[1]周 航,董宁宁,张 宏.非一致性存储结构架构下服务器物理资源竞争问题[J].南京理工大学学报(自然科学版),2019,43(05):615-621.[doi:10.14177/j.cnki.32-1397n.2019.43.05.011]
 Zhou Hang,Dong Ningning,Zhang Hong.Research on resource contention problem on nonuniform memory access server[J].Journal of Nanjing University of Science and Technology,2019,43(05):615-621.[doi:10.14177/j.cnki.32-1397n.2019.43.05.011]
点击复制

非一致性存储结构架构下服务器物理资源竞争问题()
分享到:

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

卷:
43卷
期数:
2019年05期
页码:
615-621
栏目:
出版日期:
2019-10-31

文章信息/Info

Title:
Research on resource contention problem on nonuniform memory access server
文章编号:
1005-9830(2019)05-0615-07
作者:
周 航1董宁宁2张 宏1
1.周口师范学院 网络工程学院,河南 周口 466001; 2.上海大学 计算机工程与科学学院,上海 200444
Author(s):
Zhou Hang1Dong Ningning2Zhang Hong1
1.School of Network Engineering,Zhoukou Normal University,Zhoukou 466001,China; 2.School of Computer Engineering and Science,Shanghai University,Shanghai 200444,China
关键词:
非一致性存储结构服务器 资源竞争 性能干扰 云计算
Keywords:
non uniform memory access server resource contention performance interference cloud computing
分类号:
TP30
DOI:
10.14177/j.cnki.32-1397n.2019.43.05.011
摘要:
为了提升云服务器的负载应对能力,虚拟化技术被进一步应用于云数据中心的资源调度和资源管理。然而,虚拟化技术在提升服务器资源利用率的同时,也面临着用户类型多样化以及多任务间不同资源类型需求混合而带来的挑战,服务器的物理资源竞争问题日益严峻。该文以非一致性存储结构(NUMA)服务器为对象,分别研究了一般性物理资源竞争及其多Socket之间的资源竞争问题。提出了一般性资源竞争的度量模型,并给出了包括多Socket资源竞争在内的应对策略。实验表明,该模型能够有效表征服务器物理资源的竞争程度,相应的策略算法能够降低物理资源争夺而带来的性能干扰,提升任务运行效率。
Abstract:
In order to improve the ability of handling workload,the virtualization technology has been further applied in resource scheduling and resource management. Although virtualization technology has improved the ability and performance of cloud servers,it also confronts various user types and combination of applications with different resource demands,thus the problem of resource contention has become severer nowadays. In the scope of non uniform memory access(NUMA)server,this paper focuses on the resource contention issue both on general physical resources and the specific resource between multiple sockets. This paper proposes a general model to quantify the degree of resource contention. Then an appropriate strategy is introduced consequently to cope with the resource contention problem including the contention between multiple sockets. Experiments in this paper show that this model can characterize the resource contention effectively. Meanwhile the corresponding strategy can also reduce performance interference significantly thus leads to higher efficiency of running applications.

参考文献/References:

