With so much emphasis placed on server virtualization, it’s easy to forget the importance of storage in virtual data centers. Virtualization storage requirements can complicate
Consider the varied roles that virtualization storage plays—it holds the applications and application data that help businesses function. It saves backups and disaster recovery content that protects business activities and retains the virtual machine (VM) images that virtual data centers rely on. Virtualization storage requirements have only escalated, driving up costs while raising performance and management concerns.
But relief is out there. Several technologies and practices can help organizations reduce storage utilization and overcome the hassles of virtualization storage requirements that have plagued IT administrators.
Saving space with thin provisioning
Virtualization storage requirements are often reduced by using thin provisioning—sometimes dubbed dynamic provisioning in a virtual environment. The technology assigns less physical storage than the logical unit number (LUN) is logically created for. As an example, a traditional LUN might assign 100 GB to an application, but it might take the application months—even years—to use the entire 100 GB.
Because no other application can use the allocated space, it’s essentially wasted until the application finally uses it. With thin provisioning, the same 100 GB LUN can be created but only assigned perhaps 10 GB of physical storage to start with. As the application fills the space, a company can add more physical storage over time.
The principal advantage here is that an organization can save tremendous capital expense by buying and assigning less physical storage to a far larger LUN. Consider the expense of a 1 TB LUN. Thin provisioning might allocate just 200 GB to start with—an initial space savings of 800 GB.
Although it’s an effective way to forestall large virtualization storage purchases, there are issues to consider. The first challenge with thin provisioning is proper monitoring. An application can fail if the physical storage space fills, so administrators must monitor the LUN carefully and proactively add more physical storage.
“It’s a tradeoff between vigilance and savings,” said Ray Lucchesi, president and founder of Silverton Consulting Inc. in Broomfield, Colo. “You have to be able to keep ahead of the [storage utilization] curve.”
Regular monitoring of virtualization storage may be problematic for smaller shops, often making thin provisioning more difficult to use.
How much storage and when
Just how much more storage to add—and when to add it—requires an understanding of each application and the way it is used. For example, if an application fills only its LUN slowly, an administrator can probably add physical storage in fairly small and regular increments. Applications that tend to fill their allocated space quickly or erratically will probably benefit from a larger initial allocation of storage and larger increments as the LUN fills. The key here is to work with application owners to understand each application’s storage requirements.
Replication of a thin provisioned LUN can also prove problematic. “If the thin LUN is a terabyte in size but in reality is only 200 GB, the replicated LUN will use one terabyte only it won’t replicate only what’s used,” said Pierre Dorion, data center practice director at Long View Systems in Denver. As a consequence, data protection and disaster recovery may wind up wasting virtualization storage space that thin provisioning was designed to save.
Long-term savings with thin provisioning will depend on the applications. In many cases, application owners ask for far more storage than the application actually needs, so it is entirely possible that a 1 TB LUN may fill to just 500 GB to 600 GB over time and require little, if any, additional storage. However, other applications may eventually fill the entire LUN, resulting in little long-term savings. Flexibility with storage requirements is the key benefit to thin provisioning.
Stephen J. Bigelow, a senior technology editor in the Data Center and Virtualization Media Group at TechTarget Inc., has more than 20 years of technical writing experience in the PC/technology industry. He holds a bachelor of science in electrical engineering, along with CompTIA A+, Network+, Security+ and Server+ certifications and has written hundreds of articles and more than 15 feature books on computer troubleshooting, including Bigelow’s PC Hardware Desk Reference and Bigelow’s PC Hardware Annoyances. Contact him at email@example.com.
This was first published in June 2011