Mobile Recommendation Engine for Offloading Computations to Cloud Using Hadoop Cluster

The growth of mobile applications due to the increasing popularity of smartphones and ubiquity of wireless access has fueled cloud computing to converge into mobile cloud. The intrinsic storage, processing and battery power constraints of mobile devices can be solved by adopting unlimited storage and computing power offered by cloud servers and turn the issues into a favorable opportunity for mobile cloud computing. We propose to build a mobile recommendation engine that collects context and client sensitive information’s from smart phone users for effectivelymaking a decision to offloadapplications that are compute intensive into cloud through nearby resource rich server. The information so received by the cloudis clustered and partitioned using Hadoop MapReduce framework. The provisioned Hadoop cluster extracts the information’s by mining similarity co-occurrences. The resultant output is used to build collaborative filtering based recommendation model where possible offloading scenarios are listed against each application based on the client awareness. This mobile recommender model provides the advantage of federated and automated movement with very little human interaction for making offloading decision.Earlier models use offloading into cloud only for backing up the data to cloud storages like SkyDrive or Google Drive. Hence computation offloading with recommended service through mining clients information is a novel idea for the near future assuming the growth of smart phones and social networking.