Window based security-sensitive scheduling for real-time application on cluster

Real-time tasks scheduling on cluster become more challenging when security requirement of application is considered. Running security algorithm on cluster to combat the security threat increases overall execution overhead. Increasing security strength causes decreasing task's acceptance ratio. In this study, we proposed a scheduling algorithm which balances this two conflicting requirements of real-time task to achieve maximal acceptance ratio with maximal security strength while scheduling on homogeneous cluster. Further, extensive simulation experiments which compared with SAEDF and TPSS, show that our algorithm has higher success ratio, more fairly in term of lower standard deviation of security level gained by tasks and significant improvement in overall system performance.