Re: [Ospf-wireless-design] OSPF Wireless Design Team Minutes

Hi All,

On Wed, Mar 09, 2005 at 01:49:29PM -0500, Acee Lindem wrote:
> All,
> 
> Yesterday morning Tom Henderson, Thomas Clausen, Joe Macker, Emmanuel 
> Baccelli,
> and I got together to design team issues, progress, and actions. This 
> E-mail includes my
> notes from the meeting. Please feel free to add comments and corrections.
> 
> One of the first things we discussed was how to handle the evolution of 
> requirements.
> All of us agreed that if we were going to make any progress we needed to 
> stick to the
> original charter of the design. Joe reiterated that the MANET WG is 
> actively pursing
> protocols optimized to MANET environments including reactive protocols. 
> I added
> that it would be interesting to see how either the protocols and/or 
> concepts could
> be seamlessly integrated with OSPF (however, that also is not within the 
> scope
> of the design team).

I agree that this is not within the scope of the DT right now.  However,
we must be cognizant of this fact.  Meaning, the DT should minimize the
changes to OSPF and the optimized flooding scheme, realizing that the
MANET WG may come out with a different flooding scheme down the line that
may need to be adopted.   As we all decided in the DT charter, we
should minimize the changes to OSPF, and any changes should be able to
work with standard OSPF nodes.

This is the approach that we (Cisco) have taken in our Overlapping Relays.  
The Overlapping Relays reduce the overhead in flooding as compared to 
standard OSPF (I will share some internal scalability testing that we did
in another email), without much changes to the way flooding works in OSPF.  
Essentially, Overlapping Relays suppresses redundant transmissions by 
neighbors without compromising reliability.  It uses timers and 
knowledge of 2-hop neighborhood which is already available in LSAs (this 
can also be done via Hellos if necessary), without modifying/affecting 
adjacencies, data plane, etc.  At most, it slightly delays flooding, but 
within the retransmission time.  Studies on packet delivery ratio and
throughput have shown favorable results (again, I will share our
scalability testing in another email.)

We have been receiving very positive feedback on its behavior *and* 
simplicity/compatability from our customers that have been using code 
drops of our implementation.

Thanks,
Madhavi

_______________________________________________
Ospf-wireless-design mailing list
Ospf-wireless-design@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/ospf-wireless-design