Re: I-D Action: draft-ietf-6man-ipv6only-flag-03.txt

[Sander Steffann <[e-mail]>]

Hi Mikael,

> I want this RA flag to be the input of a decision the host makes to turn off IPv4, but one of potentially several. If you all of a sudden see this, don't turn off IPv4 immediately but perhaps wait 5-10 minutes since you last had working ARP or saw DHCPv4 reply. If you see it on network attach, do try DHCPv4 a few times and see if something answers. Perhaps do higher exponential backoff on your re-tries for DHCPv4 messaging.

This would take away most (maybe all) of my objections to this draft. An "if you don't get IPv4 working then don't worry, it's intentional" flag to change retries and timers wouldn't be a problem and would actually be very helpful.

> Since I don't know much about how hosts treat IPv4 internally and we don't have an v4ops group to discuss it in, I don't think this is 6mans job to come up with exact descriptions what host operating systems do with this information.

I don't agree with that. We can describe that intention without going into details. How about:

"""
On IPv6-only networks there will be no DHCPv4 server. Hosts that fail to acquire addresses with DHCPv4 and that see this flag set in all RAs from its default gateways SHOULD treat that in the same way as a network with the DoNotAutoConfigure option set as described in RFC 2563. A host SHOULD also increase its timers for retrying DHCPv4 to avoid causing unnecessary background noise on the network.
"""

> The only thing I as an network operator want is to lessen the load of multicasts/broadcasts in the air for devices never-endingly trying to do IPv4. From what I can tell, IETF has no working group with critical mass to work on this problem. Sunset4 is gone.

I recognise this desire, and with what is described above I would be happy.

My main fear is for SMB networks which often are still IPv4-only (which also needs to be fixed, but that's a separate issue), don't have managed switches or WiFi controllers with no RA Guard etc. There are many of such networks. Those networks are vulnerable to rogue RAs already, but happy eyeballs can mitigate the impact of someone messing with that. If this flag can kill IPv4 on the network then hosts become completely disconnected and happy eyeballs won't be able to mitigate anymore.

Yes, a serious attacker can do more harm even without this flag. This flag could however make it possible for kids/students/disgruntled employees/me/etc to bring down the network with a single packet. Such networks have no capacity for debugging anyway, and because of the low number of packets necessary to sustain the attack won't know what hit them. ISPs will get calls about the internet being broken, ICT consultants will spend hours searching for the cause of the problem, they will all blame "that annoying IPv6 thingy" etc etc etc. Let's avoid all of that.

Cheers!
Sander