Rampant data growth combined with business demands for real-time analysis have created the perfect storm for enterprise storage environments. As more data floods in from mobile, social and Internet of Things (IoT) applications, IT teams are grappling with the problem of where to put it all. But simply warehousing data is no longer enough — organizations want to leverage it to improve decision-making and gain competitive advantages. That requires a storage infrastructure that can adapt to changing workloads and deliver near-instantaneous access to data.
Continuously adding storage devices and appliances won’t cut it. Organizations need an alternative to the traditional hardware-centric approach if they are going to meet capacity and application demands and keep a lid on costs.
Software-defined storage (SDS) is emerging as the answer to today’s storage conundrum. The definition of SDS remains somewhat fluid as vendors take varying approaches to the technology. But in essence, SDS separates storage management from hardware so that capacity can be pooled and automatically provisioned and allocated via an independent software stack. Rather than relying on a storage vendor for a top-to-bottom solution, organizations can deploy SDS on commodity server hardware and integrate various storage technologies based upon business requirements.
Although SDS incorporates the concepts underlying storage virtualization, there are significant differences between the two technologies. Storage virtualization consolidates multiple storage resources into a virtual pool of capacity, with policy-based management at the virtual machine level. SDS moves provisioning, monitoring, de-duplication, replication and other storage services into the software layer, where they can be applied to older devices that lack advanced functionality. This also enables more granular, policy-based control, and provides the agility to respond to changing workload demands.
As a result, SDS delivers a number of compelling benefits, including:
That said, SDS also poses some significant challenges. While SDS software will theoretically run on any general-purpose server, most solutions have hardware compatibility requirements that can create integration headaches. Also, not all commodity hardware is the same, and it isn’t always cheaper than proprietary platforms. Because there’s no set approach to SDS, you should carefully evaluate various solutions to find the right match for a particular application or workload.
Generally, you should avoid appliance-based solutions as they perpetuate the vendor lock-in that SDS seeks to avoid. However, a multivendor approach increases complexity, complicates issue resolution and could result in finger-pointing if a problem arises.
In order to select the right SDS solution, you should start by defining your operational and financial objectives, and how SDS will help you achieve them. Look for a solution that will leverage your existing compute, storage, networking and management resources as much as possible. That also applies to your team’s skill sets and expertise. Ideally, you should try the solution before you buy to determine if it’s truly compatible with your environment and delivers on the vendor’s promises.
SDS can’t solve all your storage problems, but it is well-positioned to address the twin challenges of growing data volumes and shifting application demands. By separating storage software and hardware, SDS helps create an environment that’s more flexible, scalable and responsive.