On 3/14/11 3:17 AM, Luis Carlos Maqueda Ara wrote:
According to section 3.3 (Host to Router Interaction), a host only
registers non-link-local addresses with routers:
When a host has configured a non-link-local IPv6 address, it
registers that address with one or more of its default routers using
the Address Registration option (ARO) in an NS message.
FWIW While 3.3 says that, there isn't any text which forbids a host from
registering a link-local address.
This makes me think that:
(1) The only valid source address a host has at bootstrapping, is the
link-local address.
(2) Link-local address are not registered with routers
(3) If the default router list is empty, the case is the same that at
bootstrapping: the only valid address is the link-local address.
(4) When a router becomes unreachable or lifetimes are about to expire,
then it is likely that the corresponding router already has a NCE for
that host.
Due to (1), (2), and (3), the only NCE a router may create shall contain
the host's link-local address. As this address is not registered (no NS
with ARO for registration of that address will be sent), the NCE will
inevitably expire TENTATIVE_NCE_LIFETIME seconds after its creation.
Then, I do not see many situations when the effort of creating TENTATIVE
NCEs upon arrival of RS messages is worth it. Indeed, as far as I
understand, multiple factors must occur in order that the creation of
the TENTATIVE NCE makes sense and it ends up becoming a REGISTERED NCE.
Please, correct me if I am wrong, but I would say that this creation of
TENTATIVE NCEs is in most cases undesirable and, at least, some
clarifications regarding it would be required in the draft.
Why would it be undesirable? I don't see where it would cause any harm.
Section 6.3 merely says that the router MAY create the tentative NCE.
That was added to allow for routers which need an NCE in order to send
e.g., a Router Advertisement.
Do we need to forbid that?
Erik
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan
