Hi Brandon, > Thanks Jerome. > I'll update what that issue w/ what I find. I am setting it > up so that we can > do some performance testing to compare the 2 approaches.
Very cool. > As i was digging around some more I noticed ServletCall is > obviously being used > since it is running the servlet container. I want to fully > use NIO. I am > running restlet in jetty and need to use the JettyCall > instead. It looks like > the getRequestChannel and getResponseChannel need to be > implemented to use the > NIO correct? Are there any other updates that need to be > made to run restlet > inside the jetty container and fully use NIO? That should be all you need it in theory. In practice, you will be limited by the fact that Jetty's NIO support is limited to the front-end of the connector (dealing with the socket and the thread pool). The back-end part is based on the Servlet API which forces you to see things as BIO streams. IMO, we'll only see the full benefit of NIO with Restlets when a dedicated connector, based on MINA or Grizzly (to facilitate the NIO processing). MINA is still digesting the merging with AsyncWeb but the Grizzly project recently became a top-level project (https://grizzly.dev.java.net/) and we might see some good progress on the Grizzly connector front soon. You can contribute to this new connector here: http://restlet.tigris.org/issues/show_bug.cgi?id=16 Best regards, Jerome
