Re: Confirmation to advance: draft-ietf-6man-ipv6only-flag-05

[Philip Homburg <[e-mail]>]

>> Obviously, from an operational point of view, if DHCP service is
>> important on a network, the operator is at liberty to install redundant
>> servers / relays.
>
>WHAT ARE ARE YOU SAYING?  The IETF SHOULD NOT concern itself with
>accidental misconfiguration? And SHOULD only create fragile and brittle
>protocols? The IETF SHOULD NOT seek robustness in its protocols?
>
>If that is the case I think you just said half the work in V6OPS and GROW
>is out of the scope of the IETF. And you just said your objection to the
>IPv6-Only flag is out of scope. I believe just the opposite, and I believe
>your objections to the original formulation of the IPv6-Only flag were well
>placed.

The IPv6-only flag is an additional feature. Before accepting such a 
feature it is worth asking what happens if the feature gets enabled
accidentally or maliciously.

In my opinion that is different from creating workarounds for poorly
maintained services. The normal expectation of a DHCPv4 server failing
is that there will not be any IPv4. Creating workarounds for that creates
operational complexity. 

At the recently concluded RIPE meeting, Geoff Huston suggested that the
reason some networks have very high IPv6 failure rates is because the
failures are masked by happy eyeballs. Networks that rely on IPv6
(i.e. those with NAT64) have very low failure rates.
(https://ripe78.ripe.net/programme/meeting-plan/ipv6-wg/)