Not nevcessarily. You could, for example, compute the hash based on the source IP - so all requests form one ip would be handled on one server :-)
Regards Thomas Tomiczek THONA Consulting Ltd. (Microsoft MVP C#/.NET) -----Original Message----- From: Peter Foreman [mailto:[EMAIL PROTECTED]] Sent: Donnerstag, 30. Mai 2002 09:48 To: [EMAIL PROTECTED] Subject: Re: [DOTNET] ASP.NET and HTTP request routing in a web farm --- Greg Reinacker <[EMAIL PROTECTED]> wrote: > There isn't a single load balancing server, per se. With NLB, there is a > virtual IP address of the cluster. Every server in the cluster will see > every request to the virtual IP at it's network interface. Then each > server individually execute a hash algorithm to determine if it should > service this particular request or not. Each server will come to the same > hash result, which leads to one and only one server being responsible for > a given request. I can see how that would work if they are all servicing different request types. But how does round robin work? i.e. How does a particular node know to response to a request where all nodes are applicable. You could start them off synchronised of course, but they could get out of synch. I assumed some kind of inter-node communication would take place per request. Thanks, Peter __________________________________________________ Do You Yahoo!? Yahoo! - Official partner of 2002 FIFA World Cup http://fifaworldcup.yahoo.com You can read messages from the DOTNET archive, unsubscribe from DOTNET, or subscribe to other DevelopMentor lists at http://discuss.develop.com. You can read messages from the DOTNET archive, unsubscribe from DOTNET, or subscribe to other DevelopMentor lists at http://discuss.develop.com.
