Can folks share their experience with best-case and worst-case request times for the GAE/J application tier? I think its useful to isolate the application tier performance as there is a lot of good info on datastore performance and how to optimize it (e.g. use of Memcache, etc...)
The worst case would be the initial request from a cold start. If the worst case is greater than the deadline, please include any estimate on what percentage of cold starts exceed the deadline. The best case would be for everything loaded. I may have missed existing summaries in my search so a link to that may be the answer... Here is a list of performance profiles that could be useful to track: - best and worst case for pure static request that makes it to the Google Front End. This provides the best case baseline. - best and worst case for pure Servlet/JSP request doing Hello World. This provides a baseline for isolating the low level application tier processing. - best and worst case for a pure Java framework like Wicket doing Hello World . - best and worst case for a lean dynamic language framework like Gaelyk doing Hello World. - best and worst case for a heavy dynamic language framework like Grails and JRuby on Rails doing Hello World. Thoughts on how to deal with the worst-case response time are already being discussed (http://googleappengine.blogspot.com/2009/12/request- performance-in-java.html) and include pre-compilation, optimization of the start-up of the dynamic language runtimes and the use of keep- awake pings. This thread will hopefully allow people to share specific performance data that can motivate those efforts. Best regards, Gabe
-- You received this message because you are subscribed to the Google Groups "Google App Engine for Java" group. To post to this group, send email to google-appengine-j...@googlegroups.com. To unsubscribe from this group, send email to google-appengine-java+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/google-appengine-java?hl=en.