Hello I don't know if this solution is more effictive than the other one (load balancing by using Apache server with different Tomcat instances) But the two one don't realise a random access : -cluster uses "rules" -apache uses a "simple round-robin scheduling algorithm" ; so after a disconnection the new one is not certain to be affected to the same Tomcat's instance
Jean-Claude -----Message d'origine----- De : Zohar Amir [mailto:[EMAIL PROTECTED] Envoyé : mercredi 11 mai 2005 09:14 À : Tomcat Users List Objet : load balancing Hi, I have an application server that maintains session and I'd like to make it scalable. The clients and the application server are communicating using TCP. When a client tries to connect, I'd like the load balancer to connect to a random AppServer node, and from that moment on - that node should handle the session. The client can disconnect and reconnect and the load balancer should connect it with the node that handled the session before the disconnection. Is it possible implementing these requirements using the model described in http://jakarta.apache.org/tomcat/tomcat-5.0-doc/cluster-howto.html? Is there a better architecture that you can recommend which is more suitable to what I just described? Thanks. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]