{"title":"Monitoring Runtime Metrics of Fog Manufacturing via a Qualitative and Quantitative (QQ) Control Chart","authors":"Yifu Li, Lening Wang, Dongyoon Lee, R. Jin","doi":"10.1145/3501262","DOIUrl":null,"url":null,"abstract":"Fog manufacturing combines Fog and Cloud computing in a manufacturing network to provide efficient data analytics and support real-time decision-making. Detecting anomalies, including imbalanced computational workloads and cyber-attacks, is critical to ensure reliable and responsive computation services. However, such anomalies often concur with dynamic offloading events where computation tasks are migrated from well-occupied Fog nodes to less-occupied ones to reduce the overall computation time latency and improve the throughput. Such concurrences jointly affect the system behaviors, which makes anomaly detection inaccurate. We propose a qualitative and quantitative (QQ) control chart to monitor system anomalies through identifying the changes of monitored runtime metric relationship (quantitative variables) under the presence of dynamic offloading (qualitative variable) using a risk-adjusted monitoring framework. Both the simulation and Fog manufacturing case studies show the advantage of the proposed method compared with the existing literature under the dynamic offloading influence.","PeriodicalId":29764,"journal":{"name":"ACM Transactions on Internet of Things","volume":"77 1","pages":"1 - 19"},"PeriodicalIF":3.5000,"publicationDate":"2022-03-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACM Transactions on Internet of Things","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3501262","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"COMPUTER SCIENCE, INFORMATION SYSTEMS","Score":null,"Total":0}
引用次数: 0
Abstract
Fog manufacturing combines Fog and Cloud computing in a manufacturing network to provide efficient data analytics and support real-time decision-making. Detecting anomalies, including imbalanced computational workloads and cyber-attacks, is critical to ensure reliable and responsive computation services. However, such anomalies often concur with dynamic offloading events where computation tasks are migrated from well-occupied Fog nodes to less-occupied ones to reduce the overall computation time latency and improve the throughput. Such concurrences jointly affect the system behaviors, which makes anomaly detection inaccurate. We propose a qualitative and quantitative (QQ) control chart to monitor system anomalies through identifying the changes of monitored runtime metric relationship (quantitative variables) under the presence of dynamic offloading (qualitative variable) using a risk-adjusted monitoring framework. Both the simulation and Fog manufacturing case studies show the advantage of the proposed method compared with the existing literature under the dynamic offloading influence.