云虚拟机低利用率筛查的摘要敏感性与名单可识别性

张谦; 盛圣宇; 刘童娜

doi:10.11959/j.issn.1000-0801.DXKX260579

Chinese

您当前的位置：

首页 >

文章列表页 >

云虚拟机低利用率筛查的摘要敏感性与名单可识别性

研究与开发 | 更新时间：2026-09-14

- 云虚拟机低利用率筛查的摘要敏感性与名单可识别性
- Telemetry-summary sensitivity and list identifiability in low-utilization screening for cloud virtual machines
- 电信科学 2026年
- 作者机构：
  
  1.华北电力大学经济管理系，河北保定 071003
  2.华北电力大学电子与通信工程系，河北保定 071003
- 作者简介：
  
  张谦（1977—），男，博士，副教授，主要研究方向：能源经济与电力系统管理。
- 基金信息：
  
  河北省自然科学基金资助项目(G2024502001)
- DOI：10.11959/j.issn.1000-0801.DXKX260579
  中图分类号： TP393
- 收稿：2026-08-23，
  
  修回：2026-09-02，
  
  录用：2026-09-09，
- 稿件说明：
移动端阅览
张谦, 盛圣宇, 刘童娜. 云虚拟机低利用率筛查的摘要敏感性与名单可识别性[J/OL]. 电信科学, 2026.

ZHANG Qian, SHENG Shengyu, LIU Tongna. Telemetry-summary sensitivity and list identifiability in low-utilization screening for cloud virtual machines[J/OL]. Telecommunications Science, 2026.
张谦, 盛圣宇, 刘童娜. 云虚拟机低利用率筛查的摘要敏感性与名单可识别性[J/OL]. 电信科学, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260579.

ZHANG Qian, SHENG Shengyu, LIU Tongna. Telemetry-summary sensitivity and list identifiability in low-utilization screening for cloud virtual machines[J/OL]. Telecommunications Science, 2026. DOI： 10.11959/j.issn.1000-0801.DXKX260579.

摘要

为评估遥测摘要口径对云虚拟机低利用率筛查及复核名单传播的影响，比较实例内平均CPU与5 min窗口最大CPU的95分位摘要向订阅、部署层级的传播差异。以Azure V2（2019）为主样本、V1（2017）为复核，计算生命周期加权积分歧义指数（IAI）、Top-10%稳定性证书和并列识别区间。2019年和2017年IAI分别为0.4227和0.3778，订阅层级46个阈值均未满足稳定条件；全支持域对照下，名单替换率分别为38.14%和34.56%。研究表明，摘要口径与阈值域饱和共同改变复核对象，固定破并列规则不能替代集合值识别；筛查结果仅供人工复核，不应直接触发关停、缩容或迁移。

Abstract

Low-utilization screening can select different cloud virtual machines when the telemetry summary changes

yet fixed tie-breaking rules may conceal that instability. The propagation of within-instance summary disagreement to subscription- and deployment-level review lists was examined. A 30-day Microsoft Azure trace from 2019 served as the primary sample

with the 2017 trace used for replication. For each instance

average central processing unit (CPU) utilization was paired with the 95th percentile of 5-min-window maximum CPU utilization. The analysis estimated a lifecycle-weighted integrated ambiguity index (IAI)

ranking-stability certificates

and exact tie-aware identification intervals for top-10% lists

with adjacent-threshold and full-support comparisons. The lifecycle-weighted IAI was 0.4227 in the 2019 trace and 0.3778 in the 2017 trace. At the subscription level

none of the 46 thresholds satisfied the sufficient stability condition. After full-support integration removed cutoff ties

the point-identified replacement rates remained 38.14% and 34.56%

respectively. Telemetry-summary choice and restricted-domain score saturation therefore jointly alter the units selected for review. Deterministic tie breaking cannot replace set-valued bounds

and screening outputs should not directly trigger shutdown

downsizing

or migration.

关键词

Keywords

references

CORTEZ E , BONDE A , MUZIO A , et al . Resource Central: understanding and predicting workloads for improved resource management in large cloud platforms [C ] // Proceedings of the 26th Symposium on Operating Systems Principles . New York : ACM , 2017 : 153 - 167 . DOI: 10.1145/3132747.3132772 http://dx.doi.org/10.1145/3132747.3132772 .

