Re: [Isms] ISMS/SSH and notifications

["tom.petch" <cfinss at dial.pipex.com>]

----- Original Message -----
From: "David Harrington" <ietfdbh at comcast.net>
To: <Pasi.Eronen at nokia.com>; <isms at ietf.org>
Sent: Tuesday, May 27, 2008 2:26 AM
Subject: Re: [Isms] ISMS/SSH and notifications

<snip>

> >
> > (Using the same well-known port for both source and destination
> > wouldn't work in TCP.)
> >
> The WG decided that reuse of existing connections was acceptable, so
> one  connection could be used for both R/R traffic and notifications.
> I don't think we discussed the different port numbers used and how
> that might impact this approach. Are you saying that reuse might be a
> problem? Or that the application would need to deliberately request
> reuse of a (presumably) existing connection (by specifying  the right
> destination port)? (We did discuss that the receiving port would need
> to be listening for incoming traffic of both types.)
>
> I usually work with UDP and I don't think I've ever tried to use the
> same port number for source and destination. Is using the same port
> for source and destination a problem in TCP? or just for well-known
> ports? Is this something I should be recognizing as obvious and am
> simply overlooking? ;-)
>

I have seen the same (well-known) port used as UDP source and destination quite
frequently; I have wondered why and then why not?

TCP is different and the TCP source port is usually used cyclically, starting at
some value, may be determined by the implementation but sometimes appears to be
pseudo-random.  But does it have to be?

(I remain convinced that reusing a CR/CG connection for a NO+CR entity remains
the simplest deployment, even if the CR/CG is just used for the retrieval of a
one well-known object eg snmpTsmHereiamsendmenotificationsplease )

Tom Petch

> dbh
>
>
>
> _______________________________________________
> Isms mailing list
> Isms at ietf.org
> https://www.ietf.org/mailman/listinfo/isms

_______________________________________________
Isms mailing list
Isms at ietf.org
https://www.ietf.org/mailman/listinfo/isms