Parents have told groups of siblings for generations that ‘many hands make light work’. That may not sound good for brothers and sisters who don’t want to clean the garage, but in the technology world, it may prove to be true. The latest application of blockchain and the decentralization technology model may be the creation of supercomputers that are able to perform at or above the highest level of supercomputing currently available but without the use of additional hardware.
A supercomputer is simply that – a computer that is functioning at or above the very highest level of computing currently available. The fastest computers available these days are in China and the US. The Department of Energy’s Oak Ridge Leadership Computing Facility computer, nicknamed Jaguar, and also the NICS (National Institute for Computer Sciences) shared machine, nicknamed Kraken. There’s also the Nebulae, a machine located in China’s National Supercomputing Center. All of these are extremely high cost hardware based computers that function at the highest possible levels. But with big data playing a major role for companies in the future, some are looking to invest in such solutions.
However, decentralized technology may have begun to offer solutions to the need for large-scale computing needs at budget costs. The concept is that each computer in the world is using less than its optimum computational power. By combining these computational ‘leftovers’ across the ‘fog’, as theorists call it, supercomputing could be viable with the existing hardware available in the world.
The advantages would be immediate and obvious. For starters, there would be huge amounts of processing power immediately available. Second, the up-time of the ‘computer’ would be permanent. There would be zero downtime, since the computer is really the amassing of all available and running machines in the world at any given time. Third, the system would be economically vastly superior to any available supercomputer currently available, and would be accessible by anyone willing to tap into it.
The key to implementation is likely incentivizing users. People with idle computer power on their computers could simply program them to provide computational power to the system, and earn money for it. One example of an already functioning decentralized computer is Bitcoin, but a decentralized computer that only functions in a single computation – value transfer. Other currencies like Ethereum are seeking to increase the amount of complexity in the decentralized computing by offering users the ability to do more complex transaction. However, these systems are still a work in progress.
While this system may present some really cool options for the future of supercomputing, right now there isn’t a viable algorithm available to begin making it functional. The hurdles are there, but the substructure is available to begin to think through the pieces needed for implementation. In this case, it may be possible that the many hands of the countless PCs in the world could make ‘light work’ for those seeking supercomputing abilities. It may be that simply waiting a few years will be the best play in the supercomputing race.