The Hypernet Foundation is launching the Hypernet protocol in order to make high- performance parallel computing a consumer good. Currently, only large corporations can truly benefit from massively parallel computing by exploiting their oligopoly on consol- idated computing power. However, abundant processing power from the industrial In- ternet and data from ubiquitous, inexpensive sensing technologies will soon make decen- tralized computing more powerful and less expensive than centralized models. Hypernet takes advantage of this trend through a new parallel programming model that unites the power of blockchain technology with the simplicity and robustness of distributed average consensus to enable high-performance computing on the decentralized, heterogeneous, dynamic networks of the future. This technology will finally empower consumers to take on ambitious, novel, massively parallel projects in big data, AI, and modeling with the tools to compete against the likes of Google, Amazon, and Facebook – and win.
In order to bring big data to the edge and leverage the massive latent computer power already found in the digital wild, Hypernet is developing a block-chain based resource scheduling protocol, a containerized execution environment, and an API based on the principle of distributed average consensus (DAC) to provide the capability for true par- allel public resource computing, not just grid computing. Hypernet’s API naturally fa- cilitates efficient collective operations on distributed data even under dynamic network scenarios. This API suite will allow developers to incorporate data-center-scale computa- tional capability into their distributed applications.