Debunking Myths About Software-Defined Storage

Enterprises are beginning to realize SDS’ benefits, such as easier movement of workloads from node to node, improved security and smoother data movement.

Before we get into dispelling any myths, we’ll define what we’re talking about first. Software-defined storage (SDS) is an evolving concept for data storage software to manage policy-based provisioning and management of data storage independent of the underlying hardware and its firmware. Thus, SDS is designed to work immediately with any storage hardware and file system.

SDS has been described as being too complicated for the average IT shop to run. Why? Because it doesn’t come in a box accompanied by an expert to install it. This myth is ironic. The fact is that SDS has reference architecture and close integration with familiar and easy-to-use management plug-ins, such as vCenter.

Related to the “too complicated” issue is the argument that SDS is only for the more sophisticated enterprise IT teams with specialized resources, such as classically trained Unix or CLI (command-line interface) admins. The fact is, new SDS solutions are built for every possible management style, from DevOps to CLI to simple-to-use Web consoles and APIs, to integrate with enterprise solutions. They come in many flavors, including pre-architected versions offered and supported 24/7 by major hardware OEMs that include Dell and SuperMicro. Now sufficiently mature, SDS is enterprise-ready for deployment in larger environments that require ease-of-use and ease-of-management.

SDS performance is perceived to underperform from an IOPS and latency standpoint when compared to all-flash options. However, is speed the answer to every storage question? Enterprises are looking for the flexibility to go from archival (slow) to file services and messaging applications (medium) to big data analytics and high-performance database operations (turbo). SDS can be tuned to suit the speed necessary for the task.

This relatively recent misperception, suggested by industry analysts, points out that software vendors who decry hardware vendor lock-in aren’t fully disclosing their own designs of lock-in to their specific software. In fact, SDS is based on open-source foundations, and standard tools make transferring to new hardware and/or software easier as requirements change and as technology advances.

Like an old wives’ tale that sounds truer the more it is told, this myth may owe its genesis to the size of early adopters when compared to large corporations that wait to adopt technology at later versions. But this idea is now outdated. New-gen SDS offers both scale out and scale up, delivering storage options up to multiple petabytes for organizations of all sizes and across a range of data-intensive industries, including education, financial services, manufacturing, telco and government.

At a high level, SDS is the abstraction of data services. With this abstraction comes the ability to transform not-specifically-for-storage hardware into storage systems. SDS is changing the legacy storage paradigm by fundamentally changing storage market economics. It’s time to replace the myths with the facts.