Clustering Storage

Is your data pool growing? Is the management getting difficult? If so, one of the possible answers is to cluster your storage. In a cluster, a group of independent host computers, independent nodes, work together as one system. They may share a common storage array or SAN configuration; they may have a common file system that features simpler name spacing. Whatever the components, clustered storage (modeled on clustered servers) address the problem of file sizes and data sets growing into the terabyte and petabyte ranges. Clustering provides massive throughput because of an increased port count that comes from cobbling many storage servers together into a single pool of disks and processors, all working on a similar task and all able to share the same data.

Historically there has been some confusion about the definition of clustered storage. Vendors of all sizes describe several different technologies as clustered storage, from disk pooling to virtualization. But scalability is at the heart of what clustering is about. A true storage cluster should be able to scale linearly without producing bottlenecks or a significant amount of management complexity.

Clustered storage falls into two categories: systems that combine block-based data on a storage-area network (SAN) and those that create a common file name space across NAS filers. The commonality is that the clustered storage software brings concurrent access to files that are distributed among many servers and storage devices. To date, most of the major storage vendors have released technology that can virtualize NAS systems by pooling disk capacity behind NAS engines. All claim to be developing or evaluating third-party clustering technology as well. Meanwhile, younger companies are already offering clustering software that runs across Windows and Linux. In theory, managing storage clusters should be no more difficult than managing a single array, but some users say their management interfaces could still use some tweaking. Stay tuned.