A novel scheduling strategy for an efficient deadlock detection

Deadlock is one of the most important problems in distributed and centralized systems. Many algorithms are introduced for the detection of deadlock; however some of them suffer from increasing the number of processes, while the others take long processing time. This paper studies the impact of deadlock detection on the overall performance of deadlock handling. Network performance largely relies on the accuracy of deadlock detection. This paper proposes a deadlock detection protocol based on threads, processes which are acquired, released or stopped and which thread wait for the other and causes the deadlock. Our algorithm makes the detection goes faster, reliable and schedules the threads in order to detect which thread would initiate the deadlock. Our algorithm simulation results forecasted the deadlock initiator.