Software-Defined Storage (SDS) is one of the hottest topics in the IT industry and, as a technology, provides some significant advantages over traditional storage system designs. These advantages translate into substantial benefits that arguably have driven the interest in SDS products to begin with. But not all products labeled as “software-defined” come with all those benefits. For users it’s important to understand what they’re trying to get from SDS and which solutions can deliver it.
What SDS is and How it’s Implemented
SDS is a storage technology that abstracts the storage services from the storage controller and allows them to be run on general-purpose server hardware, like any application. This enables a storage system to be ‘built’ without the need for expensive, purpose-built hardware to run the controller functions or even house the storage media. SDS has been widely promoted as a way to leverage the availability of inexpensive “commodity” server hardware, direct-attached drives and flash.
SDS has enjoyed a remarkable popularity, as evidenced by the number of products now available from storage vendors of all sizes. There are several approaches to implementation that make SDS solutions different, but they tend to be provided in one of two forms: as a pure software product or as software that’s sold with specific hardware.
Software-only SDS
The software-only approach enables the customer to use various server hardware and storage components, either purchased for each implementation, or repurposed from existing equipment. This allows the customer to configure their storage system to their exact specifications, within the parameters of the SDS software. It also can be a very cost effective way to build a storage infrastructure.
The other approach is to sell the SDS software integrated into purpose-built hardware modules, as many existing storage vendors have done. These come in clustered “scale-out” designs with storage embedded into each networked server or traditional, “scale-up” designs comprised of dedicated controller and storage modules. While this implementation method can provide a more ‘turnkey’ solution, it doesn’t offer the same number of options for the users that software-only approach does. At the same time software-only solutions can easily partner with hardware providers to offer a similar end-to-end solution.
Vendor-defined Storage
Traditional storage vendors have historically designed purpose-built storage hardware that’s tightly integrated with their software. Over the years, they’ve evolved these systems to use more standardized hardware to drive down costs (their costs, not necessarily the customers’). Now, these legacy storage vendors, as well as many start-up storage companies, are continuing this evolution using SDS.
This technology makes it easier to incorporate new components or subsystems into their storage products, leveraging the component upgrade cycle. But these systems take much of the design options out of the users’ hands, prompting the term “vendor-defined storage”.
Why SDS is so popular
SDS has become a compelling topic for IT because of what it can do – the benefits it can deliver in the data center – not from the technology, per se. So while the particular implementation of the technology is interesting, it’s not the whole story.
[to continue, click HERE]
