Publication

Abstract : Nowadays handling of big data transfer among thousands of interconnected servers plays a vital role on cloud computing environment. Big data is nothing but collection of relational data, unstructured data (human readable format), and semi-structured data, streaming data such as machines, sensors, Web applications, and social media. In existing system this concept enhances by fixing some optimal to overcome the bottlenecks occurs while data transfer on scientific cloud application. The parameters are pipelining, parallelism, and concurrency. The major problem is fixing of incorrect parameter combination leads to overloading and under utilization of network which results congestion and packet loss on data transfer. In this study, we proposed a new dynamic work load queuing methods to improvise the network data rate while balancing work load dynamically. We also invoke various scheduling algorithms to predict the unbalanced resource utilization in data center at the initial duration of each time slot which is used to reschedule the unbalanced resource. This rescheduling process recovers overutilization and under utilization on network.