In large organizations, where Computers are connected through the network, data is stored on the hard disk of the Central server. This sensitive data is shared between the users on the network. This Data stored on the Central Server gets updated either in real time or in batches. For example, the travel companies or airline companies store ticket reservation information on their central Servers.This data gets updated in real time, as soon as the transaction is completed. This data is accessed by the users to find out the status of their ticket reservation. Further, if this data is made accessible through the internet, users can see updated information on the internet as well.

However, when such large amounts of data are stored on the central computer system or server system, there is always a risk that the data might get damaged or lost. If such important data is damaged or lost it would cause great harm to the users, as well as, the Company. How can this problem be resolved? One solution is to backup the data. The data could be resorted back if there is loss or damage of data. But in simple backup systems it is difficult to restore the system immediately to the normal state. It takes time to restore the system back to the normal state as new hard disk has to be replaced in place of the faulty one, and  the data has to be restored from the backup. However, in environments where data gets updated in real time, such a time lag between system crash and system restoration entails the loss of vital information, such as the latest online transactions. Hence, you can see that merely using the backup strategy is not enough. A better solution available to ensure system reliability, is in the form of raid technology.

Raid

The full form of Raid is 'Redundant array of inexpensive disks'.In case of raid, instead of using one large hard disk to store the sensitive data you can make use of a series of hard disks to store the data.There are two popular techniques of Raid. First one is called as 'Disk mirroring' and the second one is called as 'Disk stripping'. We will shortly see how these techniques help in solving the problems of data loss, in case of large networks and real time computer system.

Disk Mirroring.

In Disk mirroring exact copies of the data are stored simultaneously on different disks. If data is damaged or lost from one disk the system automatically switches to the error free disk immediately, so that the transactions in real time are not affected.This ensures 100% uptime of the system. Disk mirroring is a fault tolerant technique. Since there is instant switch over from faulty hard disks to normal ones the system operation does not get affected.

Disk Stripping.

In Disk stripping different parts of the data file get stored on different hard disks. One main disk is used to store data only for error checking. If one disk fails the data from the error disk and data present on the other disks can be used to reconstruct the lost data. Though stripping offers to save valuable data it is not a fault tolerant technique.

This short discussion was intended to give you only an insight into the importance of raid. If you are in the job of Network administration or data processing perhaps you might now be actively interested in implementing the raid technology. It would be worthwhile reading  other Articles on raid already posted on this blog, as they contain more detail analysis.