Ever considered mining bitcoins? Well, as a business with 3,000 servers, IDrive decided to look into the idea to see if bitcoin mining was an option during non-peak times, a possible untapped revenue stream. We learned some valuable bit(coin)s of information; here’s what happened…
Most of our users backup their files at night, so our non-peak times are during the day (between 3am and 4pm). During this time, there is unused server capacity that can be used for other operations. Utilizing independent resources (since bitcoining is decentralized) such as the source code on Github, forums on Bitcoin Wiki, bitcoin.org, and the bitcoin simulator on coinplorer.com, a dedicated team of our best engineers spent a week testing the possibility and found that using our servers to mine bitcoins would affect our business model in four main areas:
Although IDrive owns 3,000 servers, we used 600 (each with a quad core processor operating at 2.8 GHz) for our simulated tests. Our study projected a year of mining at 100% processing power 24/7 and the assumption that the difficulty of mining (the calculating of hashes) would increase linearly. (We also assumed that there would be no additional charge for operating our servers consistently at full power for an entire year.*) Since mining bitcoins is designed to take a certain amount of time to ensure legitimacy, as more people get involved and more power is added, the cryptography is made more complex to slow down the process. Therefore, as time progresses, more power will be needed to mine (and, thus, more airflow for cooling is also necessary). As a result, it will take more energy to mine fewer bitcoins, with the energy needed consistently increasing.
*IDrive pays a set rate for a certain amount of electricity which we use to power servers. Mining bitcoins uses energy which increases our power expenditure. In our research, the amount of energy used for mining (during non-peak times) never exceeded our set amount, so this particular factor did not cost more. However, there are two other factors to consider. First, the servers usually operate at about 5% of their capacity during non-peak times; while mining bitcoins, the servers would operate at 100% the entire time. If you pay for the amount of electricity used, this increase in energy would cost more. Second, the increase in server usage could potentially wear out the servers faster, as they are operating at a higher capacity for longer amounts of time. This could shorten their lifespan and require replacing servers more frequently, which is another added cost. Lastly, it is possible that we could eventually exceed our electricity usage limit, which would result in our servers being shut down.
As mentioned before, it takes a longer time to mine a single bitcoin as more people get involved and the complexity of the cryptography becomes more difficult. Another point we found valuable is the fact that the time spent figuring out the details of the mining process detracted from time that could be spent focusing on our core services.
The value of bitcoins is continually fluctuating based on the market. The ROI would be unpredictable, not to mention the price paid to bitcoin miners will decrease as time goes on since there is a cap on the amount of bitcoins that can be created. Process, energy, complexity, and time will continue to rise as the ROI drops. In our simulated study, it would take a year to mine a total of 0.863 bitcoins. Using the current exchange rate, that would equal around $550.15 USD. As time went on, mining one bitcoin would continue to take longer and longer (while also using more energy). Further, considering we wouldn’t be mining during peak hours, that would cut our processing time in half, doubling the time to make a single bitcoin. So, actually, one year would result in .4315 bitcoins and $275.08 USD.
Our servers are not currently optimized or configured to mine bitcoins. Running bitcoin software on our servers would require installing the bitcoin daemon on each of them as well as re-opening parts of our network infrastructure (that we’d previously locked down) to enable the bitcoin network to periodically access and talk with our servers. We do continuous network security audits to ensure that we don’t have any areas of our network open that do not need to be. Opening up our network for something non-essential, like bitcoin, seemed like an unnecessary security risk.
In the end, we learned a lot about the interesting process of bitcoin mining, however, for us, the pros did not outweigh the cons. So, IDrive decided to stick with that we do best.