2013/10/29 by Nguyen Quang-Hung, Quang-Hung, Nguyen, Nam Thoai +3
Computer Science · #C.2.4 #C.4 #Cloud Computing and Resource Management #D.4.8 #Distributed #Distributed and Parallel Computing Systems #FOS: Computer and information sciences #IoT and Edge/Fog Computing #Networking and Internet Architecture (cs.NI) #Parallel #Performance (cs.PF) #and Cluster Computing (cs.DC)
paper · pdf · doi:10.48550/arxiv.1310.7801
openalex publication_date 2013/10/29 · openalex created_date 2022/10/07 · openalex updated_date 2026/07/28
Cloud computing has become more popular in provision of computing resources\nunder virtual machine (VM) abstraction for high performance computing (HPC)\nusers to run their applications. A HPC cloud is such cloud computing\nenvironment. One of challenges of energy efficient resource allocation for VMs\nin HPC cloud is tradeoff between minimizing total energy consumption of\nphysical machines (PMs) and satisfying Quality of Service (e.g. performance).\nOn one hand, cloud providers want to maximize their profit by reducing the\npower cost (e.g. using the smallest number of running PMs). On the other hand,\ncloud customers (users) want highest performance for their applications. In\nthis paper, we focus on the scenario that scheduler does not know global\ninformation about user jobs and user applications in the future. Users will\nrequest shortterm resources at fixed start times and non interrupted durations.\nWe then propose a new allocation heuristic (named Energy-aware and Performance\nper watt oriented Bestfit (EPOBF)) that uses metric of performance per watt to\nchoose which most energy-efficient PM for mapping each VM (e.g. maximum of MIPS\nper Watt). Using information from Feitelson's Parallel Workload Archive to\nmodel HPC jobs, we compare the proposed EPOBF to state of the art heuristics on\nheterogeneous PMs (each PM has multicore CPU). Simulations show that the EPOBF\ncan reduce significant total energy consumption in comparison with state of the\nart allocation heuristics.\n