Abstract

LAMP (Linux, Apache, MySQL/MariaDB, and PHP/Python/Perl) is a powerful and easy to use open-source web development platform which is commonly considered by many system administrators and developers as the platform of choice for deployment and development of dynamic web applications. However, root cause analysis is a challenging task with the LAMP stack because there could be several reasons (stemming from its various components and layers) behind the problems. Although there are some debugging tools for lamp stack performance analysis, it is still difficult to get a real insight into the functionality of web applications at different layers, when a bug or performance degradation occurs.

In this presentation, we introduce a unified LAMP debugging tool based on LTTng tracing tool. We also show the use of Trace Compass (a trace visualization and analysis tool) to analyse the LAMP trace data, to understand the real execution of the applications of interests, extract various runtime statistics, and detect their possible performance bottlenecks and problems. The presentation will be concluded by showing a demo of the tracing of a real website and detecting the reasons behind its slow response time.

Audience

System administrators and application developers interested in automated solutions to detect and diagnose unexpected latency occurring at multiple levels of distributed systems­.

Biography

Naser Ezzati-Jivan has been an open source developer and believer for the last 10 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.