PUFF: A Passive and Universal Learning-based Framework for Intra-domain Failure Detection

Ye，Lianjin1,2; Li，Qing2,3; Zuo，Xudong1,2; Xiao，Jingyu1; Jiang，Yong1,2; Qi，Zhuyun2; Zhu，Chunsheng2,3

10.1109/IPCCC51483.2021.9679436

2021

1097-2641

978-1-6654-4332-6

Conference Proceedings of the IEEE International Performance, Computing, and Communications Conference

2021-October

1-8

29-31 Oct. 2021

Austin, TX, USA

The increasing amount of network devices brings significant improvement to network quality but is inevitably prone to various failures. The frequent occurrence of link failures and node failures in the real-world network, causing packet losses and delay, calls for more accurate and fast detection methods. Existing network failure detection systems focus on probes and end-to-end metrics, but are limited by overhead on bandwidth or storage. Reliance on specific deployment of monitoring systems on devices like hosts also limits the feasibility and compatibility in general network topology, ignoring the potential of transferring monitoring tasks from hosts to switches. In this paper, we propose PUFF, a passive and data-driven network failure detection system based on in-network feature collection in programmable switches and machine learning algorithms. First, PUFF explores the potential use of continuous traffic changes to detect node and link failures instead of end-to-end metrics. Second, PUFF offers a software-based prototype and compares its performance with the latest passive failure detection methods. Evaluation based on simulation on real-world topology shows that PUFF can detect nearly 90% node failures and 80% link failures with less overhead in a shorter time.

Machine learning algorithms Network topology Prototypes Packet loss Feature extraction Topology Delays

通讯

英语

CPCI-S ; EI

WOS:000779177500058

20220911710139

Digital storage ; Machine learning ; Software prototyping ; Topology

Data Storage, Equipment and Techniques:722.1 ; Computer Programming:723.1 ; Machine Learning:723.4.2 ; Combinatorial Mathematics, Includes Graph Theory, Set Theory:921.4

2-s2.0-85125189310

Scopus

1.Tsinghua Shenzhen International Graduate School,Shenzhen,China
2.Peng Cheng Laboratory,Shenzhen,China
3.Southern University Of Science And Technology,Shenzhen,China

Ye，Lianjin,Li，Qing,Zuo，Xudong,et al. PUFF: A Passive and Universal Learning-based Framework for Intra-domain Failure Detection[C],2021:1-8.