Terminator: An Efficient and Light-weight Fault Localization Framework
| Autor: | Ke Pei, Jinghui Li, Chengqiang Huang, Yuxing Li, Hu Zheng, Longbo Huang |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
021110 strategic
defence & security studies 021103 operations research Computer science business.industry 0211 other engineering and technologies 02 engineering and technology Topology Network topology Phase-locked loop Packet loss Server Peer to peer computing Data center Terminator (franchise) business |
| Zdroj: | INFOCOM Workshops |
| DOI: | 10.1109/infocomwkshps50562.2020.9163055 |
| Popis: | Network fault localization has always been a major challenge for efficient data center operation. In this paper, we propose a faulty link localization framework named “Terminator,” which provides an efficient hierarchical link probing scheme for fault localization. With both local and global probing, Terminator conserves the spine link bandwidth and localizes faulty link with high accuracy. We implement a prototype of Terminator and show that it outperforms existing benchmarks, including 007 [7], TOMO [5], PLL [6], and NetBouncer [2]. In particular, Terminator achieves an average of 37.5% internal problem-solving rate and improves the localization accuracy of 007 to nearly 100% in fat-tree topologies. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|