Re: [dhcwg] WGLC on draft-ietf-dhc-sedhcpv6-21 - summary

[Ted Lemon <[e-mail]>]

It might be worth evaluating how commonly SAVI is deployed in practice before becoming concerned about this.   Also, since SAVI actually solves some of the same problems as sedhcpv6, it seems reasonable to think that the availability of two incompatible solutions may still result in more overall win.   A way to address this in the document would be to simply say that at present, the two approaches solve related problems but are mutually incompatible, so the choice must be made between one and the other.

This is also not a very serious problem in the sense that RFC 7513 proposes a DHCPv6-only address configuration model, which is not very common in practice.   In practice, SAVI First Come First Served (RFC 6620) should be compatible with secure DHCPv6, and will generally produce better results.

For sites that want absolute control over address assignment through a DHCP server, the options are either to not use Secure DHCPv6, or to implement an updated SAVI that would have to be written up that takes the assigned IP address out of the un-signed wrapper, or else uses layer 2 relay in combination with putting the IP address in the relay wrapper, as you've been advocating, Fred.