geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alan D. Cabrera" <>
Subject RE: Hot deployment
Date Tue, 20 Apr 2004 00:36:02 GMT

> -----Original Message-----
> From: Jeremy Boynes []
> Alan Cabrera wrote:
> > Simon,
> >
> > Dave and I are in the process of placing a layered pluggable
> > stack underneath Hiram's code.  It allows for a lot of nifty
> > such as Kerberos via SASL or GSSAPI; we also want to put xinetd
> > functionality in.  If I read Jeremy's email correctly, it is this
> > he is referring to.
> >
> > For remote JMX, MX4J has a perfectly good implementation and I think
> > that we should use it as is.  Are people thinking about using the
> > Geronimo network stack to support remote JMX?  It is not clear to me
> > what the advantages are of doing this.
> >
> > Do I understand this thread correctly?
> >
> Yep :)
> DB and I were not sure how it all plugged together.

Think async protocol that can be configured into a stack with each stack
having its own protocol name.

> The advantages I was thinking of were security etc. around the JMX
> invocation. If 160 handles all of this for us, then cool; I was just
> concerned that JMX over RMI/JRMP or whatever integrated with the rest
> security etc. so users don't end up having to define separate
> Seems to me that if we layer 160 on top of your stack, clients get the
> standard API and we get low-level integration - does that make sense?

Makes sense, share the same configurations/implementations.  I know that
you can plug your own transport protocol into JMX; this is pretty
straightforward.  One of the nice things that the Geronimo network stack
provides is the ability to have virtual circuits running through the
same addr/port.  We could have JMX traffic and others go through the
same TCP/IP session. 

Simone, what do you think?


View raw message