Hi Perrin -- > That sounds cool, but how important is it really? I'm not > sure any of > these solutions (including the commercial ones) do that level of > seamless failover effectively.
I know the CPE will pull faulty servers our of the pool, but it does so as a separate process (not by backing out of a bad request). One process constantly monitors servers, which informs the load-balancing process as to server availability. Constant monitoring was done by way of a request made to each server in the cluster every few seconds. There were two options. The first was a TCP request/response which was customizable. The second was by way of a request to a daemon process which was expected to return a server performance metric. This latter method was an open architecture -- the user was expected to write the daemon and implement the metric to work on a scale of 0 to 100, IIRC: 0 == server down, 1 == lightly loaded, 100 == heavily loaded. TTYL, -Jesse- -- Jesse Erlbaum The Erlbaum Group [EMAIL PROTECTED] Phone: 212-684-6161 Fax: 212-684-6226