One suggestion: I think it would be nice to consider scalability - if you have one tomcat frontend forwarding to 100 backends and acting as a load balancer - you probably can't afford one connection per thread. Many of the http forwarders I know use a blocking http client library - I think this would be the wrong approach for tomcat.
To do non-blocking you will need to work at coyote level, and probably make few enhancements to register your client connection with the same framework - i.e. use the same Selectors or APR poll. You would also need to implement a non-blocking http client - you can use the nio/apr as starting points as well, but the code is a bit messy ( IMHO ). Or you could just start from scratch - use an existing non-blocking http client. I have a small experimental async http client in sandbox, if you plan to go non-blocking it may be worth checking it out (I'll update it to what I have - it's kind-of-working ). But probably using the NIO/APR connectors would be a better starting point. Costin On Wed, Apr 15, 2009 at 1:18 PM, jean-frederic clere <jfcl...@gmail.com>wrote: > Sharmistha jat wrote: > >> So, now i would try to install a tomcat cluster and make a simple >> servlet proxy [reverse] for it, then add in load-balancing logic to it >> > > Yep thst is the idea. > > +++ CUT +++ > > >>>> I have studied a number of servlet proxies like j2ep, noodle & Http >>>>> Proxy >>>>> Servlet in past days. >>>>> >>>> Could you write a wiki on that? (Somewhere in >>>> http://wiki.apache.org/tomcat/). >>>> >>> >> i guess u mean writing wiki about proxy servlet >> will give my best shot to it, but would borrow 3-4 days for the >> task. >> > > You looked to the j2ep, noodle etc, don't you? It would be nice to describe > a little how they work or that least their main features and the url to > there code if applicable. Just like a note on other existing implementation. > > Cheers > > Jean-Frederic > > > Cheers >>>> >>>> Jean-Frederic >>>> >>>> And was puzzled about my intent of making proxy and its designing >>>>> thereof. >>>>> >>>>> So, please guide me a little bit about this. >>>>> >>>>> Thanks >>>>> Sharmistha >>>>> >>>> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org >> For additional commands, e-mail: dev-h...@tomcat.apache.org >> >> >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org > For additional commands, e-mail: dev-h...@tomcat.apache.org > >