[1] Gantz J F,Minton S,Toncheva A. Cloud computing’s role in job creation[J]. Framingham:IDC,2012,5(3):1-14.
[2]Amazon. com,Amazon elastic compute cloud(Amazon EC2)[EB/OL]. http://aws. amazon. com/ec2/,2017.
[3]idcquan. com,云计算市场潜力巨大 2017年市场规模将超690亿[EB/OL]. http://cloud. idcquan. com/yzx/114338. shtml,2016.
[4]王卅,张文博,吴恒,等. 一种基于硬件计数器的虚拟机性能干扰估算方法[J]. 软件学报,2015,26(8):2074-2090.
Wang Sa,Zhang Wenbo,Wu Heng,et al. Approach of quantifying virtual machine performance interference based on hardware performance counter[J]. Journal of Software,2015,26(8):2074-2090.
[5]Mukherjee J,Krishnamurthy D,Rolia J. Resource contention detection in virtualized environments[J]. IEEE Transactions on Network and Service Management,2015,12(2):217-231.
[6]黄纬,张建德,彭焕峰,等. 数据中心应用感知的动态资源配置研究[J]. 南京理工大学学报,2018,42(3):322-328.
Huang Wei,Zhang Jiande,Peng Huanfeng,et al. Application-aware dynamic resource allocation in data center[J]. Journal of Nanjing University of Science and Technology,2018,42(3):322-328.
[7]申京,吴晨光,郝洋,等. 面向云计算数据中心的弹性资源调整方法[J]. 南京理工大学学报,2015,39(1):89-93.
Shen Jing,Wu Chenguang,Hao Yang,et al. Elastic resource adjustment method for cloud computing data center[J]. Journal of Nanjing University of Science and Technology,2015,39(1):89-93.
[8]Lu K,Yahyapour R,Wieder P,et al. QoS-Aware VM placement in multi-domain service level agreements scenarios[C]//IEEE Sixth International Conference on Cloud Computing. Santa Clara,US:IEEE Computer Society,2013:661-668.
[9]Xu F,Liu F,Jin H,et al. Managing performance overhead of virtual machines in cloud computing:A survey,state of the art,and future directions[J]. Proceedings of the IEEE,2014,102(1):11-31.
[10]Fox A,Griffith R,Joseph A,et al. Above the clouds:A Berkeley view of cloud computing[J]. Dept Electrical Eng and Comput,Sciences,2009,28(13):1-23.
[11]Schad J,Dittrich J. Runtime measurements in the cloud:Observing,analyzing,and reducing variance[J]. Proceedings of the VLDB Endowment,2010,3(1):460-471.
[12]Barker S K,Shenoy P. Empirical evaluation of latency-sensitive application performance in the cloud[C]//Proceedings of the First Annual ACM SIGMM Conference on Multimedia Systems. Auckland,New Zealand:ACM,2010:35-46.
[13]Novakovic D M,Vasic N,Novakovic S,et al. DeepDive:Transparently identifying and managing performance interference in virtualized environments[C]//USENIX Annual Technical Conference. Paris,France:USENIX,2013:219-230.
[14]Iyer R. VM 3:Measuring,modeling and managing VM shared resources[J]. Computer Networks,2009,53(17):2873-2887.
[15]Blagodurov S,Zhuravlev S,Fedorova A. Contention-aware scheduling on multicore systems[J]. ACM Transactions on Computer Systems(TOCS),2010,28(4):82-91.
[16]Tang L,Mars J,Vachharajani N,et al. The impact of memory subsystem resource sharing on datacenter applications[C]//ACM SIGARCH Computer Architecture News. New York,US:ACM,2011:283-294.
[17]Reiss C,Tumanov A,Ganger G R,et al. Towards understanding heterogeneous clouds at scale:Google trace analysis[J]. Intel Science and Technology Center for Cloud Computing,Tech Rep,2012:84.
[18]Zhou H,Li Q,Tong W,et al. P-Aware:A proportional multi-resource scheduling strategy in cloud data center[J]. Cluster Computing,2016,19(3):1089-1103.
[19]SPEC. SPECCPU2006[EB/OL]. http://www. spec. org/cpu/,2006.
[20]Xu F,Liu F,Liu L,et al. Iaware:Making live migration of virtual machines interference-aware in the cloud[J]. IEEE Transactions on Computers,2014,63(12):3012-3025.
[21]Bienia C. Benchmarking modern multiprocessors[M]. Princeton,US:Princeton University Press,2011:235-239.

备注/Memo

备注/Memo:
收稿日期:2018-11-09 修回日期:2019-03-18
基金项目:国家自然科学基金(U1404622); 国家自然科学青年基金(61801527); 上海市科委创新行动计划(16511101200); 河南省重点研发与推广专项(192400410368)
作者简介:周航(1984-),男,博士,副教授,主要研究方向:复杂云场景下的资源调度与能耗优化,E-mail:henry@zknu.edu.cn。
引文格式:周航,董宁宁,张宏. 非一致性存储结构架构下服务器物理资源竞争问题[J]. 南京理工大学学报,2019,43(5):615-621.
投稿网址:http://zrxuebao.njust.edu.cn
更新日期/Last Update: 2019-11-30