Wolfgang Breitling wrote: >I believe it very much depends on the platform and potentially some IP >config parameters. It may also have changed over time with the advent of >the internet and the need for firewalls. If every new connection >requires a new - random - port # then there is a need to punch many >holes into the firewall. I believe that at point Oracle adopted the >option for all connections to use the same port. > >Henry Poras wrote: > > > >>Just wondering if one of my strongly held untested assumptions is wrong. I >>always thought that when you connected to Oracle via the listener in >>dedicated server mode, you would come in via the listener port (i.e. 1521, >>1526). The listener would then validate, spawn off a shadow oracleSID process >>using some arbitrary port, and the client would then be passed to this >>process with the listener butting out at that point. The client would then be >>connected to the database through that arbitrary port. >>Last week a developer and myself checked this out. He sniffed the coming and >>going of packets, and everything used only port 1521. I looked at netstat (I >>think the parameters were asp (RedHat AS3 OS) but I'm not sure. That info is >>at work but I'm not at the moment) and the only active Oracle ports were also >>1521. Are other ports used? Is this dependent on load? I searched metalink >>and found the question asked, but no answers were ever posted. >> >>Thanks. >> >>Henry >> >> >>-- >>//www.freelists.org/webpage/oracle-l >> >> >> > > > That's sort of what I was thinking, but I couldn't find it documented anywhere and I didn't want to base action on conjecture. Thanks. Henry -- //www.freelists.org/webpage/oracle-l