抽象的な

Self Learning Based Optimal Resource Provisioning For Map Reduce Tasks with the Evaluation of Cost Functions

Nithya.M, Damodharan.P

Due to the massive improvement in the usage of data’s in the real world, it becomes more burdens to handle and process it effectively. The Map reduce is the one of the more developed technology which is used to handle and process the big data/largest tasks. Map reduce is used to partition the task into sub partitions and map those partitions into the machines for processing. This process need to be done by the considering the minimization of cost and meeting deadline to improve the user satisfaction. In the previous work, CRESP approach is used which focus on allocating the map reduces tasks in the machine with the consideration of reduction of cost and deadline. However this method does not concentrate on the skew and stragglers problem which can occur while handling the largest task. In our work, we try improve the performance of resource allocation strategy by considering the skews and stragglers problem in mind. This problem of skews and stragglers are handled by introducing the partitioning mechanism. The partitioning mechanism will improve the failure of task allocation strategy.