Management of missed transactions in a distributed system through simulation

A real time distributed computing system has heterogeneously connected computers to solve a single problem. If the transactions run across different sites, it may commit at one site and may get fail at other site, leading to an inconsistent transaction. The transaction in a real time database system has deadlines to process the workloads and it need to process transactions before these deadlines expired. Series of simulation study have been performed to manage the missed transactions under different transaction management conditions such as different workloads, distribution methods, execution mode-Distribution and Parallel, impact of dynamic slack factors to throughput. The scheduling of data access are done in order to meet their deadlines and to minimize the number of transactions that missed deadlines. A new approach called as “intelligent agent” method is introduced. This agent will keep track and does recording of the status of failing transactions so that we can provide an extra slack time to improve the performance of the system.