Mining Experiment: Running 600 Servers for a Year Yields 0.4 Bitcoin

9 comments

bitcoin-idrive

Online backup provider iDrive (mascot shown above) conducted an experiment in which it used 600 quad-core servers to mine for Bitcoin. The result: Stick with custom ASICs. (Image: iDrive)

Can data centers tap unused server capacity to mine for Bitcoins? The question occurred to the team at the online backup service iDrive, which performs most of its customer backup jobs overnight, leaving its 3,000 quad-core servers idle for much of the day. So the company ran a test with 600 servers to see whether Bitcoin mining could become a secondary revenue stream.

The result: running Bitcoin mining software on those 600 quad-core servers for a year would earn about 0.43 Bitcoin, worth a total return of about $275.08 at current prices on major Bitcoin exchanges.

“Its a waste of time, so any other company thinking about mining with their infrastructure, learn from us,” said iDrive’s Matthew Harvey. “Don’t do it. You need custom machines to effectively mine bitcoins and generate a real ROI.”

The iDrive test-drive reinforced a common theme on Bitcoin mining forums: To earn money by mining, you need to invest in highly-customized computers using ASICs (Application Specific Integrated Circuits) to crunch data for creating and tracking bitcoins.

Bitcoin is sometime referred to as the “Internet of money,” a platform using cryptography and software to offer an alternative currency and payment-tracking system. At its heart is a huge distributed computing network that verifies each transaction. Participants in this online ledger are rewarded with new bitcoins, which are issued about every 10 minutes.

Miners Upgrade to Powerful Hardware

Over the past year, the computing power supporting the bitcoin network has soared. The cryptocurrency is now supported by a powerful global network backed by 150,000 petaflops per second of computing power, roughly 600 times the combined power of the all the supercomputers in the Top500 list. Practitioners of Bitcoin mining – the term for using data-crunching computers to earn newly-issued virtually currency – are adopting more powerful hardware, pooling their efforts and seeking to slash their power bills.

Most serious Bitcoin miners have graduated from CPUs and GPUs to specialized chips such as FPGAs (Field Programmable Gate Arrays) or ASICs that can be optimized for specific workloads. This has led to the emergence of a new class of hardware vendors selling custom hardware for bitcoin mining.

The horsepower required to succeed in Bitcoin is highlighted by the iDrive simulation, which used 600 servers.

“Our study projected a year of mining at 100 percent processing power 24/7 and the assumption that the difficulty of mining (the calculating of hashes) would increase linearly,” iDrive noted in a blog post describing its experiment. “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.”

They’re not the only ones who’ve contemplated repurposing powerful equipment to pursue cryptocurrency, The 4,000-core Odyssey supercomputer at Harvard was secretly used to mine Dogecoin, the ironic virtual currency used primarily for online tipping. The covert miner has had their computing privileges at the university suspended.

About the Author

Rich Miller is the founder and editor-in-chief of Data Center Knowledge, and has been reporting on the data center sector since 2000. He has tracked the growing impact of high-density computing on the power and cooling of data centers, and the resulting push for improved energy efficiency in these facilities.

Add Your Comments

  • (will not be published)

9 Comments

  1. Hello Richard, I found this article also listed here: http://newsbtc.com/2014/02/19/online-backup-service-unsuccessfully-tries-mining-bitcoin-600-servers/ Anyhow, I have reached out to their team and expressed that they need to continue this experiment and instead try cpu-only Altcoins such as Quarkcoin or Primecoin. Their servers would be better suited for that task and could collectively make 10x+ on a cpu-centered coin. NOT LTC, NOT Dogecoin, but a CPU specific coin.

  2. Dustin

    They should use the servers to mine a scrypt based coin. As there are no ASICs available for those at this time, CPUs can actually compete. GPUs are still better, but with 600 servers, they could have a decent amount of income. I believe that a decent processor can pull 5-10 KH\S, which would make it 3-6 MH\S with their servers. At that rate, they could make $20-$40 per day from mining altcoins.

