Academic Report: Police Aware In-band Network Telemetry for SDN Fault Localization

Publisher:网络空间安全学院Release Time:2019-06-12Times Views:33

Reporter: Prof. Tang Yongning, Dean of School of Information Technology, Illinois State University

Date and Time: June 13, 2019 10:00 AM

Location: Conference Room 412, Computer Building, SEU Jiulonghu Campus

AbstractIntelligent Fault localization for SDN becomes oneof the most critical but difficult tasks. In this talk, I introduceanapproach called Policy-Aware In-band Network Telemetry(PAINT) to tackle SDN fault localization. In the PAINT system,network operators define and deploy network services using ahigh-level ServiceProvisioning Language (SPL). Then, PAINT canautomatically parse the service policy to infer the causal relationshipbetween service related network components and end-to-end service-level observable symptoms. Based on the causality model, PAINT deploys monitoring instruments for the symptoms.PAINT utilizes a dynamically created Symptom-Fault-Telemetrymodel to incorporate In-band Network Telemetry (INT) actionssystematically into the fault reasoning process to improve theefficiency and accuracy of fault localization for SDN. PAINT hasbeenevaluated in a simulation environment for its accuracy and scalability with very positive results.

About the reporterYongning Tang isa professor,the headof network engineering program and thegraduate director in the school ofinformation technology at Illinois State University.His primary research areas arefault diagnosis in modern networkedsystems andartificalintelligence onnetworkoperations.