Jonathan,
We don't need the ff:fe to distinguish between short and EUI 64
stateless addresses as the U/L bit discriminates between them. We
currently use the ff:fe as an extra check.
Once we start using ND or DHCP, we will need to maintain a neighbour
table and there will be even less need for the ff:fe. I therefore don't
object to using IPv6 addrs that have the form aaaa::1:xxxx in a future
change to 4944.
Daniel.
Jonathan Hui wrote:
Hi Daniel,
On Mar 30, 2010, at 10:47 AM, Daniel Gavelle wrote:
This mean we can safely determine whether an IPv6 address has been
derived from a short address or a full EUI. We use this to determine
whether to use short or extended layer 2 addresses when transmitting.
This is a separate issue that we've discussed several times on this
list. You are assuming that there is a well-defined mapping between
IIDs and link-layer addresses in 6lowpan networks. Currently there is
no such assumption.
The IPv6 addressing architecture does not make any assumptions about the
relation between IIDs and link-layer addresses. The current 6lowpan-nd
and nd-simple drafts maintain this assumption by establishing a mapping
between IPv6 and link-layer addresses (taking the place of address
resolution in RFC 4861).
If we continue to make no assumption about the relationship of IIDs and
link-layer addresses, then you cannot safely assume that an 'ff:fe' IID
was derived from a short address.
--
Jonathan Hui
--
__________________________________________________
Daniel Gavelle, Software Engineer
Tel: +44 114 281 2655
Fax: +44 114 281 2951
Jennic Ltd, Furnival Street, Sheffield, S1 4QT, UK
Comp Reg No: 3191371 Registered In England
http://www.jennic.com
__________________________________________________
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan
