Dynamic Cloud Deployment of a MapReduce Architecture

Cloud-based MapReduce services process large datasets in the cloud, significantly reducing users´ infrastructure requirements. Almost all of these services are cloud-vendor-specific and thus internally designed within their own cloud infrastructures, resulting in two important limitations. First, cloud vendors don´t let developers see and evaluate how the MapReduce architecture is managed internally. Second, users can´t build their own private cloud-infrastructure-based offerings or use different public cloud infrastructures for deploying MapReduce services. The authors´ proposed framework enables the dynamic deployment of a MapReduce service in virtual infrastructures from either public or private cloud providers.