I have some question about the WS4EE performence.
In our application we used four different EJB WS4EE Endpoints.
After testing the application the average time, from the invocation statistic on the web-console, for the EJB RemoteInterfaces are three to four times higher than the average time of the ws4ee servlet statistics acording to the EJB-Interface.
Four times seems to much, we have no complex SOAP-Messages to parse, most messages containing only one BASE64 Encoded byte Stream
So I asked:
Is it normal that an EJB Endpoint Servlet(I know it is not really a Servlet) response needs fourtimes more responsetime than the according EJB Method ?
Some clients have problems with old netinfrastructure.
Could it be that the Servlet-Statistic measure from request income
to a complete response, so that the nettraffic time of the response to the client is in the statistic?
How measure the web-console the Servlet-Statistic?
Also welcome are tuningtips.
And carpet me if I understand something wrong.