HTTP Time Bandit
summary
A crawler measures per-URL download times, computes mean and standard deviation, and flags consistently slow endpoints for focused load trials; results are emitted as XML output and a Graphviz map. The tool reruns targeted tests with more sampling (example: 100 probes) to confirm true CPU/DB hogs and can run in parallel (demo used three instances) to drive an Apache box to ~98% CPU while the client stays near 9%. Code is published on GitHub.
AI-written summary · about
chapters (32)
- 0:00Introduction
- 0:25Who are we
- 1:07Agenda
- 1:34Why do we do this
- 2:16Taxonomy
- 7:34Exotic Authority
- 9:05Proposed Method
- 11:54Formal Method
- 12:43Linux
- 13:05Example
- 15:09Graph
- 18:16symmetric nature
- 18:52attack
- 19:02testing
- 19:42other tools
- 21:16what do we do
- 23:12load balancers
- 24:51a proxy
- 25:54protection services
- 29:22baseline
- 30:30security
- 31:52maximum connections
- 32:22mod2s
- 32:46settingstart
- 34:08Resource Consumption
- 34:37Motivation
- 34:59Evasion
- 36:17Honeypot
- 36:55Bad Times Many
- 38:29Tools
- 38:49Paper summation
- 41:17Conclusion
related talks
35mHTTP Time Bandit
34mPerformance Tools
41mMeasuring the Moving Parts of User Experience in Java/.NET applications
59mFaster Websites with WebPageTest
30mMeasuring Web Performance - Metrics & Tools
36mMoneyball for Performance Metrics
29mSite Availability is for Everybody
37mAutomated Performance Testing With WebDriver