GeNePi: A Multi-Objective Machine Reassignment Algorithm for Data Centres

Abstract

Abstract. Data centres are facilities with large amount of machines (i.e., servers) and hosted processes (e.g., virtual machines). Managers of data centres (e.g., operators, capital allocators, CRM) constantly try to optimise them, reassigning 'better' machines to processes. These managers usually see better/good placements as a combination of distinct objectives, hence why in this paper we define the data centre optimisation problem as a multi-objective machine reassignment problem. While classical solutions to address this either do not find many solutions (e.g., GRASP), do not cover well the search space (e.g., PLS), or even cannot operate properly (e.g., NSGA-II lacks a good initial population), we propose GeNePi, a novel hybrid algorithm. We show that GeNePi outperforms all the other algorithms in terms of quantity of solutions (nearly 6 times more solutions on average than the second best algorithm) and quality (hypervolume of the Pareto frontier is 106% better on average).

Exact and Hybrid Solutions for the Multi-Objective VM Reassignment Problem

“…It is largely inspired by the MRP proposed by Google, but considers objectives that are relevant to data centres' managers in non-aggregated fashion. The first attempt at modelling it and creating an algorithm to tackle it 29 was quite recent. A linear formulation for the same multi-objective problem and a study of the usability of a MILP solver were put forward 30 , but the work was only limited to the smallest instances.…”

“…A comparison between a large number of algorithms has already been performed 29 going from first-fit family techniques, metaheuristics, to hybrid-metaheuristics. A scalable hybrid-metaheuristic (i.e., GeNePi) was then proposed to optimise the multi-objective VM reassignment problem.…”

“…Although this step does not bring a large improvement in terms of hypervolume, it is important as it provides decision makers with more implementation choices. Beside these choices, we use the same parameters as in the GeNePi paper 29 . Figure 7.…”

MILP for the Multi-objective VM Reassignment Problem

“…Although this step does not bring a large improvement in terms of hypervolume, it is important as it provides decision-makers with more implementation choices. Beside these choices, we use the same parameters as in the GeNePi paper [21]. Table III shows the results obtained on the modified ROADEF instances from a 1 1 to b 1, in terms of hypervolume, number of non-dominated solutions and execution time.…”

“…In this section, we want to see whether combining a solver with a metaheuristic allows us to get better results in a reasonable time. A comparison of different metaheuristics has already been performed [21] and a scalable hybrid-metaheuristic (GeNePi) proposed to optimise the Multiobjective VM Reassignment Problem. GeNePi outperforms state-of-the-art algorithms on both quantity and quality of solutions.…”