Authors : N. Senthamarai and M. Vijayalakshmi
Abstract: Cloud computing is a most popular technology because everything like hardware, platform and software are provided as a service. It is an improving area in research which includes load balancing, virtualization and storage etc. Load balancing distributes workloads across various resources such as computers, a computer cluster and network links. The load balancing is to optimize resource usage, maximize throughput, minimize response time and avoid overload. In a cluster system, allocating resources is a critical but challenging issue. A new load balancing policy named AMRA is used which attempts to partition jobs according to the user traffic and system load and improve the performance benefits. The idea of AMRA is to allocate the job to all the servers using the system loads as parameters and dynamically tune the job size boundaries based on the current user traffic loads. AMRA then directs the jobs whose job size lie in the same boundary size to the corresponding servers. AMRA always gives high priority to small jobs and send them to the less loaded servers. The algorithm evaluates the sequential and parallel jobs based on the inputs lang. This study mainly focuses on dynamically balanced load and increases the performance of the system and reduces the overhead.
N. Senthamarai and M. Vijayalakshmi, 2016. Dynamic VM Allocation Using Adaptive Map Reduce Algorithm in Cloud Computing. Asian Journal of Information Technology, 15: 2392-2398.