Bitcoin Stack Exchange is a question and answer site for Bitcoin users, developers, and enthusiasts. It only takes a minute to sign up.
Sign up to join this community
Assuming all buyers have access to the same ASICs (ignoring that ASICs developers and manufacturers will always have the latest first, even if only used for "testing" prior to sale) energy costs would seem to promote a long term trend of mining in areas with the lowest relative costs.
I believe in the PoW model and am not interested in PoS solutions. What can be done to slow/prevent centralization in areas with the lowest energy cost?
Asking how to prevent the mining industry from being centralized into places with low electricity costs is like asking how to prevent the shipping industry from being centralized into cities that are on the coast. Both industries will tend to thrive in places where the profit margins are highest...and there's no profit margin in building a port in a land-locked city.
Since we are assuming that access to the latest ASICs is uniform for all contenders, profit margins are determined by electricity costs. However, remember that running the ASICs isn't the only thing that miners do that costs electricity. Cooling the ASICs is probably the larger electricity cost, just like it is in most data centers. Given two locations with the exact same electricity cost, but different average temperatures, the location with the colder temperature would prove to be more profitable. It would cost less electricity to cool the ASICs.
Now the problem seems to be mining centralizing into locations that have both low electricity costs and low temperatures. However, the second factor does create a variation in the trend. There are some locations that may have higher electricity, but are so cold that they don't have to use that much of it to cool the ASICs. There are other places that may be warm, but have such low electricity costs that it doesn't matter.
I'm fairly sure your question doesn't have a simple answer, but be reassured to know that there is at least another factor that will affect the trend.
One possibility among others to encourage mining decentralization is to rely on already owned/amortized hardware to mine the currency. Most people will not by ASICs to mine bitcoin, but many already have a computer, so the jump to mining is easier if that computer is somwwhat plausible as a miner. For bitcoin, ASICs are so far ahead that mining with a desktop computer is pretty much pointless.
Other algorithms have been developped to try to prevent that ovwewhelming domination by ASICs.
Early attempts include a so-called memory hard algorithm (for instance, scrypt used by Litecoin), which try to ensure that parallelization of several hashes require a linear increase in memory. In practice, scrypt failed at this as shortcuts exist allowing a memory/speed tradeoff. Newer designs such as Cryptonight are better in this respect, as there is no such tradeoff loophole. Additionally, Cryptonight make use of the AES-NI hardware on modern CPUs to ensure CPUs stay in the race compared to GPUs and ASICs. State of the art is Cuckoo Cycle (https://github.com/tromp/cuckoo), which solves some of Cryptonight's remaining problems (slow to verify).
Other approaches exist, such as chaining different hashes together, but they mainly fail as introducing complexity and potential vulnerabilities in using cryptographic primitives in ways that were not intended.
The cost of securing via proof-of-work is fundamentally linked to the cost of power.
This leaves you with the following options:
edit: The price of power for a miner is not known to the blockchain, and cannot be known currently without utilizing off-chain services. Thus, any algorithmic attempt to correct for the differences in power cost experienced by miners will require an off-chain solution, which is antithetical to bitcoin in general. If you could come up with a secure, decentralized protocol to determine the true cost of power for a given miner, you'd be getting somewhere. The reason that sounds hard is because it most definitely is! The arrangement made by the purchaser and the power company would be have to be done publicly on the blockchain in order for this sort of thing to possibly work. We are currently a looooong way from paying for power through the blockchain.
Build easy, literal "one-click" mining (GUI) into the end-user Bitcoin software, such that anyone running a full-node can instantly (or on a schedule) engage in mining activities. Since there is virtually NO profit incentive for this (given mining difficulty), it will be a purely preemptive measure to allow end users to easily and instantly mine in the event that centralization levels ACTUALLY become dangerous, OR in the event of a collapse in difficulty for whatever reason. In the latter case, the incentive again becomes an obvious monetary one. In the former, the incentive is general desire to protect the network from the ills of either possible scenario.
