Progressive recovery of interdependent services in enterprise data centers
Ibrahim El-Shekeil, Amitangshu Pal, Krishna Kant · 2016
In this paper, we discuss the progressive restoration planning in an enterprise data center environment after partial or full disruption. Repairing disrupted components takes significant amount of time and human involvement. Thus, after a major disruption, the recovery process will involve multiple stages and during each stage, the partially recovered infrastructures will be able to provide limited services at some degraded service level. However, how fast and effective an enterprise infrastructure can be recovered depends on how the recovery processes will restore the disrupted components considering the inter-dependencies in between the services, along with the limitations of expert human operators. The entire problem turns out to be NP-hard and rather complex, and we devise effective meta-heuristics to solve the problem. By taking some real-world examples, we show that the proposed meta-heuristics gives fairly accurate results compared to the optimal solution, with the ratio between the optimal and heuristic solution never exceeding 1.04.