Efficient Inline Deduplication on VM Images in Desktop Virtualization Environment

Abstract

Enterprise service is transforming from traditional physical computing nodes to virtual machines which provides better isolation and more effective use of computing ability of the hardware. However, the widely deployment of virtual machines also increase the pressure of storage significantly with the fact that each must has at list one multi-gigabytes image file to store. To address the high pressure of storage from virtual machine images, we developed a user level inline deduplication file system with Content Addressable Storage (CAS). We use the open-source framework FUSE to encapsulate the deduplication process so as to achieve portability and flexibility. Compared to an ordinary file system without deduplication, we show that our file system can save at least 30% of space of single VM image and even more of multiple VM images while achieving considerable run time performance.