ReMatch: An Efficient Virtual Data Center Re-Matching Strategy Based on Matching Theory

Abstract

A virtual data center (VDC) comprises multiple virtual machines (VMs) with communication dependencies represented as virtual links (VLs). These virtual components, i.e., VMs and VLs, often experience fluctuating demands across different resource types. In this article, we focus on addressing the issue of dynamic resource expansion that leads to the relocation of solution components (SCs), where a SC comprises a VM and its attached VLs, with either the VM and/or at least one of the VLs facing resource expansion. This is challenging because of the complexity involved in frequently relocating multiple dependent virtual components across the substrate network. This article presents a model called <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ReMatch</i> that aims at building an efficient remapping plan with reduced remapping cost and improved resource utilization for service providers (SPs) in polynomial time. The overall relocation problem is formulated as a one-to-many matching game with heterogeneous VM demands. Owing to the inapplicability of the classical deferred acceptance algorithm (DAA) and revised DA (RDA), we propose a modified version of the RDA (MRDA) to obtain a weakly stable assignment. Thorough simulation and analysis show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ReMatch</i> outperforms the baseline algorithms considering multiple evaluation metrics.