Re: Notification Message Processing Model

[Andy Bierman <ietf@andybierman.com>]

Shane Kerr wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Andy Bierman wrote:
>   
> > I am trying to understand how the protocol will have
> > to be modified to support notifications and RPCs on
> > the same channel-less connection.
> >     
>
> Rather than go into specific details, I would rather say that:
>
> - - Putting notifications and RPCs on the same channel adds complexity to
> the server.
>
> - - It also adds complexity to the client (but really only to clients that
> want it).
>
> - - It would be nice to live without the complexity if we can.
>
> I *think* this is pretty much what Andy Bierman has been saying.
>
>   

Exactly.
I've been taught the whole art of Engineering is
having just enough complexity, but no more.
(A concept totally foreign to the IETF.)


> As I see it, a client can usually multiplex two channels into one very
> easily. Can someone explain a case where this is not true? (One case
> presented in the working group where this may not be true is for clients
> that are managing tens of thousands of devices. However, I think both
> SSH and BEEP allow multiple channels on a single TCP connection, so we
> should probably consider the "massive management" case a non-problem.)
>
> Also, can someone give a specific example where putting notification and
> RPCs on the same channel would be simpler than creating a new channel?
>
> Apologies if this has already been given and I missed it!
>   

No -- you missed nothing.
I have thought about it as much as I can.
I can't come up with any use cases in which a robust
event management system would be simpler because
operations and notifications were mixed on the same channel.

I have only found bad things that can happen to your OPS
if you did this.  No real benefits at all.  I'd love to hear
why management stations need this feature.

> - --
> Shane
>   

Andy



--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>