An efficient voting and priority based mechanism for deadlock prevention in distributed systems

Abstract

In voting based approach a process getting the majority of votes will only be allowed to enter into the critical section (CS). But, this method has a disadvantage as if no process remains in the situation of achieving majority of votes then the system will remain in idle state although a considerable amount of minority processes will remain in waiting queue. In order to solve the concurrency and starvation control problems of distributed systems, the voting and priority based optimal & feasible solution is presented in this research work. In the prioritized distributed mutual exclusion based algorithm, the process having highest priority amongst all the processes of distributed system is allowed to enter into the critical section (CS). Hence, leaving the other lower priority processes into the waiting queue which results an increase in the length of waiting queue. By keeping these mentioned reasons into considerations I have proposed a new algorithm which can provide a solution for voting and priority based algorithms. Here, I have presented a combined solution for voting and priority based approaches by including some other constraints like treatment for `non-maskable interrupts' and `Shortest Job Scheduling First (SJSF)' techniques. The proposed technique allows the creation of multiple critical sections in a distributed system and prevents the processors from entering into idle state which leads towards increase in throughput.