Storage Virtualization

By Sanjay Srivastava
Updated: September 06, 2012

Storage Virtualization merges various physical storage devices in a datacenter and creates an artificial view of the storage environment. Typically, clients and servers do not come to know where the files they are using actually reside. They will also not be aware of the type of storage media that holds their data. This could be hard disks or solid state technology such as DRAM.

Storage Virtualization technology was first developed in the mainframe era and has continued in some form on microcomputers and on x86 systems. With the arrival of network based storage and a move away from client / server based storage, interest in the technology has been triggered afresh.

Storage in the storage virtualization pool can take many forms. These could be storage area networks (SAN), direct attached storage (DAS), network attached storage (NAS) and so on.

Advantages of Storage Virtualization

A key advantage comes from the concept of the assignation of a logical unit (LUN) of storage of a given size. If you have created a LUN of, say, 100 GB, and you are presently using only 20 GB of this, then the file associated with your storage will only be 20 GB. As the volume of your data increases, the disk will expand dynamically till it reaches the limit that you have defined. Only then will it request additional allocation. This permits provisioning based on actual use and not on estimates which is a big improvement on how disk storage was traditionally purchased.

An additional advantage is that if you are using 20 GB out of your disk space of 100 GB, then only 20 GB is shown against your account. This pay as you go model can result in substantial savings. However the above two are not the only benefit of Storage Virtualization. Other substantial benefits also exist. These are briefly discussed below –

• Creation of distributed file systems – to the user, remote storage devices look as if they are directly attached to the system. The local system is not aware of the location or even the type of storage device. This makes it simple to manage local users while it allows the organization to take advantage of lower costs of bulk storage and centralized administration.
• Creation of artificial storage volumes – this feature allows you to club several different storage media together to create the image of a single storage device that is much larger. You can continue to add additional storage incrementally.
• Arrays of storage volumes can be created – this has several advantages. First it allows you to implement higher level RAID configurations that are spread both logically and physically over several disks. Naturally this ensures higher storage reliability of storage.
• Permitting incompatible systems to share storage – storage virtualization makes it possible for totally incompatible systems to share the same storage devices. Mainframes, Linus, Unix and Windows all use different file systems to store their files. Storage Virtualization makes it possible to allow each of these systems to connect to the same data center but use the allotted storage independently of each other. Once again this allows for great economies of scale.

Using Storage VirtualizationUsing Storage Virtualization

Storage Virtualization is for you if your organization is looking for any of these benefits –

• High availability and disaster recovery - if your organization needs continuous access to applications and data, you cannot afford single points of failure such as a single storage server or a single access point to the network. A case in point could be implementing a RAID configuration over virtual storage that is itself created out of physical disks in different geographical locations. While there would be some performance impact, you can easily use such a configuration to store your on-line backups. 
• Improved storage performance – if you are using application virtualization or parallel processing techniques, it is near mandatory that you should also implement Storage Virtualization. This allows the read write functions to be spread over different physical storage devices which improves the overall performance of the system.
• Resource sharing – if you are deploying many virtual environments and running different operating systems, you should implement Storage Virtualization. This allows applications using different operating systems to be hosted in a common physical space
• Use with big data – If your organization deals with what is known as ‘big data’, storage virtualization is an essential tool.

The Importance of Back Up

The need for reliable backup can seldom be overstated, but with storage virtualization this becomes even more critical. If you have created logical units (LUNs) of storage for your users but have not built in a proper backup plan then you are introducing a single point of failure. Backup therefore is to be planned for when you begin thinking about your storage virtualization solution and not when you have already rolled it out. Data protection is a critical aspect of storage management in a datacenter. No responsible system administrator will be willing to be faulted on this.

To conclude, storage virtualization is not a new concept, it has been around for a long time. The very fact that it has lasted from the days of the early mainframe indicates that it is a robust concept that has much to offer. With the emergence of large datacenters, we now have the capability to take this concept far beyond what was originally imagined.