ghharrac @
ouray .
cudenver .
edu (George/Jorge) writes:
>We are currently developing an application to be used on
>the internet. The application must have the ability to
>function in conjunction with firewalls. Having looked at
>what information can freely pass through a firewall, we
>are considering using http. I am looking for white papers
>on http and keeping sessions alive for long periods of
>time. If I am correct http appears to be used for transaction
>based processing, unless java usage has changed things.
>Our primary concern is development of a CGI that can
>communicate directly with the outside user on a socket once
>the initial connection is made.
It sounds as if the service is initiated by a client behind the
firewall accessing a server out on the Internet, using a bidirectional
stream. This is a typical service supported by any decent firewall.
If you need a raw TCP pipe, just use one. Don't try to break the http
operating model. If a site wants to permit your traffic through,
they'll set aside a preassigned port for you. If they *don't* want
your application, it also lets them block it cleanly.
If the clients are out on the Internet and the servers must always be
*behind* the firewall, you need a different approach.
Rick.
smith @
sctc .
com secure computing corporation
|
|
