Thin provisioning is a method of allocating a central storage reserve on an “on demand” basis. This provides computers and servers on the network with additional storage as required, providing flexibility to the network. Thin provisioning avoids wasting network resources by constantly monitoring and reallocating which computer receives available storage space, providing a more flexible approach to local and wide area network storage solutions. The downside to thin provisioning is that it requires a constant network connection to operate successfully; when a computer is disconnected from the network, it can no longer receive access or adjustments to its allocated storage amount.
When thin provisioning is in place, each computer receives the amount of storage that it needs. Considering this from a perspective of a home computer user can help to appreciate the benefits of thin provisioning. Each computer on a basic home network is autonomous from the others; the storage space available to one computer is not available to the others, and vice versa. This can result in one computer having a significant amount of free space on a drive while another network computer’s hard drive might be close to full.
In a thin provisioning environment, the “extra” storage of the less-encumbered computer can be shifted to the computer that requires a bit more room on the drive. By constantly reevaluating which computer on the network requires additional storage, a thin provisioning system can maximize the efficiency of its available hard drive space. This improves both the performance and the economic efficacy of the network, as additional funds will not have to be spent on drive space until the entire network is actually full.
Thin provisioning has some downsides, however. First of all, understand that the storage space being swapped around is not physically located within the network computers; instead, it exists in a central location on the network within a completely separate computer. This proximity issue results in problems if the network happens to crash. During a crash, the network computers will no longer have access to any additional storage space, with the files saved in those drive remaining off-limits until the network comes online again.
Second, a network using provisioning techniques will have increased computational overhead compared to a network with fixed storage techniques. One or more computers must take on the responsibility of monitoring storage and deciding where to allocate available blocks of storage. This results in additional stress on the CPUs of the system, possibly resulting in system lag during periods of exceptionally high network activity.