Re: [Roll] updating DAO caches (was Re: Something to ADD)

[JP Vasseur <jvasseur at cisco.com>]

On Nov 19, 2009, at 12:01 PM, Richard Kelsey wrote:

> > From: JP Vasseur <jvasseur at cisco.com>
> > Date: Thu, 19 Nov 2009 11:36:49 +0100
> >
> > On Nov 17, 2009, at 5:38 PM, Jonathan Hui wrote:
> >
> > > I'll have to admit that I'm not yet convinced of the benefits in
> > > storing DAO state as it currently exists in the draft.  [...]
> > I do agree with the first statement but not with the second  
> > paragraph.
> > I do see several deployment cases where no storing DAO would lead to
> > extremely sub-optimal paths and even more importantly traffic
> > congestion when getting closer to the root of course.
>
> Can you share these cases with us?  I agree with Jonathan
> that storing DAO states will typically not provide much
> improvement in P2P routing.

Sure. I'll take the example of an inter primary+secondary substation  
network
(could be the smart metering network) that supports traffic for  
various purposes:
meter read-out, DA alarms, ... etc. Thus traffic of various nature:  
P2MP, MP2P
and P2P. There are many such networks where not storing DAO just does  
not
work since all P2P traffic will have to transit via the root  
(unacceptable delays for
alarms) and the traffic around the root will be way too high.

This is why I am saying that having an RFC that forms a DAG that  
cannot be used
for outward traffic is a non-starter.

> > I do see several deployment cases where no storing DAO would lead to
> > extremely sub-optimal paths and even more importantly traffic
> > congestion when getting closer to the root of course. This is true
> > with several networks where the amount of P2P traffic may ned up  
> > being
> > not so negligible. The beauty with the current spec is that it allows
> > both deployment models.
>
> It claims to allow both, but fails to actually do so.  That
> is the point of the example that I gave and the reason I
> keep going on about this.  While the draft says that
> intermediate nodes can cache DAOs, it doesn't say how those
> caches are updated after a node changes parents.  And
> however this is done, we need to avoid the cost of sending
> cache updates in the absense of any actual caches.

Yes I agree with you on this and this is why DAO packing is also  
important.
Thus I would propose to all work on this and make it work while still  
allowing
some nodes to not store any DAO is they want to of course.

Does that make sense ?

Cheers.

JP.

>                                 -Richard Kelsey