CSNEO:协同服务器与网络能耗优化的VM映射算法  

CSNEO: VM mapping algorithm for collaborative server andnetwork energy consumption optimization

作　　者：廖彬 张陶[2,3] 于炯 李敏[1,3] 孙瑞娜 Liao Bin;Zhang Tao;Yu Jiong;Li Min;Sun Ruina(College of Statistics&Data Science,Xinjiang University of Finance&Economics,Urumqi 830012,China;School of Information Science&Engineering,Xinjiang University,Urumqi 830046,China;Dept.of Medical Engineering&Technology,Xinjiang Medical University,Urumqi 830011,China;School of Networks Security,University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]新疆财经大学统计与数据科学学院,乌鲁木齐830012 [2]新疆大学信息科学与工程学院,乌鲁木齐830046 [3]新疆医科大学医学工程技术学院,乌鲁木齐830011 [4]中国科学院大学网络空间安全学院,北京100049

出　　处：《计算机应用研究》2021年第7期2135-2140,共6页Application Research of Computers

基　　金：国家自然科学基金资助项目(61562078,61462079);新疆维吾尔自治区“天山青年计划”资助项目(2018Q073)。

摘　　要：已有针对虚拟机映射问题的研究,主要以提高服务器资源及能耗效率为目标。综合考虑虚拟机映射过程中对服务器及网络设备能耗的影响,在对物理服务器、虚拟机资源及状态,虚拟机映射、网络通信矩阵等概念定义的基础上,对协同能耗优化及网络优化的虚拟机映射问题进行了建模。将问题抽象为多资源约束下的装箱问题与二次分配QAP问题,并设计了基于蚁群算法ACO与局部搜索算法2-exchange结合的虚拟机映射算法CSNEO来进行问题的求解。通过与MDBP-ACO、vector-VM等四种算法的对比实验结果表明:CSNEO算法一方面在满足多维资源约束的前提下,实现了更高的虚拟机映射效率;另一方面,相比只考虑网络优化的虚拟机放置算法,CSNEO在实现网络优化的同时具有更好的能耗效率。Existing research work for virtual machine mapping problem mainly aims at improving the server resource and energy efficiency as the goal. This paper considered the impact of server and network devices’ energy consumption in the virtual machine mapping process. Based on the definition of physical server,virtual machine resource and state,virtual machine mapping and network communication matrix,it modeled the virtual machine mapping problem which collaborative energy consumption optimization and network optimization. The problem was abstracted into the packing problem with multi-resource constraints and the QAP problem with quadratic allocation. It designed virtual machine mapping algorithm CSNEO based on ACO and local search algorithm 2-exchange to solve the problem. Compared with 4 algorithms such as MDBP-ACO and vector-VM,the experimental results show that: CSNEO algorithm achieves higher virtual machine mapping efficiency on the premise of satisfying multi-dimensional resource constraints,while comparing with traditional virtual machine placement algorithm which only considers network optimization,CSNEO has better energy efficiency while realizing network optimization.

关 键 词：虚拟机映射 能耗优化 蚁群算法 局部搜索算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]