An efficient load balancing strategy based on mapreduce for public cloud

Cloud computing has emerged as the most promising technology concept within several structures regardless of their industries. However, the exponential growth experienced by the Cloud has significantly complicated the administration of Cloud platforms. One aspect of the sustainability of Cloud development remains the performance of services supplied by Cloud providers. Through this paper, we propose to tackle the particular feature of performance optimization within the Cloud model by introducing a load balancing architecture based on the MapReduce concept. Indeed, the geographic extent of physical resources which are implemented in different datacenters of the Cloud is at the same time strength in terms of fault tolerance and availability for critical resources, but also a weakness in terms of management of the load balancing of the system. Indeed, the proposed architecture will take advantage from the MapReduce principles to handle the massive number of available resources in order to find out the most appropriate load balancer regarding the requirements of users' requests. The proposed architecture aims to improve both the response time and the fault tolerance within the Cloud.