cschneider wrote > I wonder why ECF and Amdatu would be so much slower. Maybe there is just > something wrong with the setup.
We suspect the problem with Amdatu was, that it does not use connection pooling which makes it suffer heavily from many very short invocations. As for ECF, it's likely that there is something wrong with the setup, especially since it gets slower the more I use it. The benchmark is now available on github[1], maybe Scott can try it out and clearify. For the ECF test I was running two Karaf 4.0.4 instances on the same host with the ECF feature installed. The property 'zoodiscovery.autoStart=true' was set in system.properties and I had a Zookeeper 3.4.6 running on the same machine. The test was performed with "benchmark:run 1000 5 localhost" (1000 invocations, 5 threads, RMI baseline test against localhost) Johannes [1] https://github.com/jutzig/remote-osgi-benchmark -- View this message in context: http://karaf.922171.n3.nabble.com/Proposal-Lightweight-standalone-remote-OSGi-implementation-for-karaf-cellar-tp4045343p4045409.html Sent from the Karaf - Dev mailing list archive at Nabble.com.
