Adapt the central server algorithm for mutual exclusion to handle the crash failure of any client (in any state), assuming that the server is correct and given a reliable failure detector. Comment on whether the resultant system is fault tolerant. What would happen if a client that possesses the token is wrongly suspected to have failed?

Question

What will be an ideal response?

Answer 1

The server uses the reliable failure detector to determine whether any client has crashed. If the client has been granted the token then the server acts as if the client had returned the token. In case it subsequently receives the token from the client (which may have sent it before crashing), it ignores it.

The resultant system is not fault-tolerant. If a token-holding client crashed then the application-specific data protected by the critical section (whose consistency is at stake) may be in an unknown state at the point when another client starts to access it.
If a client that possesses the token is wrongly suspected to have failed then there is a danger that two processes will be allowed to execute in the critical section concurrently.