Re: [DNA] Simple DNA [Fwd: I-D Action:draft-krishnan-dna-simple-03.txt]

[Bernard_Aboba@hotmail.com]

> I have a question.
>
> Assume an unknown prefix arrives from an unknown router
> before any NA arrives from a known router.
>
> Shall a DNA host immediately configures a new address &
> send a BU (in case of MIPv6) or
> wait a little further for a might-be NA from a known router?
>
> If the latter is the case,
> how long shall the host wait? A second?

[BA] This is a good question -- and one I think that the draft should
answer.

> thanks for your clarification. I had assumed that the case of Steps 1-3
> is not common, whereas you assume it is.

[BA] Perhaps it would be useful for the Applicability section to discuss
this assumption and the situations in which it would be valid (and not
valid).

For example, if we are talking about a vehicle moving on a highway,
the assumption might not be valid, whereas if we are talking about
a consumer who frequently moves from the home, a work office to
the (same) coffee shop, then it would be.

> ok. However, I wish we use the term 'valid address' in more clear
> and precise way, lest there should be confusion.
>
> The term 'valid address' may mean
> 1) an address which a host can use on the currently attached link or
> 2) an address with valid lifetime.

[BA] I agree that the term needs to be defined in the terminology section.
In DNAv4, the term "operable address" is used, with the following 
definition:

"   Operable address
      In this specification, the term "operable address" refers either
      to a static IPv4 address, or an address assigned via DHCPv4 that
      has not been returned to the DHCP server via a DHCP RELEASE
      message, and whose lease has not yet expired."

This seems close to 2) above.

> Is 'manually assigned address' different from 'statelessly autoconfigured 
> address'?
>
> I assumed that latter includes the former. In case I miss something,
> kindly let me know.

[BA] As used in DNAv4, the term "manually assigned" meant an address that 
was fixed
and not assigned via DHCPv4.    So the question is whether such a concept 
makes sense
for DNAv6.  As Suresh noted, if we assume that it only makes sense to assign 
an address
that is valid given the advertised prefixes,  then the handling of a 
"manual" address would
not be very different from a "statelessly autoconfigured address", assuming 
that it is placed
into the address cache with an infinite lifetime.