A Novel Low-Overhead Recovery Approach for Distributed Systems

Algorithm 1

Nonblocking Algorithm.

At each process 𝑃_𝑖(1≤i≤𝑛)

 if 𝑃_𝑖 receives 𝑀_𝑐

      takes checkpoint 𝐶^𝑖𝑥;

   sends its 𝑉^𝑖𝑥(_sent) and 𝑉^𝑖𝑥(_recv) to the initiator process 𝑃^∗;

      // all such vectors from each 𝑃_𝑖 are used by 𝑃^∗ to determine the lost messages

         sent by the processes during (𝐶^𝑖𝑥-𝐶^𝑖𝑥−1) in the event of a failure continues its normal operation;

   else if 𝑃_𝑖 receives a piggybacked application message <m, $> && 𝑃_𝑖 has not yet received 𝑀_𝑐

       for the current execution of the checkpointing algorithm

        takes checkpoint 𝐶^𝑖𝑥 without waiting for 𝑀_𝑐;

        sends its 𝑉^𝑖𝑥(_sent) and 𝑉^𝑖𝑥(_recv) to the initiator process 𝑃^∗;

      // all such vectors from each 𝑃_𝑖 are used by 𝑃^∗ to determine the lost messages

         sent by the processes during (𝐶^𝑖𝑥-𝐶^𝑖𝑥−1) in the event of a failure continues its normal operation;

      // processes the received message 𝑚 and ignores 𝑀_𝑐, when received later