{"title":"A Proof of Optimality on EDF Scheduling in Sink-tree Packetized Networks","authors":"Xuan Liu, Bingyang Liu, Shoushou Ren","doi":"10.1109/HPSR52026.2021.9481798","DOIUrl":null,"url":null,"abstract":"Earliest Deadline First (EDF) scheduling, is known to obtain the optimality of deterministic delay performance upon the single link. However, due to the generality of the network with multiple nodes and multiple flows transmitted within it, the optimal online scheduling is impossible. In this paper, we prove that EDF can still provide the optimal delay performance in a class of practical multiple-nodes-and-multiple-flows network scenarios, such as 5G Cloud VR/AR applications in up-link cases, namely in sink-tree networks multiple flows sharing the root node as the same destination node. The proof uses the delay-based schedulability region to quantify the performance of different scheduling policies. Finally, the proof indicates that EDF can achieve the largest schedulability region in sink-tree networks.","PeriodicalId":158580,"journal":{"name":"2021 IEEE 22nd International Conference on High Performance Switching and Routing (HPSR)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-06-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 IEEE 22nd International Conference on High Performance Switching and Routing (HPSR)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/HPSR52026.2021.9481798","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Earliest Deadline First (EDF) scheduling, is known to obtain the optimality of deterministic delay performance upon the single link. However, due to the generality of the network with multiple nodes and multiple flows transmitted within it, the optimal online scheduling is impossible. In this paper, we prove that EDF can still provide the optimal delay performance in a class of practical multiple-nodes-and-multiple-flows network scenarios, such as 5G Cloud VR/AR applications in up-link cases, namely in sink-tree networks multiple flows sharing the root node as the same destination node. The proof uses the delay-based schedulability region to quantify the performance of different scheduling policies. Finally, the proof indicates that EDF can achieve the largest schedulability region in sink-tree networks.