ABSTRACT
To avoid making big data equivalent to "big price" it is significant to investigate how to deploy the VMs so as to lower both communication cost and deployment cost for streaming big data processing (SBDP). The authors motivate such a bicriteria problem by considering task scheduling and resource allocation for SBDP in geo-distributed datacenters. They introduce some representative big data processing frameworks in Clouds, with a special emphasize on SBDP. To support efficient data processing and analysis, service providers need to characterize the features of the different big data services and propose new processing models. The geo-distributed datacenters are running rich and complex tasks to support a variety of user requirements. To improve the service performance and guarantee the quality of service at the same time, task scheduling and resource allocation algorithm in a Cloud platform is a fundamental issue in achieving efficient Cloud services.
