Given the fact that no one managing storage these days seems to know what to expect next, there’s a critical need for maximum flexibility when it comes to scaling storage.
This requirement has given rise to a scale-out storage movement where the goal is not only how quickly a particular storage platform can add capacity, but also how to do it in a way that allows performance to linearly scale as well.
The latest storage vendor to jump into the scale-out storage fray is Nimble Storage, which today announced a 2.0 upgrade to the core operating system that manages its storage systems. According to Radhika Krishnan, head of product marketing and alliances, the 2.0 upgrade makes it possible to cluster Nimble Storage CS400 Series arrays together in a way that linearly scales both capacity and performance.
That “scale-to-fit” storage architecture is unique in that at no time does it require an IT organization to over provision storage, says Krishnan. Each array can be dynamically added to a cluster as required. Much of this is accomplished using a Cache Accelerated Sequential Layout Architecture (CASL) that combines flash and high-capacity disks to make sure performance and capacity scale in a linear fashion. In effect, the Nimble Storage approach creates storage pools in a cluster that can be used to segment applications and workloads on different arrays, says Krishnan.
The days when IT organizations could carefully plan their storage needs are pretty much over thanks to the rise of virtualization and the advent of Big Data. But it’s not like most IT organizations can afford to throw storage hardware at the problem. Storage systems need to be intelligent enough to allow IT organizations to bring extra capacity on line in way that doesn’t wind up slowing application performance. That’s obviously a lot easier said than done when you consider all the controllers and arrays that data might be distributed across.
Nimble Storage is hardly the only storage vendor trying to address this problem. But whatever the approach ultimately taken, it’s pretty clear that given the increasingly dynamic nature of enterprise application requirements, most of the storage systems in place today are based on architectures that are rapidly becoming obsolete.