"Virtualization" refers to placing systems, storage, and/or network resources into logical groups or virtualized "pools". These resources can then be exploited by applications and databases that require additional the processing, storage, and/or bandwidth services available in the pool.



When most people think of virtualized environments, they think about distributed server, storage, or network pools that are logically configured to look like one large device (for instance a group of servers can be pooled to appear as one large server; likewise a group of storage arrays can be configured to look like one large array; and a group of network switches can be configured to look like one large switch that can provide variable slices of bandwidth). But virtualization also can refer to single devices (for instance, a PC or server can be configured to run multiple, distinct operating environments as "virtual machines"; or a single storage array can be logically organized to appear as several distinct storage arrays; and so on).



The primary benefits delivered through virtualization are:



- Increased utilization of systems/storage/network resources (sometimes doubling or tripling existing utilization rates); and,

- Simplified management because multiple devices can behave as a single manageable device (resulting in fewer people-related management costs).

But, there are a half-a-dozen secondary benefits that can be delivered through virtualization including better return-on-investment in existing information systems, reduced total-cost-of-ownership, improved business resiliency, better systems/storage availability, increased flexibility and freedom-of-choice (if using a standards-based approach to virtualization), and better preparedness to the application/database service requests of tomorrow's service-oriented architectures.



The state-of-the-market in virtualization is that systems virtualization is very mature (systems have been virtualized for over thirty years now); storage virtualization is less mature; and network virtualization is comparatively in its infancy. In this article, we'll take a closer look at the current state of storage virtualization.

Storage Virtualization: Three Approaches



There are three approaches to storage virtualization:



1. Array-based;

2. System-based; and,

3. Switch-based.

Each approach has very distinct benefits such as in the areas of performance and manageability. Therefore, choosing the right approach has important ramifications on what service levels your IT department can provide, and on systems, storage and network management costs.



To decide which approach to use, IT managers must weigh one critical factor: "where should the intelligence be hosted to handle storage virtualization?" Here's why:



1. Hosting the storage virtualization intelligence at the array level enables the array to rapidly locate storage resources on the array and make those resources rapidly available to applications and databases that are requesting services (this should have some significant performance advantages if - as some IT analysts predict - applications are someday hosted on dedicated storage arrays very close to the storage they need). Additionally, storage vendors that provide array-based virtualization usually have very strong storage management tools that can be used to either manually or automatically manage storage use - simplifying storage management and usually leading to reduced storage management costs.



2. Hosting storage virtualization at the systems level usually does not result in significant performance degradation because applications usually reside on servers - hence

• Email
• Print
• Share
  • Slashdot
  • Digg
  • Stumble
  • Reddit
  • Newsvine