One of the fastest-growing segments of the software-everything universe is software-defined storage (SDS). Vendors new and old – including my company – are busy staking out their claims in this fast-developing market. The basic concept is simple: abstract the storage services (e.g. management, data protection, data placement, I/O, etc.) from the hardware that provides these services.
When deployed intelligently, SDS has the potential to be a reliable solution for certain workload and customers large and small. Potential benefits include greater flexibility, scalability, CapEx and OpEx savings, and reduced complexity.
The State of Storage
Enterprise storage today can be fragmented with many moving parts, including different technologies (e.g. SAN, NAS, Flash, Object, etc.), multiple vendors, tools, and management software. Information types are changing too, with structured and unstructured data, rich or complex data, Big Data and the Internet of Things. Unlike the cost of storage hardware, which continues to decline on a per-byte basis, the total IT budget involved in running that hardware – and software – often times cannot feasibly grow as fast as the soaring storage requirements for many organizations. That makes automation even more important and key to easing storage management as organizations struggle to do more with less.
For most organizations, storage is isolated in a silo, separated from other silos like compute and networking. Increasingly, storage isn’t a standalone array, but a combination of arrays, servers and even memory. This can add complexity and make management and change challenging, thus making convergence, integrated storage, compute, and networking solutions more attractive. All of these elements underscore why SDS is gathering momentum. We’re still in the early stages of it, with a variety of solutions being lumped under the SDS banner.
Software Defined-Storage Defined
Much like “cloud computing,” there are many variants to an SDS definition. The most common consensus defines SDS as technology that separates the storage software and services from the underlying hardware for increased flexibility, scalability, automation and cost benefits.
The Storage Networking Industry Association (SNIA) also offers a list of what it believes SDS should include. These essential elements are: automation; standard interfaces; virtualized data path; and scalability.
To propel storage to where it needs to be, we believe SDS must provide the services available on storage hardware, i.e. snapshots, deduplication, replication, and thin provisioning on a software layer that can be deployed on industry standard servers. We believe the best approach is built on three principles: abstract data from the hardware; integrate storage, compute and networking; and orchestrate via software. The intent is to provide flexible solutions that can be easily added to any environment without ripping and replacing existing infrastructure.
By creating a unified pool of hardware resources and adding automation and monitoring tools,
SDS transcends storage virtualization. It moves functions out of the storage appliance and places them close to compute, enabling better load balancing, reducing operational task loads, and improving responsiveness and flexibility.
How is SDS Used Today?
While interest is surely growing, we’re still in the early stages ahead of widespread SDS adoption. For the first time, IDC, in 2014, measured the size of the SDS platform market, specifically platforms that deliver the full suite of storage services via a software stack that uses, but is not dependent on, industry standard hardware built with off-the-shelf components.
According to IDC, SDS platforms will continue to grow faster than any other market segment in the file- and object-based storage market. It will be driven primarily by a rich and diverse set of data-intensive use cases across multiple industries and geographies. (http://bit.ly/1mvErpH).
While SDS can run on any industry-standard hardware, deploying SDS on a tried-and-tested enterprise-class hardware with the appropriate hardware configurations, and working with a vendor with global services and support are critical to an enterprise-quality SDS implementation.
Customers have told us they want flexibility without compromising the quality and reliability of the storage solution, and they are seeking vendors that can deliver pre-tested bundled solutions, from appliances to end-to-end reference architectures. As users opt for newer SDS solutions, they also are looking to evolve their current storage environments. A natural progression likely will lead to traditional storage vendors offering more SDS-like benefits, including increased flexibility, automation and scalability in traditional storage solutions while users seek new solutions that bridge traditional IT with new models.
In addition to SDS, there are plenty of hot storage innovations today, such as flash optimization, storage-as-a-service and cloud storage. However, without separating the software from the hardware and providing the freedom of choice, agility, scalability, manageability, and affordability of the other virtualization and software-defined portfolio, these areas alone will continue to fall short of solving today’s storage opportunities.
New approaches, technologies and solutions are great, but at the end of the day, it’s about addressing customer needs and putting the right data in the right place at the right time and for the right cost. The success of SDS will depend on how well it does this. In the end, it’s about greater flexibility and freedom of choice.