Mishra-Srimani false alarm and waits again for another time-out

Mishra-Srimani algorithm is a token-based mutual exclusion algorithm on
a distributed system. Here we present Mishra-Srimani algorithm and how to works
this algorithm, its message complexity, synchronization delay, advantage &
disadvantage. This algorithm extended the algorithm of Sujuki and Kasami. When
a process j wants to enter its critical section it sets a time-out after
sending its request to all other nodes. If the token does not arrive within the
time-out, process j checks if another node has started the regeneration of the
token. When j waits for a message indicating it can reset its time-out and wait
for the token again. When no other node has started the regeneration procedure,
j commences it. It sends a PRIVILEGE_CHECK message to all nodes and waits for
their respons. If the token is not lost, then j sends a message to all other
nodes. It was a false alarm and waits again for another time-out for the token
to arrive. If a node k receives a PRIVILEGE_CHECK message, it either sends a
REP_CRITICAL message to j if k has the token, or sends a NO_CRITICAL message to
node j if it has not started a token regeneration procedure. If node k already
started to regenerate the token when a PRIVILEGE_CHECK message arrive. When a
process k that has started the regeneration of the token will know that the
token is lost and no other node is trying to regenerate it. It receives no
REP_CRITICAL messages. Then it sends CREATE_PRIVILEGE messages to all correct
nodes and gets an UPDATE message from each one of them. Node k sends message to
all nodes indicating recovery of the system and starts executing its critical
section25. The synchronization delay of Mishra-Srimani algorithm is 2T and
message complexity is (L*N)+(N-1) where L>21. There are some advantages
and disadvantages in Mishra-Srimani algorithm. This algorithm incorporated a
fault tolerance mechanism to recover the system from a single node failure24.
The algorithm assumes that transfer delays are finite but unpredictable.
Mishra-Srimani algorithm considers node failure, the elimination of duplicated
tokens25. The system state is reconstructed. Here a lost token can be
regenerated and the state of each site can be reconstructed. At any given time,
there exists only one central coordinator in the system.