Efficient fault tolerant consensus using preemptive token

Abstract

The problem of agreement has been widely studied by many research groups in crash-stop as well as Byzantine model for distributed transaction systems. In one of our previous works [12], we assumed a trustworthy Transaction Manager, TM amenable to the job of view creation, detection of faulty primary as well as backup replicas, and to evacuate them from the transaction processing system. The TM provides an efficient failure-resiliency in the protocol; however, to eliminate the dependency on a single TM, later, we introduced the concept of Distributed Transaction Manager DTM [13], which distributes the responsibilities of TM among 3f+1(f are faulty) replicas. Both of the above mentioned protocols employ proactive approaches that is, essentially, preemptive in nature. It acts, in advance, in order to mitigate the tentative failures in the transaction processing system. The present article insinuates a system model to optimize the performance of DTM in terms of overall message overhead, latency, and throughput. A deterministic solution has been proposed to synchronize the processes via eventual bisource where the connecting links (incoming and outgoing) of at least one non-faulty process with all other processes is eventually timely.