Re: A common problem with SLAAC in "renumbering" scenarios

[Jan Zorz - Go6 <[e-mail]>]

On 04/02/2019 01:51, Mark Smith wrote:
> So I think it is worth finding out exactly why ISPs are doing this, to
> then work out what the best solution or solutions are.
> 
> Are they doing it
> 
> 1. because they're used to giving residential IPv4 customers dynamic
> addresses by default, so they've carried that practice over to IPv6
> without realising there are different and avoidable consequences?

Yes.

> 
> 2. their employer charges a premium for static address assignment in
> IPv4, and their employer assumes or expects, possibly implicitly, to
> do the same with IPv6?

Yes.

> 
> 3. because PD privacy e.g. German example?

Yes.

> 
> 4. to suit occasional flash renumbering?

Yes.

Cheers, Jan

> 
> 5. ???
> 
> In some of these cases, education may be a better, easier and quicker
> solution in comparison to the years it'd take to deploy CPE behaviour
> changes.
> 
> I put options in radvd that tried to mitigate this nearly 10 years ago
> (DeprecatePrefix, DecrementLifetimes), as radvd was popular in IPv6
> CPE I was working with at the time. Assuming radvd is still popularly
> used, if those options haven't been widely enabled and are not well
> known, then that might indicate an education and expectations problem.
> 
> Ultimately though, the residential ISP I worked for decided to give up
> the specialness and price premium of static addresses and gives all
> customers their own persistent stable PD prefix tied to PPPoE
> username. I've had the same one at home for 7 years.
> 
>>
>>>> The CPE should -- if possible -- be faithful to its LAN hosts, and
>>>> advertise if previous contracts between the CPE and the LAN hosts are
>>>> void. i.e., if the CPE does  not get leased the same prefix as before,
>>>> it shoudl notifiy its "clients". However, possibly for simplicity sake,
>>>> CPEs don't record what
>>>> information was previously advertised on the LAN -- they are not
>>>> required, so.... when they reboot, they may not not be in a position to
>>>> notify hosts accordingly.
>>>>
>>>> That's the environment hosts operate in -- no matter whether you or me
>>>> like it.
>>>>
>>>> In that environment, hosts can and should be smarter.
>>>
>>>
>>> Multipath transport layer protocols for the win. They're splitting
>>> identifier semantics off from IP addresses.
>>
>> That's mostly irrelevant -- because the thing here is to get rid of the
>> invalid addresses which tends to be benefical for anything running on IP.
>>
>>
> 
> I don't think it is. Invalid addresses become far less of an issue
> when a transport layer connection doesn't depend on a single address
> at each end and that it is valid all of the time.
> 
> Regards,
> Mark.
> 
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> [e-mail]
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
>