EVERMAN B , GAO M , ZONG Z . Evaluating and reducing cloud waste and cost: a data-driven case study from Azure workloads [J ] . Sustainable Computing: Informatics and Systems , 2022 , 35 : 100708 . DOI: 10.1016/j.suscom.2022.100708 http://dx.doi.org/10.1016/j.suscom.2022.100708 .

KUMBHARE A G , AZIMI R , MANOUSAKIS I , et al . Prediction-based power oversubscription in cloud platforms [C ] // 2021 USENIX Annual Technical Conference . Berkeley : USENIX Association , 2021 : 473 - 487 .

BARBALHO H , KOVALESKI P , LI B , et al . Virtual machine allocation with lifetime predictions [J ] . Proceedings of Machine Learning and Systems , 2023 , 5 : 232 - 253 .

毛东峰 , 贾曼 , 何晓明 , 等 . 网络遥测技术及其在网络自动化运维中的应用 [J ] . 电信科学 , 2021 , 37 ( 2 ): 154 - 163 .

MAO D F , JIA M , HE X M , et al . Network telemetry technology and its application in automatic network operation and maintenance [J ] . Telecommunications Science , 2021 , 37 ( 2 ): 154 - 163 . DOI: 10.11959/j.issn.1000-0801.2021012 http://dx.doi.org/10.11959/j.issn.1000-0801.2021012 .

余竞航 , 赵一辰 , 王凌 , 等 . 基于虚拟网络嵌入的异构算力网络资源管理 [J ] . 电信科学 , 2024 , 40 ( 10 ): 86 - 99 .

YU J H , ZHAO Y C , WANG L , et al . Heterogeneous computing network resource management based on virtual network embedding [J ] . Telecommunications Science , 2024 , 40 ( 10 ): 86 - 99 . DOI: 10.11959/j.issn.1000-0801.2024230 http://dx.doi.org/10.11959/j.issn.1000-0801.2024230 .

祝淑琼 , 徐青青 , 李小涛 , 等 . 算力度量与任务调度：物联网端侧设备策略研究 [J ] . 电信科学 , 2024 , 40 ( 4 ): 122 - 138 .

ZHU S Q , XU Q Q , LI X T , et al . Computational measurement and task scheduling: a study on IoT edge device strategies [J ] . Telecommunications Science , 2024 , 40 ( 4 ): 122 - 138 . DOI: 10.11959/j.issn.1000-0801.2024084 http://dx.doi.org/10.11959/j.issn.1000-0801.2024084 .

武振宇 , 赵占军 , 卜忠贵 , 等 . 智算中心推理业务能力评估方法研究 [J ] . 电信科学 , 2025 , 41 ( 7 ): 200 - 206 .

WU Z Y , ZHAO Z J , BU Z G , et al . Research on evaluation methods for inference business capability of intelligent computing center [J ] . Telecommunications Science , 2025 , 41 ( 7 ): 200 - 206 . DOI: 10.11959/j.issn.1000-0801.2025167 http://dx.doi.org/10.11959/j.issn.1000-0801.2025167 .

REISS C , TUMANOV A , GANGER G R , et al . Heterogeneity and dynamicity of clouds at scale: Google trace analysis [C ] // Proceedings of the 3rd ACM Symposium on Cloud Computing . New York : ACM , 2012 : 1 - 13 . DOI: 10.1145/2391229.2391236 http://dx.doi.org/10.1145/2391229.2391236 .

VERMA A , PEDROSA L , KORUPOLU M R , et al . Large-scale cluster management at Google with Borg [C ] // Proceedings of the 10th European Conference on Computer Systems . New York : ACM , 2015 : 1 - 17 . DOI: 10.1145/2741948.2741964 http://dx.doi.org/10.1145/2741948.2741964 .

BRUGNACH M , BOLTE J , BRADSHAW G . Determining the significance of threshold values uncertainty in rule-based classification models [J ] . Ecological Modelling , 2003 , 160 ( 1/2 ): 63 - 76 . DOI: 10.1016/S0304-3800(02)00309-5 http://dx.doi.org/10.1016/S0304-3800(02)00309-5 .

