Enhancing Market-Based Scheduling Algorithm on Globally Distributed Web Servers Using Least Suitable Sealed Bid Technique

Scheduling of a multiple distributed servers is considered as a complex problem.considered as NP-complete problem,where no single efficient algorithm guaranteed to produce optimal results.This thesis investigates on how to find optimal solution for distribute system,by implementing market based scheduling Algorithm(MBSA).On implementing the MBSA, a new auction technique which is the least suitable sealed bid auction will be introduced.it is found that least suitable sealed bid technique will give the close-to-optimum solution.In the implementation, cooperative agents were used as a middleware between web servers and stand-alone schedulers. Each of these agents is connected directly to a different scheduling server. These agents work cooperatively with each other through the Internet using message exchanging.An improved algorithm was implemented to decrease the amount of message exchanging between agents, which increased the performance of the system.

Preview PDF 549082_t_fsktm_2006_2.pdf Download (135kB)

Actions (login required)

View Item