We all know that bitcoin miners spend their time earning money by computing the huge number of sums that secure the bitcoin network. The problem is that this sizeable resource doesn’t have any other use. It also takes a lot of energy to run.
However, there are alternatives to this way of working, offering more intrinsic purpose to all that computational power. The latest one comes from a research team that wants to use a cryptocurrency network to store archival data. Say hello to permacoin.
What is proof of work?
Bitcoins are 'mined’ using computing power, in a process known as 'proof of work'. The idea is to balance the time it takes to produce new bitcoins, so that each of them is produced in equal time (that's currently set to 25 bitcoins every 10 minutes or so).
To do that, all of the miners on the bitcoin network try to solve a mathematical problem, and the difficulty of that problem varies depending on how many people are trying to crack it.
As more people try to mine bitcoins, the difficulty increases, meaning that it takes more computing power to solve the problem. The proof of work necessary to earn a bitcoin rises, taking more computing cycles for the same outcome. This cycle of increased consumption has been growing exponentially since late last year.
Using storage instead of computing power
The folks behind permacoin come from the University of Maryland, Cornell Tech and Microsoft Research. Led by Andrew Miller, a computer science PhD student at the University of Maryland, the team proposes that we use storage, rather than CPU cycles, to secure a cryptocurrency network – providing a useful way to back up our public data in the process.
Instead of burning through proof of work which has no intrinsic value beyond the network, Miller and his team want miners to store pieces of a large archive of data that society would like to see preserved – the Library of Congress, perhaps.
They suggest this can be achieved by having miners prove that they are storing those pieces of data, in the same way that the bitcoin network makes its miners prove that they have solved mathematical problems.
Miners still have to prove that they have solved a mathematical problem, but the problem needn't get more difficult as more miners joined the network.
Instead, the miners must refer to a piece of code stored locally on their computer to solve the puzzle. If they successfully solve the problem, then the algorithm can deduce that they are storing that data – at least for a short time. Thus, all miners must be storing a piece of the archived data to participate by mining permacoin.
There’s another smart aspect to permacoin known as 'erasure coding', which increases the size of a file slightly, padding it with extra data. Then, even if some sections have been erased, that extra data makes it possible to retrieve a file in its entirety, within limitations.
The case for a distributed archival network
This all sounds plausible, but why bother? Storage is cheap, after all, and in the permacoin academic paper, Miller’s team admits that the amount of storage theoretically available seems relatively low compared to conventional cloud-based storage.
Miller conservatively estimates that bitcoin miners have spent $80m on equipment. If they’d spent that on RAM instead, they could have bought around four petabytes (4 million gigabytes) of RAM storage capacity, he says.
That isn't how much the network would store, however. In the paper, Miller quarters the 4 PB figure, just to be on the safe side, concluding that bitcoin’s users could safely store a 200 TB file if everyone had bought storage instead of computing muscle power.
Two hundred terabytes is pretty small potatoes in the storage world (although estimates say it is sufficient to store the aforementioned Library of Congress print collection), and online archiving is already available, via services such as Amazon’s Glacier.
Miller insists that what permacoin lacks in storage capacity, it makes up for in risk mitigation and diversity.
"The main advantage of permacoin isn't just the amount of data being stored, it's the diverse and decentralized way in which it's stored," he said.
"Even the Amazon Glacier storage service, which offers replicated storage in multiple geographic locations, is confined to Amazon's administrative domain [ie: it's under their control]," Miller continued. "The goal of permacoin is to inherit the same kind decentralized security provided by bitcoin currently, with no central party responsible for the data."
Miller’s approach is slightly different than other decentralized archival projects that have been around for a while, such as Stanford University’s LOCKSS (Lots Of Copies Keep Stuff Safe), for example. That has software which monitors participating computers to check that they’re still holding the files they’re listed as holding, thus ensuring that enough pieces of data are around at any one time to reconstruct a file.
While LOCKSS relies on altruistic participants within a special interest group, permacoin’s approach incentivises miners to store by enabling them to produce more permacoins.
Jeff Garzik, one of bitcoin's core developers, is a big fan of Miller and said that the idea has potential:
"Andrew Miller is a sharp guy. That bootstraps a lot of my personal trust into the system, sight unseen. Proof-of-storage can work."
Who might use it?
Miller prefers to call it 'proof of retrievability', because the permacoin network can prove its ability to get archived files back. Whatever the underlying technology is called, it could be a good way to solve some of the more pressing problems facing organizations with large data stores – especially those who are operating on a non-profit basis.
Take the Internet Archive, for example, which aims to provide "universal access to all knowledge". That now has more than 10 petabytes spread across multiple data centres. Brewster Kahle, who founded the Archive in 1996, said that projects like permacoin have promise:
"The public data in the World Wide Web and elsewhere is growing rapidly, and the cost of hard drives is not falling as fast, so the overall cost of archiving the published works is growing. While it is still manageable, it is something that sustained funding can help insure long lasting and up-to-date collections."
"The Internet Archive and the Wayback Machine could be well served with a distributed system," he added.
There are other alternative systems to CPU- or ASIC-intensive proof of work, such as solarcoin, which only uses miners to process the transaction of pre-mined coins on its network. The coins are exchanged for proven production of solar energy.
Primecoin, produced by 'Sunny King', searches for a specific type of prime number, and has just achieved its fifth world record doing so. However, King admits that these numbers are primarily of theoretical relevance:
"The main practical benefit is that it provides incentives to the development of primality testing related ASIC computing devices, which is a great benefit to cryptography and computing industry in general."
Computational power can be used for other things, such as distributed protein folding and searching for patterns in extraterrestrial radio signals. The problem is that computing tasks like these aren’t the same types of task as solving the cryptography problems intrinsic to cryptocurrency production, and no one seems to have worked out a way to marry the two so far. This means curing cancer or AIDS, while also mining digital coins isn’t a reality.
However, storing data while mining digital coins is.
"We designed our [scheme] to closely match the performance and security of bitcoin," concludes Miller. "If the same amount of money were invested towards equipment and power in permacoin, it would have the same security as bitcoin does now."
We don’t see anyone changing the core bitcoin proof of work in the near future, though, and permacoin hasn’t yet been implemented. Miller and his team have contributed the theory – will anyone take the next step?
Digital archive image via Shutterstock