There are three steps towards the next-generation data center.
First is consolidation of the storage, computing and network resources into “pools” that can be allocated for any application. The second step is to standardize these resources so that they can be used interchangeably - for example, you can substitute one server for another without changing the application. The last step is to virtualize and “tier” the resources so that you can deploy an application with the correct level of isolation, reliability, performance and security.
Many companies have already made significant progress in realizing this vision of the next-generation data center. Enterprise IT executives have had the most success using storage-area networks (SAN) to implement a strategy of consolidation and virtualization. The success in storage consolidation and virtualization can provide broader insights into the power of the next-generation model.
One characteristic that makes storage easier to virtualize is that one storage block is effectively the same as any other storage block. If you ignore the performance and reliability differences between storage subsystems, the block-level storage they offer is undifferentiated, and can be used by any application for any storage purpose. If you compare storage with computing, you can see that a basic computing resource is dependant on the operating system, the amount of memory, and the patches that have been deployed. It is much harder to create a “computon,” an undifferentiated computing resource that can be used for any application.
As a result of the standardization of the basic storage unit, storage consolidation and virtualization is often the first step taken towards the next-generation data center. Some of the advantages companies reap from storage virtualization are:
* Much higher utilization of resources. By pooling storage across applications it can be more efficiently allocated, without the need to allocate extra storage for future growth for each application.
* Lower complexity and therefore lower operational costs. One large company we spoke with had increased its storage from 1.5 petabytes to almost 4 petabytes without growing the storage team of 14 administrators.
* Much higher availability. Since storage blocks can be moved from one storage subsystem to another without stopping the application, storage virtualization increases the availability of applications.
* Faster application deployment time. Storage can be allocated for a new application in a matter of minutes.
* Isolation. Using virtual SANs, storage administrators can isolate one application from another, while still using a single shared pool of storage.
* Abstraction. Using gateway devices, storage administrators can offer file-based storage (e.g., NFS, CIFS) as an abstraction of block-level storage. That means one storage pool, two storage services (block and file).
All of these benefits that companies gain from storage consolidation and virtualization give us insight into the advantages that are to be gained with other forms of virtualization. Computing virtualization is much more difficult to apply because of the differences between different operating systems and platforms, but the potential benefits are clear.