Re: [DNA] A comment on draft-pentland-dna-protocol3-00.txt

[Brett Pentland <brett.pentland@eng.monash.edu.au>]

Hi Tero,

I'm not sure about that one.  If the host has been previously assigned
an address using DHCP, then it could easily extract a prefix from that
to use as a landmark.  The problem would be how the routers found out
about the prefixes.  I guess they have to know about them but I'm not
sure how DNA would be supposed to access them.  Would it be reasonable
to take them straight from the routing table?  I don't know if that's
a good idea.  Other than that, we could have an applicability statement
that just makes such configuations out-of-scope but that's not a very
satisfying answer.

Any thoughts?
Brett.

Tero Kauppinen (JO/LMF) wrote:
> Brett,
> 
> I found this while digging into the DT mail archive. Erik educated me 
> once with the following:
> 
> "Even if DHCP is used for address autoconfiguration, there is normally 
> significant benefit of telling the hosts the on-link prefixes.
> 
> If the routers don't advertise any on-link prefixes, then the hosts will 
> have to send all packets to a router (even when the destination is on 
> the local link). RFC 2461 does have a redirect message which the routers 
> can use in this case to tell a host "the destination is on link - at 
> this L2 address". But for a normal broadcast medium, having the hosts 
> know the on-link prefixes and using NS/NA to find the L2 addresses is 
> more efficient."
> 
> However, it's still valid to have a network where the router advertises 
> no prefix in the router advertisement but as pointed out above it is not 
> an optimal way of configuring the network.
> 
> /Tero
> 
> Tero Kauppinen (JO/LMF) wrote:
> 
>> Hi Brett,
>>
>> Hopefully this has not already been brought up in one of those mails 
>> on the mailing list. I tried to quickly look through the issues in 
>> other mails and the eye didn't catch this one.
>>
>> We were discussing this on the design team mailing list sometime ago, 
>> but I don't remember what was the conclusion. Anyway, the case when 
>> the router is not advertising any prefixes, e.g. the case when stateful
>> autoconfiguration is applied, is not clearly stated in the 
>> specification how to proceed (i.e. obviously no landmark prefix can be 
>> picked, CPL list can't be constructed). Or is it enough just to state 
>> the following?
>>
>> 6.2.6.3  DNA and Statefully Configured Addresses
>>
>> ...
>>    signaling.  If, however, a non-DNA RA is received, the host SHOULD
>>    use the DHCPv6 CONFIRM message as described in RFC 3315 [8] rather
>>    than wait for additional RAs to perform CPL, since this will reduce
>>    the amount of time required for the host to confirm whether or not it
>>    has moved to a new link.
>>
>>
>> /Tero
> 
>