Simulation Of Distributed Fault Containment Algorithm Using Randomized Scheduler


Fault containment is an important feature of stabilizing distributed systems. A distributed system is stabilizing if there exists a finite sequence of moves that leads to a legal configuration. Generally, a distributed system will stay in a stabilizing state once it reaches that configuration unless there is a fault. The probability of a massive failure is very low, and a single fault is more likely to occur compared to multiple failures. However, containing single failures is more important in distributed systems because of the improvement of system reliability. Fault containment algorithms we implemented helped restore the legitimate configurations of the system from any state. This paper presents some experiments we have performed in single fault scenarios using a randomized scheduler. Furthermore, we have conducted simulation experiments. The simulation results give an insight into the efficiency of our algorithms. Variations in the number of nodes and the degree of faulty node are some of the factors that have been considered in the experiments, and they were graphically and numerically analyzed to provide insightful information.


Blazer Summer Research Institute and Honors College at Valdosta State

This document is currently not available here.