FAULT AWARE LOAD BALANCING ALGORITHM FOR CONTENT DELIVERY NETWORK

With the increasing use of data sharing, traffic over the internet has increased significantly. There is a need to effectively to manage the load over the servers and maintain the overall system performance with better Quality of Service (QoS). To maintain better QoS, Content Delivery Network (CDN) is used. CDNs offer services that improve network performance in terms of utilizing the bandwidth, improving accessibility, maintaining correctness through content replication and reducing load on servers. The limitation of existing CDN load balancing algorithms is that it considers servers and the systems as non faulty which increases the probability of request been allocated at faulty server. With the increase in number of requests, the failure probability increases due to long waiting queue which increases the network load and processing time. To overcome this problem, a fault aware load balancing algorithm for CDNs is proposed that improves the QoS and reliability of the system. In this paper, the effect of network failure on QoS and reliability of the system is studied in the presence of high request rate and network traffic. Performance of existing load balancing algorithm is investigated and compared in faulty environment. Moreover, the performance of the proposed algorithm is compared with reported techniques. The experimental results demonstrate that proposed algorithm provides better robustness and resilience to fault without affecting the QoS. Further, a dynamic fault model is proposed and implemented which takes care of changing failure probability with load and provided better result as compared to static fault models.