The most popular and comprehensive Open Source ECM platform
First there was server virtualization technology like VMware and Citrix Xen. But there hasn’t been much equivalent activity around storage virtualization. It’s really been the software to support storage virtualization that’s been lacking. Storage hardware costs have dropped significantly. Commodity storage now feature low-cost high-capacity drives. For example, SCSI disk prices have dropped and SATA disks have grown in capacity. Cloud-based and Web 2.0 applications plus the dramatic increases in the amount of unstructured data which is being stored has been incentive for vendors to look more closely at the software that drives storage virtualization.
Gluster is an Open Source software that was designed to do exactly that. It is taking on the problem of storage virtualization. Gluster takes commodity storage hardware components, just about anything, and is able stitch those pieces together to build multi-petabyte clusters. The hardware components can range anywhere from JBOD, SAN, Fibre Channel, DAS, 10GbE, to Infiniband. The software creates a File System called GlusterFS, and the capacity and performance of the software can be scaled by adding new hardware.
Gluster was a company started five years ago from technologists at Lawrence Livermore Laboratory. At Livermore the group worked to build a super-computer out of commodity components and open source software. Gluster provides an open source alternative to scale-out architectures and file systems like ZFS, Lustre and GFS.
The secret of the success of the Gluster technology is the fact that they bypass something called the metadata server (MDS). They have written their algorithms to embed in the Gluster Storage Platform the intelligence to know how to automatically find the location for where a piece of data is stored.
Jack O’Brien, senior director of marketing for Gluster, said that the advantage of Gluster is that “it’s modular and easy to configure. When you need new features, just add a new module. You get great performance with small or large files, or even with random files.”