NetMARKS: Network Metrics-AwaRe Kubernetes Scheduler Powered by Service Mesh
Lukasz Wojciechowski, Krzysztof Opasiak, Jakub Latusek, Maciej Wereski, Victor Morales, Taewan Kim, Moonki Hong
摘要
Container technology has revolutionized the way software is being packaged and run. The telecommunications industry, now challenged with the 5G transformation, views containers as the best way to achieve agile infrastructure that can serve as a stable base for high throughput and low latency for 5G edge applications. These challenges make optimal scheduling of performance-sensitive containerized workflows a matter of emerging importance. Meanwhile, the wide adoption of Kubernetes across industries has placed it as a de-facto standard for container orchestration. Several attempts have been made to improve Kubernetes scheduling, but the existing solutions either do not respect current scheduling rules or only considered a static infrastructure viewpoint.To address this, we propose NetMARKS - a novel approach to Kubernetes pod scheduling that uses dynamic network metrics collected with Istio Service Mesh. This solution improves Kubernetes scheduling while being fully backward compatible. We validated our solution using different workloads and processing layouts. Based on our analysis, NetMARKS can reduce application response time up to 37 percent and save up to 50 percent of inter-node bandwidth in a fully automated manner. This significant improvement is crucial to Kubernetes adoption in 5G use cases, especially for multi-access edge computing and machine-to-machine communication.
问问这篇 Paper
问问你的智能体。
Lune 读过与它相关的顶会 Paper,每个回答都会注明依据哪几篇。
引用它的顶会 Paper2
- Canal Mesh: A Cloud-Scale Sidecar-Free Multi-Tenant Service Mesh ArchitectureEnge Song, Yang Song, Chengyun Lu, Tian Pan 等SIGCOMM 2024 · 被引用 30 次
- BREAK: A Holistic Approach for Efficient Container Deployment among Edge CloudsYicheng Feng, Shihao Shen, Xiaofei Wang, Qiao Xiang 等INFOCOM 2024 · 被引用 11 次
相关 Paper
- Enabling SLO-Aware 5G Multi-Access Edge Computing with SMECXiao Zhang, Daehyeok KimNSDI 2026 · 被引用 4 次
- NeuRO: Inference-time Profiling and Orchestration of ML Applications at the EdgeArshad Javeed, György Dán, Viktoria FodorINFOCOM 2026 · 被引用 1 次
- Tailored Learning-Based Scheduling for Kubernetes-Oriented Edge-Cloud SystemYiwen Han, Shihao Shen, Xiaofei Wang, Shiqiang Wang 等INFOCOM 2021 · 被引用 93 次
- Real-Time Flow Scheduling in Industrial 5G New RadioTianyu Zhang, Jiachen Wang, Xiaobo Sharon Hu, Song HanRTSS 2023 · 被引用 6 次
- CoreKube: An Efficient, Autoscaling and Resilient Mobile Core SystemJon Larrea, Andrew E. Ferguson, Mahesh K. MarinaMobiCom 2023 · 被引用 21 次
