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

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

On 12/02/2019 01:42, Mark Smith wrote:
> I was thinking earlier today that dynamic /64s on the PPPoE virtual
> link would also suffer from the same issues that dynamic PD prefixes
> does, and therefore thinking that I'd do static /64s on the PPPoE link
> too (and aggregate them upstream of the group of BNGs).

Surprisingly, this is not causing major problems. Similar to IPv4, where 
changing the public IPv4 address on the CPE that is doing NAT to the 
internal LAN doesn't have considerable effect on the hosts inside (well, 
some sessions could drop on change, but that's the price you pay with 
NAT). In IPv6, if source and destination address is the same - if 
network in between changes and re-establishes quick enough - that should 
work (and it works in deployments out there ;) )

> 
> However, different prefixes across link disconnect/reconnect is
> probably less of an issue when the device is a host itself - the
> disconnect/reconnect is most likely going to be caused by the host
> being rebooted/powered off/powered on, so the all of the transport
> layer and application state that would be disrupted by a re-addressing
> is discarded anyway.

Indeed.

> 
> Still, if the link is flakey or interrupted somehow, then a dynamic
> /64 prefix on the host to BNG link would still be disruptive to the
> host and its applications. For the best customer/end-user experience
> it would be best to have stable/static /64 prefixes on the WAN link
> too.

That (in theory) would be ideal, but reality sometimes needs some 
compromises ;)

Cheers, Jan