Storage

01:20 PM
George Crump
George Crump
Commentary
50%
50%

Software-Defined Storage: What Hardware Players Must Do

If 2014 is the "year of software-defined storage," hardware makers must take these steps to stay relevant.

At some point this year you are bound to hear the claim that 2014 is the "year of software-defined storage" (SDS) and that storage hardware companies are dead. Certainly I expect 2014 to be a significant year for SDS, and I am a fan of the concept, but I don't think we should be ready to pronounce storage hardware dead. Nevertheless, the potential success of SDS creates an interesting question: How should storage system designers respond to the growing threat of SDS?

Build great hardware
Storage hardware's primary response to SDS should be to build great storage hardware. Even if all the services that a vendor typically provides are eventually replaced by SDS, customers will need cost-effective and reliable systems to store their information. There remain certain realities in storage that should provide opportunity for hardware designers.

First, storage system performance is more important than ever, and most storage vendors are using flash-based storage to meet the performance demands of the datacenter. The key is to design a storage system that can tap into the full performance potential of flash. This allows storage system designers to create a custom system designed to match the flash medium's zero-latency nature. That architecture might give them a significant differentiator when compared to other systems that take an off-the-shelf approach. It could also mean designing specific storage software like deduplication and compression not commonly found in today's software-defined offerings.

Build better software
As we point out frequently, all the various data services are not created equal, and there is certainly opportunity to deliver a better software data-services set than what is available from the software-defined vendors. For example, there are storage system vendors that provide thin provisioning technology that has little impact on performance and requires no pre-allocation of space. This advantage could be important enough that you decide to standardize on a platform for that specific capability. The same holds true for other data services like snapshots, replication, compression, and deduplication. If a hardware vendor can provide a better quality of data services it can overcome the "openness advantage" of a software-defined solution.

[Want to learn more about different types of SDS? See Software-Defined Storage: How To Pick The Right Flavor.]

Let software do the grunt work
Another approach would be the opposite: Let software do the grunt work and stop developing the data-services basics. With the abundance of compute power available at the server layer, maybe it's time to let basic storage services like volume management, provisioning, snapshots, and replication become commoditized in software. This would allow storage hardware to do the services that are more compute intensive like deduplication, compression, and automated tiering or caching.

Let users have a choice
Software-defined storage is not a panacea. Embracing SDS means taking more of a "build it yourself" approach to storage management, something that not all datacenter staff will have time for. Many will want or need a turnkey approach. Others may want to dabble in SDS -- but only after they have addressed the core storage needs of the business with a storage system. The logical approach for storage system vendors is to continue to provide integrated storage hardware, but open their platforms up to accept hardware or software from other vendors. In essence, they will become SDS vendors for their customers when the customers want them to.

George Crump is president and founder of Storage Switzerland, an IT analyst firm focused on storage and virtualization systems. He writes InformationWeek's storage blog and is a regular contributor to SearchStorage, eWeek, and other publications.

Developments in object storage technology have made integrating inexpensive, flexible cloud storage easy for organizations of all sizes. As new options have arisen, we've seen improvements in performance, security, availability, and compliance. Yet in this competitive market, only good planning and careful implementation will yield a strong ROI. Join us for the Cloud Shift: Object Storage Services webinar on how to achieve maximum benefit and cost savings from object storage. It's available on demand. (Free with registration.)

Comment  | 
Print  | 
More Insights
Comments
Threaded  |  Newest First  |  Oldest First
Lorna Garey
50%
50%
Lorna Garey,
User Rank: Ninja
1/6/2014 | 6:23:50 PM
Reports of death ...
Or, people could simply realize that when the trade press declares 2014 the year of [inset tech here], that means that the typical enterprise CIO might decide to read a few whitepapers. Heck, we've all been declaring tape dead for decades!
YottabyteDuane
50%
50%
YottabyteDuane,
User Rank: Apprentice
1/8/2014 | 3:14:13 PM
Customer Wins
Whether 2014 is the year of SDS or not, one thing is for sure...the customer wins! SDS has the potential to become the key technology that enables businesses to stop feeding the storage monster. As hardware advances are made, like for example, hashing done at the chip-level, all the features will at least be configurable in software. This year, businesses should begin to invest in SDS for a portion of their data strategy and grow it as their skills and comfort evolve. 
Cartoon
Slideshows
Audio Interviews
Archived Audio Interviews
Jeremy Schulman, founder of Schprockits, a network automation startup operating in stealth mode, joins us to explore whether networking professionals all need to learn programming in order to remain employed.
White Papers
Register for Network Computing Newsletters
Current Issue
Video
Twitter Feed