XIE M , SINGH K , ZHANG C H . Confidence intervals for population ranks in the presence of ties and near ties [J ] . Journal of the American Statistical Association , 2009 , 104 ( 486 ): 775 - 788 . DOI: 10.1198/jasa.2009.0142 http://dx.doi.org/10.1198/jasa.2009.0142 .

ADOMAVICIUS G , ZHANG J . Classification, ranking, and Top-K stability of recommendation algorithms [J ] . INFORMS Journal on Computing , 2016 , 28 ( 1 ): 129 - 147 . DOI: 10.1287/ijoc.2015.0662 http://dx.doi.org/10.1287/ijoc.2015.0662 .

SOLIMAN M A , ILYAS I F , BEN-DAVID S . Supporting ranking queries on uncertain and incomplete data [J ] . The VLDB Journal , 2010 , 19 ( 4 ): 477 - 501 . DOI: 10.1007/s00778-009-0176-8 http://dx.doi.org/10.1007/s00778-009-0176-8 .

WANG C , YUAN L Y , YOU J H . On the semantics of top-k ranking for objects with uncertain data [J ] . Computers & Mathematics with Applications , 2011 , 62 ( 7 ): 2812 - 2823 . DOI: 10.1016/j.camwa.2011.07.047 http://dx.doi.org/10.1016/j.camwa.2011.07.047 .

ASUDEH A , JAGADISH H V , MIKLAU G , et al . On obtaining stable rankings [J ] . Proceedings of the VLDB Endowment , 2018 , 12 ( 3 ): 237 - 250 . DOI: 10.14778/3291264.3291269 http://dx.doi.org/10.14778/3291264.3291269 .

MANSKI C F . Partial Identification of Probability Distributions [M ] . New York : Springer , 2003 . DOI: 10.1007/b97478 http://dx.doi.org/10.1007/b97478 .

BONTEMPS C , MAGNAC T . Set identification, moment restrictions, and inference [J ] . Annual Review of Economics , 2017 , 9 : 103 - 129 . DOI: 10.1146/annurev-economics-063016-103658 http://dx.doi.org/10.1146/annurev-economics-063016-103658 .

SANKARAN A , KARLSSON L , BIENTINESI P . Ranking with ties based on noisy performance data [J ] . International Journal of Data Science and Analytics , 2025 , 20 ( 5 ): 4363 - 4384 . DOI: 10.1007/s41060-025-00722-1 http://dx.doi.org/10.1007/s41060-025-00722-1 .

MICROSOFT AZURE . Microsoft Azure traces: Azure Public Dataset V2 [EB/OL ] . ( 2019 )[ 2026-08-20 ] . https://github.com/Azure/AzurePublicDataset/blob/master/AzurePublicDatasetV2.md https://github.com/Azure/AzurePublicDataset/blob/master/AzurePublicDatasetV2.md .

MICROSOFT AZURE . Microsoft Azure traces: Azure Public Dataset V1 [EB/OL ] . ( 2017 )[ 2026-08-20 ] . https://github.com/Azure/AzurePublicDataset/blob/master/AzurePublicDatasetV1.md https://github.com/Azure/AzurePublicDataset/blob/master/AzurePublicDatasetV1.md .

DAVISON A C , HINKLEY D V . Bootstrap methods and their application [M ] . Cambridge : Cambridge University Press , 1997 . DOI: 10.1017/CBO9780511802843 http://dx.doi.org/10.1017/CBO9780511802843 .

CAMERON A C , MILLER D L . A practitioner's guide to cluster-robust inference [J ] . Journal of Human Resources , 2015 , 50 ( 2 ): 317 - 372 . DOI: 10.3368/jhr.50.2.317 http://dx.doi.org/10.3368/jhr.50.2.317 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于计算机视觉的机柜设备和U位识别方法

基于“智云上海”云宽带网络速联功能的技术实现

跨域通算智算资源效能采集与提升研究

云宽带环境下基于浅压缩的云交互质量优化方案研究

基于M-LAG技术的天翼云专线改造方案