Hi Vicky,
Am 11.06.2015 um 16:36 schrieb Victoria Mercieca <vmercieca0@xxxxxxxxx>:
Hi Lotte,
I didnt think we mentioned OrigSeqNum in RREP Generation. But yes... if we
are using the seqnum TLVs to indicate what the address is (if its OrigAddr or
TargAddr) then we will need to include both OrigSeqNum TLV and TargSeqNum TLV
in RREQ and RREP. If we dont know a value for the seqnum, do we say to use
zero or say not to include a value?
In the RFC5444 representation section, we could say to include no value in
the TLV if the seqnum is zero or unknown from our data element (or will it do
this automatically? because absence of a value seems to mean zero in a few
other places?)
Also, how are we going to declare which addresses are the error destinations
in the RERR? Is the fact we have a SEQ_NUM TLV enough there? Currently we say
to include SeqNum if a seqnum is known for that route, so now we would have
to ensure to send a SeqNum TLV...
Regards,
Vicky.
On Thu, Jun 11, 2015 at 1:58 PM, Lotte Steenbrink
<lotte.steenbrink@xxxxxxxxxxxx> wrote:
Hi,
it just ocurred to me that, while we do have an OrigSeqNum TLV, 7.2.1. RREP
Generation
never instructs the implementor how to set it (it appears that I've
implemented this feature by accident, since my code does set and fill the
OrigSeqNum TLV :D). Therefor, I'd propose to add the following:
6. If available, include the last known sequence number associated with
OrigAddr.
Does this make sense?
Regards,
Lotte