In a shift that has the potential to turn the way IT organizations think about how IT infrastructure is deployed in the data center, DriveScale today unveiled a platform that plugs storage devices directly into an Ethernet switch.
The solution is designed around commodity hardware, and DriveScale CEO Gene Banman says rather than plugging storage into a server, the use of a network switch will make it simpler for IT organizations to process and store data at a much more massive scale.
Fresh off of raising an additional $15 million in funding along with the signing of an alliance with Ingrasys, a unit of Foxconn that will build the DriveScale systems, Drivescale also counts former Sun Microsystems luminaries Scott McNealy and James Gosling among its advisors.
Banman acknowledges that the biggest challenge DriveScale will face will be getting IT organizations to think differently about scale-out architecture. That conversation, however, will be easier to have in Big Data application scenarios where existing compute and storage platforms are already challenging to handle.
In addition, Banman notes that DriveScale provides all the benefits of software-defined infrastructure and doesn’t force IT organizations to scale out compute and storage in tandem. Instead, they can scale out compute and storage independently of one another simply by adding more drives to the switch. The tools for managing that environment are then made available as a software-as-a-service (Saas) application hosted by DriveScale.
DriveScale will be initially available only with magnetic disk drives. But in the near future, the company plans to add support for solid-state disk (SSD) drives. Pricing starts at $6,000 for the adapter needed to plug drives into the switch. There is a $2,000 per node annual subscription fee and a $20 per disk annual fee.
It’s unclear at the moment how willing IT organizations are going to be to adopt what amounts to a completely different data center architecture. But given all the data many of those organizations now routinely wrestle with, many will probably be curious enough to consider new and different options to existing data center architectures that are starting to creak under the strain of Big Data.