To date, the implementation of enterprise storage systems has evolved around traditional storage array architectures. Typically, these are single or dual controller systems with cache and disk in RAID configurations. Enterprise IT administrators scale storage capacity upward by simply adding more disks to the existing array controller or controller pair—a technique now referred to as “scale-up.”
Clustering technology applied to storage system architectures now offers the “scale out” alternative: scale a single storage system in both capacity and performance by simultaneously adding or replacing modular controllers and disks. In fact, one can actually combine the ability to scale out—adding nodes to the cluster—with the option to scale up individual nodes, a technique we call “bi-directional” scaling.
There are many situations where have the option to scale the same enterprise storage system out, up or both is a better way forward than continuing to rely on the traditional scale up model. Here we compare the approaches, pointing out the significant operational and economic advantages of the new scaling out paradigms. These include:
Sponsored by NetApp