Naser Ezzati-Jivan has been an open source developer and believer for the last 12 years. He is currently a research associate in the DORSAL lab at École Polytechnique, Montreal, Canada. His research interests include performance debugging and analysis of distributed systems, operating systems, and networking, areas in which he has published numerous scientific articles.
English session - Intermediate
We present an automatic analysis method that clusters requests, based on their response times, extracts and organises their features in a tree-based data structure, and mines the association rules among the different groups of requests. This aims to comparatively analyse the groups of slow web and database requests to find the possible root-causes of their latency problems. Three real world use-cases will show the usefulness of this method.