Storage is becoming ever more difficult to manage, and storage virtualization offers the key to reducing that complexity. That was the message from storage virtualization expert Marc Staimer, president of Dragon Slayer Consulting, at TechTarget's Storage Decisions seminar in London this week.
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Staimer identified data growth and the growing complexity of storage systems as the key drivers behind user interest in storage virtualization.
He said, "Storage is growing exponentially, not linearly, as a technology, and it is the only thing in the data centre that is constantly being consumed. Bandwidth, for example, gets used and reused; storage does not."
Staimer added: "Storage is consumed because people save their data. You may delete some, overwrite some, but generally you are constantly increasing it. Since that's an exponential growth, the storage goes with it. Therefore, the amount of storage you have to manage is going up, human beings aren't scalable and people aren't hiring lots more administrators, hence management is becoming a greater and greater concern."
Staimer's presentation covered three main areas: virtualization for file (NAS) storage and the problems of NAS sprawl; virtualization of SAN-based storage and switch-based storage virtualization; and virtualization in the infrastructure, which included looking at converged fabric/network standards such as FCoE.
Many attendees at the event were looking for solutions to ease the management of increasingly complex storage environments.
Gareth Byrd is architect/manager EMEA with the Publicis Groupe, based in the London office of the Paris-based international advertising group. Data growth, acquisitions and client demand for more complexity had driven him to look at storage virtualization to ease the management load.
He said, "We need to control storage growth, reduce costs through efficiencies across a disparate range of offices and locations in a fast-changing environment. Storage is a huge management task, and we need to control growth and match the appropriate storage to the correct data. Unstructured data and its growth has a huge management cost associated with it, and there are efficiencies there but they are difficult to find.
Others said they're interested in storage virtualization as a potential means of abstracting multiple storage systems into one manageable entity. For Jim Benson, transportation engineer with the IBI Group, an urban design and architecture firm in London, the challenge is to link two SANs to make availability gains.
He said, "We're carrying out a project on behalf of one of our clients in which they want to implement EMC SANs with virtual machines, and the high-availability side of things is really driving it."
For others, storage virtualization potentially offers ease of management over large storage subsystems.
Mick Hancock, systems administrator with the Royal Borough of Kingston, said, "We recently purchased a new Xiotech SAN, which was quite a big investment at £400,000, so I'm here to get some ideas about how to keep it running sweetly, how to administer the storage and how to integrate it into our virtualization environment, which is growing rapidly."
He added, "We're looking for ease of management, to be able to react to situations easily, to spread the load of management rather than have it all fall on one person."
For an infrastructure manager with an advertising company who wished to remain anonymous, the attraction of storage virtualization is the potential of some products to effect the movement of data between storage tiers as a way of dealing with the growth of unstructured data.
He said, "We've already been down the route of storage virtualization and had experience of one of the products, the F5 ARX, which was a bad experience, but we do recognise that storage virtualization is a really powerful concept. We're really trying to find out what's out there. Our challenge is unstructured data growth and a lack of control over user data, with spiralling storage costs. So, the key interest for us is tiering and moving data onto lower-cost storage."
The Storage Networking Industry Association defines storage virtualization as "… abstracting, hiding, or isolating the internal functions of a storage (sub)system or service from applications, host computers or general network resources, for the purposes of enabling application and network-independent management of storage or data."
SNIA identifies three main infrastructure locations where storage virtualization can take place. They are host-based virtualization, such as the use of logical volume managers on servers; storage-based virtualization, which encompasses the likes of RAID, LUN masking, thin provisioning, etc.; and network-based virtualization, in which devices between servers and storage effect, for example, virtualization of LUNs across heterogeneous storage systems, tiering, archiving and replication.
Despite the plethora of areas in which storage virtualization can be applied, it hasn't yet set the market alight as a distinct product category. When SearchStorage.co.UK questioned readers in its storage Purchasing Intentions survey earlier this year, it found 30% of survey takers had virtualized some or all of their capacity. However, the majority had either not virtualized any storage (43%) or were only likely to evaluate it in 2010 (27%).
Respondents' ideas for the type of virtualization product they intend to purchase were quite fragmented, with 15% looking at host-based virtualization software, 12% evaluating virtualization appliances, 10% weighing virtualization software to run on storage arrays, 9% assessing virtualization software to run on a standalone server and 7% considering virtualization on a fabric switch.