The Beyond Storage Virtualization panel discussion in the SiliconAngle program theCube from VMworld 2011 brought together experts and bloggers from HP, IBM, and Wikibon, to discuss their visions of the future of the storage industry. And for the most part they seemed to agree on a common vision.
Wikibon CTO David Floyer started the session by proclaiming that the virtualization architect will take control of storage as well as servers, although the network organization will try to wrest control of storage from him.
Marc Farley, blogger and former storage consultant, now with HP, agreed and added that as flash storage built into the server becomes increasingly important, it will fuel that trend toward a simplified IT organization.
IBM Global Storage Efficiency Evangelist Steve Kenniston also agreed and added that originally systems admins ran the entire data center. It was only later that operations were broken up into today’s three smokestacks. So virtualization, which will refocus management on the entire system, is the pendulum swinging back.
Wikibon Chief Analyst David Vellante noted that storage-as-a-service, essentially the concept pioneered by Storage Networks, seems to be coming back. The experts agreed that the difference this time is multi-tenancy in storage, which allows economical storage provisioning with adequate security and quality of service among multiple users. The question this raises for users, said Kenniston, is which data is valuable enough to be kept in-house and which can be moved off to the cloud where storage is less expensive.
They each had their own favorite trend from the conference, however. Floyer was excited by VMware’s vision of getting rid of LUNs in favor of storage volumes controlled by the VM, which he predicted would become reality in the next three-to-four years. Farley was interested in storage operations over extended distances and the new standard that VMware is collaborating on. Kenniston was most interested in storage efficiency. “The fact is a disk drive can’t get faster, and we can’t put any more data onto it. So what techniques can we use to make it more efficient, and how do we identify the most important data and what do we do with it?”