[SWMP] Re: faster field messages

[[e-mail]]

Jay:

I have not been thinking about the whole issue of doing something on one 
channel and using it on another.  This will require careful thought, but 
it could be very important.

As I mentioned elsewhere 
(http://www1.ietf.org/mail-archive/web/swmp/current/msg00045.html) we 
might need to subscribe on a reliable channel for notifications that will 
be delivered on an unreliable channel.  This sort of semantic mixing of 
the multiple channels was not spelled out in the paper, nor is it 
addressed in my follow up.  (In fact, I assumed the channels were 
independent and could not use each other's information.)

We need to think this through.  It is likely to be difficult to get right. 
 Nevertheless, it might be exactly the sort of innovation that would make 
swmp invaluable.

jfp

"Jay C. Weber" <[e-mail]> wrote on 08/29/2007 12:09:17 PM:

> [e-mail] wrote:
> So each side gets to declare its own mapping and the other is 
> obligated to keep track of that.  Servers can use the same id for 
> everyone becasue they get to declare the id they'll use.  Can I 
> declare a mapping on a reliable channel (TCP) and use it on an 
> unreliable channel (UDP)?  That would be helpful, since the first 
> use establishing the mapping should be protected from loss. 

> Exactly the way I was thinking, the first message gets sent reliably
> and enables a stream of very-low-latency, potentially-unreliable, 
updates.
> 
> jay

_______________________________________________
SWMP mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/swmp