A Scalable Algorithm for Readers-Writers Problem

The readers-writers problem is quite popular in the field of parallel and distributed systems; there are numerous ways in which the algorithm has been developed and implemented. The thesis aims to propose and implement an improved solution for the fair scalable reader writer problem. The solution is implemented by an API provided by Java which is the RMI (Remote Method Invocation). The end user who can be both reader and writer has complete control of the respective processes even in the critical section which won't be possible in the other case. The main problem solved is bringing scalability with the help of hash map, the existing doubly linked list approach will be costly, since each time the process has to signal that it whether enters or exits the critical section, whereas here, since the processes are independent, signaling is done only by the writer. Adding to it, each process has a hold and stop control that can be executed in the critical section.