Re: [v6ops] SLAAC renum: Problem Statement & Operational workarounds

[Fernando Gont <[e-mail]>]

On 30/10/19 11:07, Ole Troan wrote:
> Ted,
> 
>>> We check both, the handling of the IA_PD and the error message.
>>
>> And to be clear, you mean that if the CPE asks for a prefix delegation and doesn’t get the prefix it previously had, it deprecates it as described in L-14?  When this deprecation happens, what ways are being tested for it to happen?
>>
>> E.g., is it the case that the client sends an IA PD containing an IA Prefix option, is the server returning the IA Prefix option containing the same prefix, with a status code encapsulated in it?   What status code?   Or is it returning an IA Prefix option with a different prefix?   Or is it doing both?
> 
> You are thinking about this wrong.
> The simple implementation of uRPF/BCP38 is: on LAN ingress to do a lookup in the FIB of the SA, if the resulting adjacency/interface matches the ingress interface allow packet otherwise drop packet and send ICMP message. This behaviour is required regardless of renumbering or not.
> The ISP is required to do this on their ingress interfaces too.

You mean required as in "MUST"? If so, where?


-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492