Storage Virtualization (part 2)
Already, we can see a few of the elements that have been mentioned earlier. The haphazard dispersion of files has created an especially complex storage environment, all the while allowing the company to make the most out of their available disc space. The mapping table connects the seemingly haphazard physical locations with their more intuitive, easier to understand logical locations, and the user is able to more fully grasp what files they have stored.
The user makes a request for a file that is saved on LUN-1 (Logical Unit 1) and the mapping table connects that file to a logical block on disc array 2 (LBA-2). The file is then retrieved from LBA-2 and delivered to the user as though it came from LUN-1 4. This, in a nutshell, is how storage virtualization is supposed to work. A complex, yet efficient storage environment is simplified by a virtualization client so that the user doesn’t have to mess with piecing all of the pertinent files back together.
So where can you get a storage virtualization software client? A few companies that vend such agents are Altiris, IBM, and the EMC Corporation. EMC, a highly regarded name in the digital storage business since 1979, has a product called Invista, which is a virtualization agent that promises to “reduce planned downtime and other disruptions” and “increase the flexibility of your storage infrastructure” 5. As a company that delivers $11.6 billion worth of revenue in a year, the EMC Corporation has clearly got something on making a profit in the digital storage business.
IBM, another heavy hitter in the world of digital storage, has a virtualization agent called POWER™ Virtualization that promises to allow organizations a way to take full advantage of space on their IBM System p servers. Some of other advantages that POWER provides are reducing network complexity (of course), faster implementation timeline with a reduced risk of business disruption (sound familiar?) and the ability to redirect IT resources toward business-critical applications 6.
Both IBM and EMC’s product descriptions highlight the positive aspects of storage virtualization, and describe only what advantages it will provide for your company. So what are some of the disadvantages of storage virtualization? No man-made technology is ever perfect, so what is it about storage virtualization that is less than ideal?
Because the implementation of data virtualization is fairly new (hence the expression “emerging technology”), no standard way of doing things has yet arisen. This lack of standards complicated the possibility of interpretability between two organizations that may be employing different flavors of virtualization 7 For instance, IBM’s POWER™ is designed to work with IBM System p servers. If one company wanted to merge records with another organization that already had an infrastructure set up that is composed of different devices, the virtualization agent would prove to be more of a hindrance than a help.
Another instance in which virtualization would prove to be a bit of a nuisance is the potential complication of data replication. When storage virtualization is implemented, only the virtualization agent knows where the data is physically stored. To effect a large-scale back up, the pertinent files have historically needed to be on physically contiguous discs for a traditional backup 8. To implement a backup with a virtualized system, while possible, requires extra work that involves some predetermined procedure of regathering all of the necessary files and copying them. With storage virtualization, large-scale backup is more time consuming, but with a well-designed procedure, it can prove to be of minimal consequence.
Even though the whole point of storage virtualization is to prevent managers from having to mess with the physical storage environment, it does occasionally need to be dealt with. As you might expect, with the limited attention that goes into the physical environment with virtualization in place, it can get pretty confusing. Like a cardboard box that people have been throwing a bunch of stuff into, the physical environment can take quite a bit of time to organize. The only redeeming factor where physical management is concerned is that it doesn’t have to be done often. It mostly just comes up when a company is looking to switch storage solutions or giving other organizations or entities access to their records.
Lastly, when a company decides to implement storage virtualization, they run the risk of adding another layer of data that they need to keep safe and backed up. Metadata is the proverbial “glue” that connects the logical data location with the physical data location, thereby make the virtualized system work 9. In the event of some disaster wiping out your mapping tables, it is vitally important to a company too have backup tables that they can use to get their system back up and running again. If that metadata gets lost, the company then faces the harrowing task of trying to reconstruct it by sorting through the complicated physical environment. This risk is an easy one to avoid, but it is important enough to consider before deciding to switch to a storage virtualization network.
Running at approximately $900 per server virtualized, storage virtualization is a relatively inexpensive way to simplify the way your company retrieves data. With new companies rapidly coming forward and providing their own virtualization clients to the market, prices for virtualized solutions are expected to continue to fall 10. With emphatic potential for system improvement, and manageable and avoidable risks, the ledger leans more heavily towards storage virtualization being an advantageous solution. As the technology matures, and standards are agreed upon, one should not be surprised to see storage virtualization become more and more commonly implemented among businesses who routinely store and access large amounts of data.