Joint Computing and Storage Resource Allocation Based on Stable Matching in Data Centers
Qiao Chu, Lin Cui, Yuxiang Zhang · 2017
As the fundamental of cloud computing, efficient scheduling for both computing and storage resource is important for effectiveness of applications in data centers. In this paper, we jointly consider the scheduling for both computing and storage resource in data centers. To solve this coupled placement problem, we apply and extend the three-sided stable matching theory to model the problem to be three-sided matching among computing, storage resources and applications. With carefully defined preference lists for each side and the stability of their matching, we proposed an effective SMB (Stable Marriage Based Algorithm) scheme, which is guaranteed to be able to always output a stable matching for computing and storage resources as well as applications (Virtual Machines). Extensive evaluation results through POX + Mininet show that the proposed SMB scheme can effectively allocate computation and storage parts of different topology with various distance factor. The cost is reduced by at most 29.41% compared to other approaches.