Abstract: Demand on big data is being rising day by day and also growing heavy burden on computation, storage and communication in data centers, which cause significant expenses to data center providers. So, cost reduction became an issue for the upcoming big data. One of the primary feature of big data is coupling of data and computation as computation assignment. Three obligations like data placement, task assignment and data movement impact the rate of facts centers. In this paper we study how to reduce the cost using joint optimization of these above three factors for big data service in geographically spreaded data centers. Right here we recommend 2-d markov chain to describe time to finish a particular undertaking with consideration of data transmission and computation to derive average challenge finishing touch time in closed time. The problem with the mixed integer nonlinear programming solved by linearizing it.
Keywords: big data, markov chains, data centers, nonlinear programming, geo-distributed.