirements of Bitcoin must at least equal
> the effort attackers are willing to devote.  What if the reward for an
> attacker exceeds the mining reward?  What if the attacker is willing to
> steal electricity i.e. if the attacker does not actually pay for the
> energy they use in their attack?

The price of compromise of Bitcoin is very high.  An attacker needs to
individually use as much energy as the entire network combined to
produce fake transactions, which may then be noticed.  Alternatively,
they could mine blocks to secure the network, and produce as many
block rewards for themselves as they want, without anybody getting
upset at them.  At the moment, this would provide them with about
$12.6 million USD, _per day_.  To secure the network instead of
compromise it.

The Bitcoin developers can directly control the energy usage of
mining, and hence the energy usage required by an attacker, by
adjusting the block reward, which as far as I can tell is where the
security trade-off you discuss comes from.  So, if everybody agreed
that Bitcoin was using too much mining energy, developers could simply
lower the block reward and the energy use would decrease
proportionally.  Security would decrease too!  So if this energy use
for you is not worth the security provided, you should make a case for
this.

The block reward is decreased on a regular basis, anyway.  So energy
usage would exactly halve in 2020, except that the price will also
rise due to decreased supply, raising the energy use at the same time.

>> They rise only when bitcoin has enough demand to warrant paying for
>> them.  These things are not true of conventional financial systems.
>
> Right, because for the conventional financial system, the incentive is
> pretty consistent: banks will always have an incentive to *reduce*
> their costs and become more efficient.  If Visa could halve the energy
> it spends on security and still securely process the same volume of
> payments, there is no doubt that they would do so.  On the other hand,
> the existence of a newer, more energy-efficient mining technology would
> not actually reduce the energy consumption of Bitcoin.

That specific scenario is correct, because that would also reduce the
energy consumption required for compromise.  But if a new technology
came out that is more secure and efficient than the current double
sha-256 proof of work, Bitcoin could switch to this new technique,
agree to decrease the block reward, and have the same degree of
security for a smaller energy usage.  Which the miners are
incentivized to encourage, because their excess energy availability
would give them increased profits until the network slowly adjusted to
compensate.

The miners in bitcoin are also incentivized to directly reduce their
energy costs to maximize their profits.  The article references a
coal-powered Bitcoin mine; but that mine is not making as much profit
as the solar-powered mines which do not have to pay for coal.

This discussion seems more about opposing opinions rather than facts.
This is probably because the article itself provides very few facts.
It does not compare the energy usage of Bitcoin security with that of
the global economy (bitcoin is likely _less_ energy expensive overall,
even if projected to a comparable scale), and it depicts the numbers
in a way that is not conceptually meaningful, but rather skews them to
appear large.

Karl


More information about the cryptography mailing list