A Novel Approach to balance the load between nodes of distributed computing
AbstractThe cost/performance magnitude relation of networks of workstations has been perpetually rising. This trend is anticipated to continue within the close to future. the mixture peak rate of such systems typically matches or exceeds the height rate offered by the quickest parallel computers. Thus, distributed computing systems are a viable and fewer pricey various to parallel computers. However, a significant problem in coincidental programming of a distributed system is a way to alter programming and cargo equalization of such a system which can include heterogeneous computers. A distributed system may be viewed as a group of computing and communication resources shared by active users. Once the demand for computing owner will increase the load equalization drawback becomes necessary.We will state the load equalization drawback as follows. Given the initial job arrival rates at every pc within the system notice Associate in nursing allocation of jobs among the computers in order that the time interval of the complete system over all jobs is reduced. Often, jobs in an exceedingly distributed system may be divided into totally different categories supported their resource usage characteristics and possession. As an example the roles that belong to one user will type a category. Or else, we will distinguishdifferent categories of jobs by their execution times. Reckoning on what percentage job categories square measure thought-about we will have single category or multi-class job distributed systems. During this paper we have a tendency to take into account the load equalization drawback in single category job distributed systems.
 Bridgewater S.A. Jesse, Boykin Oscar P., and Roychowdhury P. Vwani, “Balanced OverlayNetworks (BON): An Overlay Technology forDecentralized Load Balancing”, IEEETransactions on Parallel and DistributedSystems, Vol. 18, pp. 1122-1133, No. 8, August 2007.
 Bahi J.M., Vivier C., and Couturier R.,“DynamicLoadBalancing and Efficient LoadEstimators for Asynchronous IterativeAlgorithms,” IEEE Trans. Parallel andDistributed Systems, vol. 16, no. 4, Apr. 2005.
 Cortes A., Ripoll A., Senar M., and LuqueE.,“Performance Comparison of Dynamic Load-Balancing Strategies for DistributedComputing,” Proc. 32nd Hawaii Conf. SystemSciences, vol. 8, p. 8041, 1999.
 Cybenko G., Dynamic load balancing fordistributed memory multiprocessors, Journal ofParallel Distributed Computing, Vol 7, pp. 279-301, 2001.
 Dhakal S. , “On the optimization of loadbalancing in distributed networks in the presenceof delay, Advances in Communication ControlNetworks,” LNCSE vol. 308, pp. 223–244,Springer-Verlag, 2004.
 Dhakal S., Hayat M. M.,Elyas M., GhanemJ.and Abdallah C. T., “Load Balancing inDistributed Computing Over Wireless LAN:Effects of Network Delay” , IEEECommunications Society / WCNC 2005, Vol. 2, pp.
 Dhakal Sagar , Hayat Majeed M., PezoaJorgeE., Yang Cundong, and Bader DavidA.,”Dynamic Load Balancing in DistributedSystems in the Presence of Delays: ARegeneration-Theory Approach” IEEETransactions on Parallel and Distributed Sytems,vol. 18,no. 4, pp. 485-497, April 2007.
 Gelenbe E. and Kushwaha R., “IncrementalDynamic Load Balancing in DistributedSystems”, submitted for publication, 1993.
 Ghanem J. et al., “Load balancing in distributedsystems with large time delays: Theory andexperiment,”Proceedings of the IEEE/CSS 12thMediterranean Conference on Control andAutomation (MED ’04), Aydin, Turkey, June2004.
 Hayat M.M., Dhakal S., Abdallah C.T., Birdwell J. D., and Chiasson J, “Dynamic Time Delay
Models for Load Balancing. Part II: StochasticAnalysis of the Effect of Delay Uncertainty,”
Advances in Time Delay Systems, vol. 38, pp.355-368, Springer-Verlag, 2004.
This work is licensed under a Creative Commons Attribution 4.0 International License.