Re: [DNA-BOF] [Requirement for Optimized DAD in IPv6 Mobility] isneeded ?

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

Hi Daniel,

This is a good start to the DAD requirements.


Soohong Daniel Park wrote:
> Hi all
> 
> Recently, I am trying to write DAD issues like 802.11, Optimized, 
> Applicability Statement, etc.
> During DNA meeting, I sensed that DAD issues should be discussed at DNA.
> In order to use the Optimized DAD in IPv6 mobility, I believe we have to
> 
> consider requirement of node in advance.
> 
> I've written rough document as follows
> http://home.megapass.co.kr/~natpp00/requirement.txt
> 
> After look into, please let me know your view on this.

I think that section 3.2 is restrictive in that
it can be used to prevent usage of the defined
Optimistic DAD system.

Do you mean that the optimized DAD scheme MUST NOT
modify standard (peer hosts?) neighbor cache operation

If you actually mean neighbor cache entries, then
for Optimistic DAD, in the case of collision, there
may be a neighbour cache entry modification, although
this will last for under a second.

I think that it may be worth while saying that the
Neighbor Cache entries MUST NOT be modified in a way
which is unrecoverable through standard RFC2462 DAD
operation.

Additionally, I think that we're defining requirements
for DAD optimization, so I don't think we can delegate
this task to SEND WG.

I think that as a requirement, the DAD optimization scheme
SHOULD NOT introduce any new attacks on ND and the
resulting schemes should be no worse than existing
RFC 2462 DAD.

You may add a statement indicating that:

These schemes SHOULD interwork with Secure Neighbor
Discovery DAD mechanisms as defined in SEND WG.

It is OK to move all of the security discussion to
the security considerations section.  This section
may have subsections.

How does this sound?

Greg