Data Storage on Clients for Implementation of Network RAID

Abstract

The client server model has become one of the most central ideas of network computing. Most business applications written today use the client server architecture. Various types of data are stored on these servers which can be accessed by the clients through a front-end application. Hence to concentrate all data on the central server, we require huge storage space on the central server. On the other hand storage space on the clients is wasted. If we have 500GB space on each client, then such a huge amount of space may not be required by each client. Storage space of around 100-200GB may be enough for use on each client. The remaining space may remain unutilized. Hence this storage space can be utilized for all data that needs to be shared among all the clients. By mirroring the data on multiple clients, dependency for data on a particular node is reduced. Also because of mirroring of data, uninterrupted availability of data is ensured which results in improvement in business continuity and disaster recover capabilities from system failure and data loss. Hence task of central server is to consolidate and access data stored in multiple physical locations where data is actually stored. The features of the proposed model include reliability and availability of data, low power consumption, no initial setup for a server and low maintenance. This is more advantageous than the traditional client server model and can replace the enterprise storage solutions.