And I’m back! with another essay from the University of Nicosia’s Digital Currency MOOC after a brief hiatus. I haven’t posted as frequently due to a couple lectures lacking essay prompts and my falling a bit behind in the MOOC to attend the Bitcoin in the Beltway conference in downtown Washington, D.C. (my backyard and former home). I may try and write a post on the conference before too long, though I have more of the MOOC to catch up on as well. As always, you can find the rest of the Digital Currency Course Essay Series here.
For this essay, the prompt was: Recently and according to this distribution, Ghash.io has (again) concentrated a very sizeable percentage of the hashing power of the network. Why is this happening and how could it be avoided to prevent a potential “51% attack”?
Ultimately, Bitcoin mining is a business, and mining operations that are run like a successful business are the ones that will survive, earn better profit margins, and be able to reinvest those profits in expanding and improving hardware. The centralization of pluralities of Bitcoin hashing power in just a few mining pools is a natural consequence of the business conditions that mining pools lay out before potential new pool members. This comparison of Bitcoin mining pools on the Bitcoin wiki has a handy chart showing selected data about various Bitcoin mining pools, although the chart is a bit dated and does not contain information on some of the larger players currently on the scene (such as Discus Fish), it will nonetheless help illustrate some key points.
First, note that in the data, there are only 9 pools with hash rates of at least 100 TH/s: BitMinter, BTC Guild, Eclipse Mining Consortium, Eligius, GHash.IO, Multipool, P2Pool, PolMine, and Slush’s pool. Of these top 9, only 1 pool, EMC, keeps the transactions fees associated with a correct block. The other 8 pools all share transaction fees proportionally with miners, along with the block reward. Transaction fees are currently dwarfed by the block reward: fees currently total approximately 11 bitcoin per day (according to Blockchain.info); whereas, the block reward churns out a staggering 25 bitcoin every 10 minutes. Over time as the block reward halves again and again, transaction fees will become a more important incentive for miners and the pools that greedily withhold transaction fees from their constituent miners will find it increasingly harder to maintain a significant share of mining power.
Going back to our comparison of pools on the Wiki, we note that fees charged by pool operators are another key distinguishing factor. Five of the nine largest pools advertise either 0% fees for payment of shares or issue promises of no other fees, and GHash.IO, the largest pool, advertises both. One other mining pool, PolishPool, offers no PPS fees and promises no other fees. According to the ranking, PolishPool only has about 5 TH/s hashing, but language barriers may account for much of that variance. Here again, miner’s rational self interest of maximizing profit tends to centralize Bitcoin mining on the pools that promise to take the smallest cuts out of rewards for successful block creation.
Since we have identified the cause of mining centralization as a few pools offering superior pricing schemes, the solution to centralization logically follows. More pools should consider lowering their fees and offering to share transaction fees. This would lessen the economic motivation towards centralization. Of course, this is easier to articulate than to implement since mining pools also incur costs to operate and therefore must charge some fees as a going concern.
References:
1. Bitcoin Wiki, Comparison of mining pools
https://en.bitcoin.it/wiki/Comparison_of_mining_pools
2. Blockchain.info, Total Transaction Fees
BlakeChain 