A Tabu Search Algorithm for Optimization of Survivable Overlay Computing Systems

Recently, distributed computing paradigm is gaining much interest since both industry and academia require large computational power to process and analyze huge amount of data. As distributed computing systems - similar to other network systems - are vulnerable to failures, survivability guarantees are indispensable to provide the uninterrupted service. Consequently, in this work we consider a survivable computing systems based on a 1+1 protection mechanism. The goal of the system is to provide scheduling of tasks to computing nodes and dimensioning of network capacity in order to minimize the operational cost of the system and satisfy survivability constraints. Since the problem is NP-complete and computationally demanding, we propose an effective heuristic algorithm based on Tabu Search approach. Extensive numerical experiments are run to verify effectiveness of the heuristic against optimal results provided by CPLEX and other heuristic algorithms.