Alex Henthorn-Iwane

Subscribe to Alex Henthorn-Iwane: eMailAlertsEmail Alerts
Get Alex Henthorn-Iwane: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


“Software-Defined Storage” is the flavor of the month for vendors hawking storage gear. While it can be a useful way of thinking about storage, it’s not something you can buy, but a set of capabilities. We recommend not getting hung up on formal definitions of “software-defined storage.” Instead, evaluate products on whether they can deliver the key benefits it promises. Those include cost savings, scalability, flexibility, and insuring agility by eliminating single points of failure.

Birth of a Buzzword

VMware popularized “Software-defined Storage” as part of their Software-Defined Data Center vision, which is the next step in their product evolution and still far from reaching the market. Over the past decade VMware has, of course, built a leadership position in server virtualization, providing a software layer to abstract the logical server from the physical hardware it runs on. This makes it much easier to deploy servers when needed, to share hardware with other applications, and to move workloads for load balancing or availability. The economics are compelling and well over half of x86 servers are now virtualized.

However, the expensive storage and networking gear attached to these virtual servers is typically still allocated to individual as if they were physical devices. To make matters worse these storage arrays typically lack flexibility, portability, or scalability, and remain one of the most costly aspects for corporate datacenters.

Customers found that these inflexible resources stood in the way of continued cost savings and the ability to dynamically provision IT resources. Several years ago VMware started publicly musing about extending the abstraction layer from servers to networking and storage, calling it their Software Defined Data Center vision. Through product announcements and acquisition they’ve begun fleshing out this strategy. Many competitors have jumped on the bandwagon, rebranding their point solutions as “Software-Defined Storage.”

The Public Cloud: Already There

Public cloud players like Google and Facebook have already built many of these “Software-Defined Storage” capabilities but only for their own use. The formula is familiar: Proprietary software aggregates commodity hardware into a storage pool and provides some level of data protection and resiliency. Google, for example, developed its Google File System a decade ago. Facebook recently replaced NetApp storage with commodity hardware and its own Haystack file system for its image library. This architecture also efficiently uses flash storage devices to accelerate metadata reads and writes, while using low cost SATA drives for longer-term storage. The bad news is that most customers lack the engineering talent to build and maintain this virtual infrastructure for themselves, and it is limited to very specific workload profiles.

That’s unfortunate because this model can bring similar, or greater, benefits on the storage front than on the server front. They include:

1)     Cost reduction through more efficient use of storage and lower cost hardware

2)     A “fluid” architecture that increases flexibility and reduces deployment time. By abstracting application data from storage hardware, it becomes much easier to scale out workloads, move them among physical devices, and ensure high availability.

Everyday customers need the scalability and performance benefits of these “public-cloud like” storage infrastructures, along with storage services such as data protection and service level management. Whether or not they’re called “Software-Defined Storage,” these solutions will also need to be agnostic across hardware and hypervisor platforms, and be compatible with legacy storage infrastructure.

The Public Cloud for the Rest of Us

Sanbolic has been working on the challenge of distributed computing and storage since 2000. Our Melio clustered file system and volume management technology delivers the benefits of “Software-Defined Storage” to hundreds of enterprise and government customers globally. We are also working with the leading Flash vendors to enable high availability and scale out utilizing flash within in servers and storage arrays, while also enabling the performance of Flash storage across all application workloads.

Sanbolic’s Melio5 is the only software-enabling platform to provide storage and data management capability across legacy SAN as well as DAS and server side flash storage. Hardware and hypervisor agnostic, Melio unifies customers infrastructure and can run on on-premise and in the Amazon and Rackspace cloud infrastructure and provide SAN/NAS scalability for thousands of servers and storage devices and Exabytes of data.

When you hear “Software-Defined Storage” we encourage you to think beyond the jargon and focus on the benefits it claims to provide. We then invite you to consider Melio5 to deliver the public-cloud levels of cost reduction and scalability – not to mention the data protection – you need in an enterprise-proven solution that is shipping today

Follow me on twitter @mvmsan

Related posts:

Getting Pitched on Storage? Ask ‘Em These Questions

SSD and Server-side Flash Scale Out: What to Look For

Solving Capacity Challenges with SSD and Server Side Flash

Web 2.0 infrastructures ready to move to the corporate data center

Why SSD and Flash Need More Brains to Reach the Corporate Data Center

Webinar OnDemand: Change the way you look at server-side SSD and Flash Storage with Sanbolic and Fusion IO

Experience for yourself why OVER 700 customers globally have chosen Sanbolic over more costly and complex alternatives. Melio for FREE

Read the original blog entry...

More Stories By Momchil Michailov

Momchil Michailov is CEO and Co-Founder of Sanbolic. He has served as CEO since 2000. Prior to founding Sanbolic, Momchil was a co founder and CEO of Number One GM, Inc. where he oversaw the company from a SAN hardware distribution start-up to a leader in the broadcast workgroup software space. The company was acquired by Autodesk in 1999. Mr.Michailov brings over 15 years of storage expertise and has a background in technology and media production.