Momchil "Memo" Michailov, CEO of Sanbolic, reached out to me recently to discuss how the combination of flash storage devices and software defined storage (read: a virtualized storage environment) is changing the storage hardware business. I always enjoy conversations with Memo. He likes to challenge current beliefs and point out new ways to look at events and technology.
Memo believes that the storage business, as we currently know it, is going away. He points to nearly $1 billion in acquisitions of suppliers of storage virtualization technology to point out that innovation is trumping the old rotating storage business.
What is software defined storage?
Software defined storage is the use of storage virtualization technology to separate the control of storage from where and how the data is actually stored.
This innovation makes it possible to place objects where it makes sense and use storage devices that match the requirements for storage capacity, performance, latency, reliability and cost that each workload presents.
It is then possible to use low cost storage elements when required; to use high cost, high performance storage elements when required; to move data objects as needed to meet service level requirements without requiring operator intervention; and to hide this movement, sometimes called "auto tiering," from IT staff and workloads alike.
Also, software defined storage enables the pooling of many storage devices to present a single huge storage capability. These elements may come from different manufacturers, and they may be any mix of solid state, rotating media or tape as needed by the workload.
Finally, software defined storage automates the allocation, provisioning, use, de-provisioning and de-allocatation of storage to meet the needs of workloads and also minimize the physical storage used.
What does software defined storage mean to organizations?
When it is possible to build vast storage configurations using any combination of system memory, flash storage, rotating media and tape that address the performance, reliability and archival requirements of workloads, it is likely that the need for expensive and complex storage servers and combined with racks of disks will go away.
Organizations, after all, want an effective storage mechanism not disks. Disks just happen to be the best way to store volumes of data today.
When it is simpler and less costly to purchase a large number of 1U systems and use their memory and any included rotating media as a huge, flexible, easy to set up and manage storage utility, organizations are likely to take that route rather than to continue to purchase today's storage servers, storage area networks and rotating media.
In Memo's view, this means that smaller, innovative suppliers of storage virtualization technology are likely to win out over today's large suppliers of storage equipment.
If we take a look at the history of information technology, it is easy to see that organizations seldom abandon old approaches. They, instead, add new approaches when there is a need for a new workload and keep the old systems as they are. The older systems are updated with new processors, memory, storage, networking and the like on a regular cycle, but as long as what they're doing is needed, they will continue to be used.
An old joke
While listening to Memo, I found myself thinking of an old joke about the CEO of a well-known tool manufacturer speaking to his sales force about the performance of the company at the end of a year. The CEO told the sales folks that he had good news and bad news and asked what they wanted to hear first. The crowd yelled that they wanted to hear the good news first.
The CEO told them that their sales had reached an historic high. He thanked them for their work and promised a nice bonus in their next pay envelope. The crowd then yelled "what could possibly be bad news?"
The CEO told them that no one wanted their drills. The crowed shouted "that's crazy, drills are our number one product, customers are buying them like crazy."
The CEO told them that folks bought drills because they wanted holes. Customers don't really want drills he told them. He went on to say that when someone comes along that offers a better way for customers to get the holes they need, their drills would fall by the wayside.
He promised to invest in research to keep on top of the best ways to create holes. He advised them to always listen to customers about their need for holes and not get too arrogant about the success of the company's drills.
What happens when there is a way to store objects without traditional media
Sanbolic is right to point out that when it is possible to store data objects without having to deploy rotating media, companies are likely to slowly turn away from traditional storage media. The key word here is "slowly." Once a technology makes it into the data center, it stays there for a very, very long time. Companies often operate workloads developed 30 years ago as part of their overall IT infrastructure.
The revolution is coming, but it is likely to move in slow motion.
What do you think?
- Artisan Infrastructure offers wholesale only storage as a service
- Dell shipping new all-flash storage option for data-intensive apps
- How does flash storage fail?
- Tecplot: When storage becomes the bottleneck