[DNA-BOF] Applicable Environments for DNA

[Greg Daley <greg.daley@eng.monash.edu.au>]

Hi Jari,

This is a follow up to your post asking for
elaboration on the environments where DNA may be useful.

> 
>> * Define the problem scope, and environments where network
>>   attachment detection is desirable.
> 
> 
> Scope has to be defined. Can you give an example of
> what "environment" would mean in this case? Some properties
> of link layers perhaps?
>

Hmm, I guess that the wording is so foggy
as to mean nothing.

I think that the enviornments would be those
situations where such detection may be used to effect
in L3 signalling, or to feed L4 action (concentrating
on L3 though because there are candidates for this
already).

Originally I was thinking that Movement detection for
roaming hosts is a clear domain for this work,
although I've been encouraged to think more diversely
on this topic.

For example, even if IP mobility signalling or DHC is not
required, a host joining a link may need to discover
that it needs to join multicast groups.

Principally, the system needs to desire a rapid
configuration or response to attachment.

The effect of link types will also be an issue.
For example, if a PPP session will permanently (?)
be available on a particular interface, then
there's no need to undertake attachment detection,
since this interface will never receive L3 messages
indicating that configration is required.

Some network topologies make the work of network
attachment detection more interesting/complex.  It
will be worth defining those topologies where
this is likely to be an issue or a non-issue.

My opinion about this is not fully formed yet
though.


The problem scope does have to be defined.
We may be able to do this succinctly without
oversimplifying it, but there may need to be
some more discussion as to what we're aiming
to do generally.

Your post regarding configuration is a good
way to start this discussion.

I'm thinking that the aims section of the
BoF description will have to be rewritten...

Greg