  3. Having modeled this out twice - once at small scale and once at megawatt scale, it is EXTREMELY difficult to mine and make money. There are many reasons for this - some I put in a two part blog post http://wp.me/p3Vlt2-5Z - because of the heat from the rigs, the cost of electricity, facility shortcomings, and the way bitcoin works. A miner never makes as much as they do the first month. So the longer they are a tenant, the more risk an operator has they won't get their rent paid. If the tenant is a mining company that signs a lease but has their customers bring their own rigs and the tenant defaults, then the landlord tries to put liens on rigs their tenant doesn't own. Messy. I haven't given up hope that mining could be profitable but with the current rigs, current options of facilities, and current costs of power combined with the nature of bitcoin & diminishing returns after the first 30 days, it'll be quite the riddle for the immediate future.

  4. Jeff

    What a facepalm. I can't believe they thought it was a good idea to do this, let alone publish the results. If they have a cyclic workload for bandwidth and cpu, why not try to match it with some similar activity that's out of phase, by provisioning the servers as a CDN to run during the day (or whatever their idle window is) to do something like video streaming? Netflix for one is apparently willing (after the recent announcement) to spend money on well placed capacity.

  5. Any nerd on the forums or IRC coulda told you not to mine from cpu servers :)

  6. cpuman

    It really took them a year to realize this would be wildly unprofitable?

  7. Time

    This ultimately would be bitcoins downfall. The amount of energy is wasted. Because to verify future transaction would take even more processing power. The world won't have enough energy to keep it going. But it came at the right time, to serve a right purpose. That is to offer an alternative to currencies which take no energy to create, thus does not store much value over time. But once trust is restored back to world currencies, once it's backed by something tangible and a store of value, then bitcoin will have to struggle with it's design. That it's very power hungry and thus can be considered a waste of energy. Which goes back to gold. Gold and other precious metals are so rare in the universe because the amount of energy it takes to create them is infinite and that energy only exists for a split moment in time. Gold acts as a store of energy and thus a store of value. Not only the energy it takes to mine it, but the rare universal energy that was used to create it. It doesn't wastes energy to continue to exists the way bitcoin does. However currencies take no energy to create more of, and would not be a good store of value over time.

  8. Michael - totally agree. I figured it out with a friend on Skype in 20 minutes. CPU's do too many things besides crunch numbers and is like usiing a cement truck to run errands in London. And 'Time' it is good to have an existential view on things sometimes, however predicting that we don't have enough energy to support Bitcoin is heavy on opinion, we only have to wait until 2030 until no more Bitcoin can be made/mined. In 16 years it is my opinion that there will a technology that isn't even conceived that will rear its head in 7-10 years and and fix the problems, if Bitcoin doesn't implode first. This is an excerpt of an email exchange I was having with the LA Times: Bitcoin is an interesting proposition depending how you look at it. One one hand, Bitcoin is interesting in that it is decentralized, transparent, global, community controlled, and anyone can create the currency. On the other hand it smacks of a different scheme in that it is tightly controlled by the early adopters/users, reduces the proliferation and creation over time, and makes it increasingly difficult to mine and cover costs - http://wp.me/p3Vlt2-5Z - so as time marches on there is less incentive to make and ultimately use it. Who does this control benefit? The early players who made their BTC for next to nothing and sell it at 100x the value. So from a story perspective I ask the question - will BTC be the sacrificial lamb of the model because it was first? Was it conceived and thought out without thinking through the scale of a global currency? Why is it harder and why does the currency get tighter controls and more scarce instead of fewer controls, more transparency, and rapid proliferation and adoption so everyone can access and use it inherent in its design? I love the notion but question the execution of this fiat currency. It appears as if BTC was started by techies to try to fix the problems they saw in financial markets - central control, wealth concentrated in too few hands, little access for the common man. Yet if we play out what is happening, is BTC not headed for the same fate with different beneficiaries of the concentration of wealth (techies vs. wall street/banks)?

  9. Wow, so nice! I am really thankful to you for providing me such a wonderful in formation and i am waiting for that.It has described about how liquid cooling server is used to cool the server?