Last week VMware released vSphere 4.1, an impressive minor release for its virtual infrastructure which introduced a number of remarkable new features. One of them is called Memory Compression:
Compressed memory is a new level of the memory hierarchy, between RAM and disk. Slower than memory, but much faster than disk, compressed memory improves the performance of virtual machines when memory is under contention, because less virtual memory is swapped to disk.
See Understanding Memory Resource Management in VMware ESX 4.1 for more details.
While virtualization.info can’t say when the IT industry started researching the memory compression technique, we certainly can report about Nitin Gupta, a former member of the VMware’s Technical Staff part of the ESX Resource Management team from India, who mentioned memory compression on his personal blog in March 2009.
Gupta is working on this technology since 2007, turning it into a GPLv2 open source project available for free on Google Code under the name of compcache (aka Compressed Caching for Linux).
More than that, Gupta left VMware in April 2010 and submitted compcache to the Google Summer of Code (SoC) 2010 program.
SoC is a global program that offers student developers stipends to write code for various open source software projects.
Google accepted compcache based on the Gupta’s paper Memory Compression for Virtualized Environments, where he describes how to apply his generic memory compression software to hypervisors.
Gupta already contributed code to the Linux kernel and Xen, so while his paper doesn’t specifically mention any hypervisor, it’s clear where compcache will end up.
Assuming this whole thing is perfectly legal, it will be interesting to see how long will take for Citrix, Oracle or even Red Hat to hire the guy.