Overview
This course on Differential Network Analysis aims to teach learners how to analyze network changes to predict undesired alterations in end-to-end forwarding behavior. The course focuses on introducing a new control plane analysis paradigm called Differential Network Analysis (DNA), which emphasizes quantifying differences in affected end-to-end forwarding behaviors. The course covers topics such as differential control plane simulation, modular three-stage framework, incremental computation, and customized graph algorithms. The intended audience for this course is network operators, engineers, and researchers interested in efficiently analyzing network changes and ensuring network stability and performance.
Syllabus
Intro
Existing control plane verifiers
Changes are often small Analyzing from scratch is wasteful
Hard to select properties to reverify
Make differences first class citizens!
Differential control plane simulation
Modular, three-stage framework Incrementally
Evaluation: end-to-end
Evaluation: three stages
Evaluation: generality & scalability
Taught by
USENIX