AN EFFICIENT LOAD BALANCER FOR CONTENT DELIVERY NETWORKS

A content delivery network or content distribution network (CDN) is a large distributed system of servers deployed in multiple data centers across the Internet. The CDN’S purpose is to serve content to end-users with high availability and high performance. Content delivery networks have become the order of the day with servers specializing in specific contents to be delivered. The issues are that the distributed nature of architecture a content delivery network poses causes control issues which should be monitored and mainly due to its complexity of geographical distance factor. The problem dealing with different servers receiving requests and passing on received requests to the correct content server is tricky as the client or server may not be closest to it and the routing may fail if the content server is busy with intermediate server taking more bandwidth. So an effective routing protocol may reduce the delay, ensure the content is served and all this is based on a set of parameters. So handling flash crowds with sudden surge in requests also takes up bandwidth and causes failure in both passing on the request or delivering the content requested. This leads to QOS (quality of service) issues. A better response time server is based on geographical distance from the client is chosen on the basis of network proximity. Whereas the overall system throughput is optimized through load balancing across a set of content file servers. There are no clear identification of factors used by commercial systems in the literature, possible concurrence is for reducing response time. The current study proposes an algorithm that dynamically estimates server load and balances the work load in real time for Content Delivery Networks by integrating all aspects of the protocol.