MapReduce Performance in Federated Cloud Computing Environments

Large scale scientific and engineering applications, and cloud auditing generate huge amounts of data. MapReduce framework coupled with cloud computing is emerging as the viable solution for distributed big data processing. Specifically, if data is generated from distributed sources and computation is also distributed then multiple clouds need to be set up to minimize data transfer, which introduces us to federated distributed or multi-domain clouds. In addition to security concerns of general clouds, distributed clouds expose new challenges to the performance of cloud based applications including cloud auditing and analysis. This book chapter focuses on a method to deploy distributed clouds and evaluates the performance of various cloud based applications over distributed clouds. It also proposes a method to optimize the performance of cloud based applications over high speed networks.