Regarding end2end tests for longwell. There are some tools that could be useful for your purposes, such as: IPerf, netflow or ntop.
By means of these tools, network metrics (one way delay, roundtrip time) could be measured, so you could further determine application processing time (provided you configure NTP for timing synchronization). Regards, //david lópez Stefano Mazzocchi wrote: > Andrew Plotkin wrote: > >> We've been looking at various ways to improve Longwell's performance. This >> starts with figuring out what the *current* performance is, so I have put >> together a simple Longwell benchmarking tool. You run it from the command >> line; it does a (configurable) list of queries against your Longwell >> server, corresponding to the activity that would be generated by a user >> adding and removing facets. For each user "command", you get a timing >> measurement in milliseconds. >> >> This is a (nearly) end-to-end test. The timing info includes network time, >> but not the time taken by the user's browser to download and display >> individual item resources (such as images). >> >> https://simile.mit.edu/repository/longwell-bench/trunk >> >> See https://simile.mit.edu/repository/longwell-bench/trunk/README for >> details of use. >> > > Andrew, very nice! > > This gives me a way to understand the impact of Banach smooshing on the > overall longwell performance. > > Also gives me a way to stress test longwell on the latest sesame API, > perfect timing. > > Expect results here real soon. > > _______________________________________________ General mailing list [email protected] http://simile.mit.edu/mailman/listinfo/general
