Scheduling of Tasks on Componentizing Kernel During Evolution of Tasks

Abstract

This thesis proposes how to find a feasible schedule with minimum system degradation during the evolution of the software components on Componentizing kernel. On the componentizing kernel, the software component can expand and their system utilization can be increased at run time. This might lead to the system to be un-schedulable. We proposed a method to reschedule the software components at run time by changing their offset (by increasing the release time). But increasing the release time of the software components might cause performance degradation of the system. We maintained the performance of the system by introducing weight for each transaction. To find the best solution in short period of time, we developed a heuristic algorithm.