Hi, I wrote a simple test of performance using camel on ServiceMix 5.4 Here are the details : - Both of my tests are using an uncompressed request/response - The client is a load test under jmeter with a call frequency of 20 req / s - The provider is a Mock SOAPUi with one reply of 20k - During the load test : no resources was overloaded. - Jmeter, SOAPUi, ServiceMix are talking together on the same computer over the loopback interface. - There is no other bundle than these required to my test on servicemix
First test detail: - Call from JMeter to SOAPUI (without ServiceMix intermediate) - Test Duration : 30s - Max Time result : 190ms Second test detail : - Call from JMeter to SOAPUI with ServiceMix intermediate With Camel route <camelContext xmlns="http://camel.apache.org/schema/blueprint";> <route> <from uri="jetty:http://0.0.0.0:8081/?matchOnUriPrefix=true"/> <to uri="jetty:http://localhost:8080/?bridgeEndpoint=true"/> </route> </camelContext> - Test Duration : 30s - Max Time result : 1290ms Second test detail : - Call from JMeter to SOAPUI with ServiceMix intermediate (without camel) with java CXFRS bundle OSGI (using OSGI Container for transport) and Jetty client (for calling localhost 8080) - Test Duration : 30s - Max Time result : 210ms As you can see, camel test have some time results 10 times over the two others. This is very strange. For more : I ran new tests with camel-restlet or camel-cxfrs on "from" endpoint and http component for "to" endpoint. But all results keep the same. I also tried to define the size of jetty component' threadpool (Max 200 / Min 50). But with no best results. Do you have any idea? -- View this message in context: http://camel.465427.n5.nabble.com/Performance-Issues-with-Camel-tp5766380.html Sent from the Camel - Users mailing list archive at Nabble.com.
