Hi Michael, On 24/05/2016 23:45, Michael Behringer (mbehring) wrote: >> -----Original Message----- >> From: Anima [mailto:[email protected]] On Behalf Of Brian E >> Carpenter >> Sent: 23 May 2016 22:45 >> To: Anima WG <[email protected]> >> Subject: [Anima] Anima and renumbering >> >> I made an unplanned experiment yesterday. Because of a discussion in >> 6man, I set up my laptop to run with only temporary IPv6 addresses (RFC >> 4941 addresses). And then I ran my tests of the GRASP prototype for a while. >> After half an hour or so, GRASP stopped working, because the laptop's >> address changed automatically, so the cached discovery results became >> invalid. >> >> So, one conclusion is that my implementation should be more aggressive >> about aging out the discovery cache. But it raises a question about how >> resistant we need the Anima infrastructure to be against renumbering. It >> seems to me that an autonomic mechanism really needs to repair itself in >> that case. Any thoughts? > > Hmmm... I'm not an IPv6 expert, so please bear with me. > > https://tools.ietf.org/html/rfc4862#section-5.3 states explicitly that "A > link-local address has an infinite preferred and valid lifetime; it is never > timed out."
Correct, although there is one corner case. If you build your link local address using the MAC address, and you are running with randomized MAC addresses, I suppose the link local address could become invalid. > > https://tools.ietf.org/html/rfc4941#section-3.3 implies that the procedure > applies only to global scope addresses: > "This document extends [ADDRCONF] as > follows. When processing a Router Advertisement with a Prefix > Information option carrying a global scope prefix for the purposes of > address autoconfiguration (i.e., the A bit is set), the node MUST > perform the following steps:" > > Therefore, link local addresses, in my reading, do not use temporary > addresses a la RFC 4941. Correct so far? Correct. > I would further claim that addressing inside the ACP does not require > temporary addresses, since the ACP is a secured overlay network. I believe that's correct, but probably it should be stated explicitly in the addressing specification, since stable vs temporary addresses is a very hot topic in 6man at the moment. > Therefore, if GRASP runs on link local addressing in the data plane, and if > we use static, greater-than-link-local scope only inside the ACP (which is my > assumption), we should be fine, right? Yes. But in the case that people deploy AN components without an ACP, which we have always said would be possible but not recommended, I think the issue could arise. Acually the possibility of data plane renumbering is a very strong argument for the ACP approach with ULA addressing, since it allows stable addresses for the ANI whatever happens in the data plane. So maybe this is something for the next rev of the reference model. Brian _______________________________________________ Anima mailing list [email protected] https://www.ietf.org/mailman/listinfo/anima
