Overview
This course aims to teach learners about coordinating network-wide monitoring in heterogeneous and dynamic networks. The learning outcomes include understanding sketch-based monitoring techniques, profiling hardware capabilities for sketching algorithms, optimizing measurement task placement and resource allocation, and deploying monitoring solutions in large networks efficiently. The course covers topics such as network monitoring advancements, handling network heterogeneity, utilizing sketch structures for profiling, and evaluating the performance of the monitoring framework. The teaching method involves presenting theoretical concepts, practical challenges, and evaluation results. This course is intended for network engineers, network administrators, and professionals interested in network management and monitoring in complex environments.
Syllabus
Intro
Advances: sketches & programmability
Network-wide monitoring
Increasing trend towards heterogeneity
Overlooking heterogeneity is costly
Place sketches & allocate resources?
HeteroSketch in a nutshell
Profiler: Goal & Challenge
Sketch structure simplifies profiling
Micro-benchmarks: Device complexity
3 Phases: Micro-benchmark
Challenge: Scalability & Dynamics
Clustering strategy matters
Evaluation: Profiler
Evaluation: Optimizer scalability
Summary & Future work
Taught by
USENIX