CSS 11150 Functionality

[andyhsu]

If you have a CSS acting as a loadbalancer for a number of web servers, how do requests for content get processed by the loadbalancer? Does it:

a) Pass the actual request along to the servers behind it

b) Act like a proxy and retrieve the content from a server and pass it back to the requesting client

I noticed in the advance configuration guide that there is a reverse-proxy function available. Is this what is necessary for option b? If so, can the caching part be disabled? We're using our CSS to load balance e-commerce webservers; it wouldn't make sense to have it cache the dynamic content such as a shopping cart.

--A.Hsu

[jlimbo]

To answer your question the CSS passes the packet to the servers behind it. All it does is use an algorithm like "round robin" to decide WHICH server it will send it to.

If you wanted to prevent the termination of the requests via proxying, then yes reverse proxy is one way. It will terminate on the caching device instead. If the content is indeed dynamic then it should be flagged in the http header with something like "pragma" or the "no cache" field is set. This is set by who ever the is writing the web content.