Re: draft-shirasaki-dualstack-service-02.txt CPE assigned Host Address Lifetime issue

[JINMEI Tatuya / 神明達哉 <jinmei@isl.rdc.toshiba.co.jp>]

>>>>> On Mon, 16 Feb 2004 13:35:21 +0900 (JST), 
>>>>> SHIRASAKI Yasuhiro <yasuhiro@nttv6.jp> said:

>> If so, I'd say the host is not fully compliant (at least not compliant
>> to RFC3484), and network operators should not assume such a behavior.

> Exactly.

> I got your point, but if so, in the situation where delegated prefix is
> volatile (like PD with prefix-pool, as Green mentioned), how CPE should do?

> - CPE sets preferred lifetime in RA with very small value (ex. 1 min)
>   and advertises with a short period (20sec?).
> - CPE has a stable storage and advertises an old prefix with pltime
>   = 0, when CPE reboots accidentally and a delegated prefix is changed.

For the reboot case, perhaps a desirable behavior is:

- detect the old prefix is now invalid (as described in Section 12.1
  of RFC3633)
  
- advertise the old prefix with zero lifetimes for a while.  The zero
  preferred lifetime will deprecate the old address immediately.  The
  zero valid lifetime will at least reduce the corresponding valid
  lifetime to two hours, and the old address will be removed in two
  hours.

					JINMEI, Tatuya
					Communication Platform Lab.
					Corporate R&D Center, Toshiba Corp.
					jinmei@isl.rdc.toshiba.co.jp