Scalability of GlusterFS in a small cluster for providing fast access to large amount of data

Abstract

With growing of data volumes, providing of fast access to information becomes a challenge. There are two ways to solve it: scale up (getting more power machines without increasing of its number) and scale out (getting larger number of machines). Scaling up is easier way, but it is often expensive and there is always a physical limit at current point in time when a more powerful hardware cannot be purchased for any price. Scaling out is usually cheaper, however it requires more skills to set up such systems and reach anticipated scale level. This article will be dedicated to one of scale out tools - distributed file system GlusterFS.