Data Center Tech Blog

Puzzle pieces

Unless you have been living under a rock for the last few years, you have heard the industry buzzword “software-defined” thrown around for everything from networking, to data centers, to storage. For many, “software-defined” is just that – a buzzword – with the question about whether it will ever actually develop into anything tangible being hotly contested. However, the debate hasn’t stopped with whether it will become tangible or not. In the case of Software-Defined Storage (SDS), the debate has also centered on what it actually is, or means, as exhibited by recent articles like this one in InformationWeek.

Ultimately, the answer to these questions will be answered over the coming years. However, I think SDS this isn’t just another buzzword, but will ultimately re-define how enterprise organizations view and approach their data center infrastructure, as well as how they can benefit from storage moving forward. So, why hasn’t the industry already begun to see SDS grab hold and significantly alter the course of storage? Quite frankly, there have been few production level products ready that deliver on it, particularly from the large established vendors. For example, VMware’s VSAN is just now moving to general availability, EMC’s ViPR started shipping just a few quarters ago, and most of the solutions in this market are available from startups. Furthermore, traditional HDD-based architectures just weren’t sufficient enough, from a performance perspective standpoint, to make software-controlled environments beneficial.

Domino pieces

However, over the last 9 – 12 months, we have seen various dominoes fall, creating an environment where SDS can be successful. One of these influencing elements is how SSD technology has evolved, reaching the proper balance between cost, performance and capacity needed to supplant HDDs in various enterprise environments. Software solutions, like SanDisk®’s FlashSoft, have come to market to help take best advantage of the performance and latency that SSDs provide, enabling the creation of powerful new application throughput. This is all coming together to create new opportunities for software-controlled architectures, showing the promise of SDS.

I think we are about to reach an industry crescendo, which will be realized by SDS moving from industry buzzword, to the new reality. Whether running your storage on commodity hardware or automatically placing your workload on differentiated classes of storage without the application owner having to become a storage expert, I see a tipping point coming that will drive mass-adoption of SDS in the enterprise. SSDs are a facilitator for SDS as they enable quick distribution of workloads across the different classes of storage. Some in the industry seem to agree, as evidenced by John Martin’s recent article in ComputerWorld which states, “in the end, the only way to effectively combine the speed and performance of solid-state storage with the scale and price advantages of capacity-optimized storage is by using a software-defined storage infrastructure.”

I believe that there is no “if” in the regards of the adoption of SDS, and that it is only a matter of time. This does challenge the traditional product model of established storage system vendors which over many years have built up clearly differentiated storage product lines. SDS puts those distinctions into the background as it intelligently selects for the application owner what class of storage to use rather than this being dictated by the storage hardware vendor. Or, as IDC stated in their “2013 Worldwide Software-Based (Software-Defined) Storage Taxonomy” report, “it is in the long-term interest of incumbents to seize this opportunity to shed their hardware-centric mind frame.” Some hardware vendors are actually embracing this change, such as EMC’s purchase of Scale IO, and the introduction of their own offering called ViPR (disclaimer: I know a bit more about the EMC offerings as I actually used to work at EMC prior to joining SanDisk).

Will SDS ever live up to the hype? Where do you see SDS having the most impact in the future? Let me know what you think on Twitter: @FlashStorageMan

– Brian Cox

FlashSoft Applied: Speeding Up The Internet of Things

New CloudSpeed Lineup Eliminates Costly “One-Size Fits All” Approach to Storage Infrastructures

subscribe blog