Research Article
EBA: An Adaptive Large Neighborhood Search-Based Approach for Edge Bandwidth Allocation
@INPROCEEDINGS{10.1007/978-3-031-24383-7_14, author={Qinghong Hu and Qinglan Peng and Jiaxing Shang and Yong Li and Junjie He}, title={EBA: An Adaptive Large Neighborhood Search-Based Approach for Edge Bandwidth Allocation}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 18th EAI International Conference, CollaborateCom 2022, Hangzhou, China, October 15-16, 2022, Proceedings, Part I}, proceedings_a={COLLABORATECOM}, year={2023}, month={1}, keywords={Edge computing Edge bandwidth allocation 95th-percentile billing}, doi={10.1007/978-3-031-24383-7_14} }- Qinghong Hu
Qinglan Peng
Jiaxing Shang
Yong Li
Junjie He
Year: 2023
EBA: An Adaptive Large Neighborhood Search-Based Approach for Edge Bandwidth Allocation
COLLABORATECOM
Springer
DOI: 10.1007/978-3-031-24383-7_14
Abstract
As a promising computing paradigm, edge computing aims at delivering high-response and low-latency computing, storage, and bandwidth resources to end-users at the edge of the network. However, those edge-based novel applications (e.g., live broadcast, edge cloud game, real-time AR/VR rendering, etc.) are usually bandwidth-consuming, which has made a considerable contribution to the operating costs of edge application providers. Meanwhile, the bandwidth pricing modes of edge infrastructure providers are also complicated. Therefore, how to allocate the bandwidth demands of edge users to suitable edge servers to minimize the monetary cost of edge application providers becomes an important issue. In this paper, we consider the widely adopted 95th-percentile bandwidth billing mode and Quality-of-Service constrained edge bandwidth allocation problem, and propose a neighborhood search-based approach, shorts for EBA. It firstly employs a network flow-based heuristic to find a feasible initial solution quickly. Then, an adaptive large neighborhood search-based method is utilized to perform iterative optimization, which contains hill climbing and simulated annealing mechanisms. Therefore, the proposed EBA approach can expand the searching space, accelerate the optimization convergence speed, and avoid falling into local optimal. Experiments based on a real-world edge bandwidth consumption dataset illustrate the effectiveness of the proposed approach.
