الفهرس | Only 14 pages are availabe for public view |
Abstract The wide utilization of the Internet and the accessibility of high performance computers and distributed systems as minimal effort item segments deeply affect the manner in which we utilize computers today. These specialized offices have come about to the likelihood of utilizing topographically dispersed and multiproprietor resources to take care of huge scale issues in science, building, and business. Ongoing examination on these themes has prompted the rise of another world view known as grid computing. Performance and utilization of the grid depends on a complex and excessively dynamic procedure of optimally balancing the load among the available nodes. More effective marketing, along with new revenue opportunities, enhanced customer service, improved operational efficiency, competitive advantages over peer organizations and huge business benefits are the outcome of the analytical findings. The organizations performance is raised to the maximum using big data which transforms the tremendous amounts of data into knowledge. In this thesis, a novel two-dimensional figure of merit is presented to describe the network effects on load balance and fault tolerance estimation to improve the performance of the network utilizations. The enhancement of fault tolerance is obtained by adaptively decrease replication time and message cost. On the other hand, load balance is improved by adaptively decrease mean job response time. Finally, design and implement a new paradigm framework for big data clustering which utilizes grid technology and bionic based algorithms. Analysis of genetic algorithm, ant colony optimization, and particle swarm optimization are implemented regarding to their solutions, issues and improvements concerning load balancing in computational grid. Consequently, a significant system utilization improvement was attained. The experimental results confirm the feasibility of the suggested model in real time applications. |