Multi-Objective Bottleneck-Aware Allocation of Multiple Resources

Abstract

The ever-growing demand for cloud resources places the resource management at the heart of design and decision-making processes in the cloud computing environment. In this paper, we consider multi-objective allocation to optimize the max min( i x ), maximize job numbers, and maximize resources utilization simultaneously. Firstly, a greedy online framework is presented to allow the scheduling decisions to be made based on any well-defined value function. In order to tackle the possibly conflicting objectives, we propose a fuzzy-based priority approach to explore the tradeoffs of two or more objectives at the same time; secondly, a novel algorithm is designed to find the nearest integer solution efficiently while maintaining the constraints and tightly bounding the optimal solution. In addition, this algorithm has very desirable runtime and solution quality properties when the number of tasks and machines become large.