COST MINIMIZATION TECHNIQUES FOR BIG DATA PROCESSING IN DATA CENTERS

Abstract

In commonly we are using data as single cluster to store in central approaching and its moving with inefficient or infeasible due to lack of different imitations as lack of wide area transmission capacity and the low latency essential of data processing. In big data processing across all the information processing to specific geo-distributes centers working and computing resources. Whatever we are managing distributed data using Map Reduce functions or computations are data will distributed on datacenters not really to solve different types of technical issues. In data centers how to appropriate data surrounded by selection of geo-distributes to minimize the transmission cost, how to resolve the VM (Virtual Machine) managing approach that offers very high performance and low cost this working criteria will select datacenters as outcome of the reduce for big data analytics jobs. In this paper, these difficulties is tended to by adjusting bandwidth cost, stockpiling cost, processing cost, migration cost, and latency cost, between the two Map Reduce stages across datacenters. We figure this intricate cost streamlining issue for information development, asset provisioning and reducer choice into a joint stochastic number nonlinear improvement issue by limiting the five cost factors all the while.