Usually in these sort of scenarios, I've seen something like this:

INTERNET
----------------------------------
FIREWALL
----------------------------------
DMZ   Web Server, Local Dir, etc.
----------------------------------
FIREWALL
----------------------------------
APP SERVER, DATABASES, etc.
----------------------------------

I would assume the swing stuff makes it a pain but you
could serialize swing requests into XML/SOAP and tunnel
them through a proxy.



>From: Tom Jansto <[EMAIL PROTECTED]>
>Reply-To: [EMAIL PROTECTED]
>To: [EMAIL PROTECTED]
>Subject: system configuration question using ejb's
>Date: Sun, 21 Apr 2002 08:09:43 -0500
>
> > hey folk,
> > this may be a bit off topic, so i'll apologize up front if so.  we have
>an
> > application that uses a combination of browser and swing based code to
> > access our ejb application.  we have the ejb server (jboss) setting in a
> > dmz.  the dmz is a straight forward deployment (cisco switch setting
> > behind a router, feeding three firewalls that are load balanced, these
> > feed another load balancer that distributes the load over 4 jboss linux
> > boxes.  oh yeah, there are some hardware ssl accelerators hanging off
>the
> > inner most load balancer for both the browser and client app traffic.)
> > anyways, what we've had to do is crank the firewall session way up to 24
> > hours, (nokia recommends 30 seconds), but that was the only way to make
>I
> > knew of (that made sense) to keep from having to periodically
>re-establish
> > the db connection pools into our database.  we still have to make sure
>we
> > push a 'ping' from the connection pool down to the db itself.  the
>problem
> > we are seeing is that the db connections are showing as being closed
>from
> > the db's view, but the firewall logs don't show any drops or losses.
>the
> > question boils down to this, is there something missing in the way this
>is
> > configured?  we need both browser and client app  access into the ejb
> > server, so we need the server in the dmz (otherwise we hit the rmi
>behind
> > a nat issue).  but the firewalls and load balancers all seem to prefer
> > stateless connectivity flowing through them.  has anyone gotten reliable
> > results, and if so, can y'all point me to appropriate sources, to get
>this
> > to be more reliable?  thanks, I realize there are a hundred additional
> > details that can be delved into, i'm just trying to determine where to
>go
> > next in the system architecture.
> >
> > tom jansto
> > [EMAIL PROTECTED]
> >
> >
> >
>
>===========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff EJB-INTEREST".  For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".
>




_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to