{"title":"意图驱动的路径状态监控,以实现集中的状态感知流转向","authors":"Christoph Hardegen","doi":"10.23919/CNSM55787.2022.9965141","DOIUrl":null,"url":null,"abstract":"Running state monitoring for network switches and links enables the derivation of a path-based data view. Thereby, various monitoring intents like utilization and latency awareness can be targeted during data collection. Whereas each of these objectives relies on a particular set of state metrics, different monitoring methods may be run to gather the data basis serving as decision input for subsequent analysis purposes. In addition, path conditions have an impact on the state observed for individual packet streams being forwarded along a specific path. While path level state data is of relevance, e.g., to evaluate past load ratios in order to run state-aware and efficient path determination, flow level state helps to monitor flow experience conditions like achieved throughput or perceived latency, e.g., to track the compliance with flow-based requirements. This paper presents a modular architecture for path state monitoring that considers port counter query, network probing and in-band network telemetry as methods for demand-driven data collection and focuses on utilization and latency awareness as monitoring intents. State data is collected by a centralized controller in collaboration with distributed modules deployed in a switch’s data plane to run data tracking, wherefore programmable switches are used as operational basis to ensure a flexible monitoring protocol. Evaluations show that continuously collected data snapshots allow to track accurate path state trends that – w.r.t. path state-aware traffic steering – can be leveraged to improve flow-based load distribution across available path capacities and to resolve inefficiencies like imbalanced path load or congestion.","PeriodicalId":232521,"journal":{"name":"2022 18th International Conference on Network and Service Management (CNSM)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-10-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"Intent-Driven Path State Monitoring to Enable Centralized State-Aware Flow Steering\",\"authors\":\"Christoph Hardegen\",\"doi\":\"10.23919/CNSM55787.2022.9965141\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Running state monitoring for network switches and links enables the derivation of a path-based data view. Thereby, various monitoring intents like utilization and latency awareness can be targeted during data collection. Whereas each of these objectives relies on a particular set of state metrics, different monitoring methods may be run to gather the data basis serving as decision input for subsequent analysis purposes. In addition, path conditions have an impact on the state observed for individual packet streams being forwarded along a specific path. While path level state data is of relevance, e.g., to evaluate past load ratios in order to run state-aware and efficient path determination, flow level state helps to monitor flow experience conditions like achieved throughput or perceived latency, e.g., to track the compliance with flow-based requirements. This paper presents a modular architecture for path state monitoring that considers port counter query, network probing and in-band network telemetry as methods for demand-driven data collection and focuses on utilization and latency awareness as monitoring intents. State data is collected by a centralized controller in collaboration with distributed modules deployed in a switch’s data plane to run data tracking, wherefore programmable switches are used as operational basis to ensure a flexible monitoring protocol. Evaluations show that continuously collected data snapshots allow to track accurate path state trends that – w.r.t. path state-aware traffic steering – can be leveraged to improve flow-based load distribution across available path capacities and to resolve inefficiencies like imbalanced path load or congestion.\",\"PeriodicalId\":232521,\"journal\":{\"name\":\"2022 18th International Conference on Network and Service Management (CNSM)\",\"volume\":\"53 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2022-10-31\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2022 18th International Conference on Network and Service Management (CNSM)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.23919/CNSM55787.2022.9965141\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 18th International Conference on Network and Service Management (CNSM)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.23919/CNSM55787.2022.9965141","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Intent-Driven Path State Monitoring to Enable Centralized State-Aware Flow Steering
Running state monitoring for network switches and links enables the derivation of a path-based data view. Thereby, various monitoring intents like utilization and latency awareness can be targeted during data collection. Whereas each of these objectives relies on a particular set of state metrics, different monitoring methods may be run to gather the data basis serving as decision input for subsequent analysis purposes. In addition, path conditions have an impact on the state observed for individual packet streams being forwarded along a specific path. While path level state data is of relevance, e.g., to evaluate past load ratios in order to run state-aware and efficient path determination, flow level state helps to monitor flow experience conditions like achieved throughput or perceived latency, e.g., to track the compliance with flow-based requirements. This paper presents a modular architecture for path state monitoring that considers port counter query, network probing and in-band network telemetry as methods for demand-driven data collection and focuses on utilization and latency awareness as monitoring intents. State data is collected by a centralized controller in collaboration with distributed modules deployed in a switch’s data plane to run data tracking, wherefore programmable switches are used as operational basis to ensure a flexible monitoring protocol. Evaluations show that continuously collected data snapshots allow to track accurate path state trends that – w.r.t. path state-aware traffic steering – can be leveraged to improve flow-based load distribution across available path capacities and to resolve inefficiencies like imbalanced path load or congestion.