In-memory computing was invented a decade ago to process real time data ranging typically to hundreds of “gigabytes” of data per server. Today, the size of real time data has been growing exponentially. Real time applications now need to provide instantaneous response after working on many “terabytes” of data. That’s where traditional DRAM infrastructure is breaking down. Real time applications of today demand memory infrastructure of many terabytes, at lower cost without changing the applications. Furthermore, as more and more data are placed in memory, higher availability is expected from the memory infrastructure.
Big Memory Computing is the next generation of memory infrastructure designed to meet this challenge. IDC defines Big Memory as a market segment consisting of memory hardware and Big Memory software that creates a new persistent memory tier. The market is explosive with revenue for persistent memory projected to grow at a 248% CAGR to reach a revenue of $2.6B in 2023. MemVerge delivered Memory Machine, world's first Big Memory software, virtualizing DRAM and PMEM and creating a pool of software defined memory. Memory Machine delivers bigger memory at lower cost and DRAM-speed, without changing the applications. Memory Machine also performs persistence-on-demand with its patent-pending ZeroIO™ Snapshot technology, transforming the volatile memory into a highly available memory tier.