On 03/15/2013 07:10 AM, Michael Richardson wrote:
Okay, so there would be standards work to do multi-master.
I thought so, but I wasn't clear.
Are there any multi-master implementations using proprietary protocols?
I think that multi-master isn't that important for homenet.
I think that having a stealth master for the reverse ZONE should be a MUST.
Homenet and code point of view, the important requirement would actually be:
CER MUST be able to be authoritative name server for
at least one reverse (ip6.arpa) zone, and at least one
forward zone.
CER SHOULD support zone transfers to a secondary server,
which SHOULD be configurable manually, and SHOULD be
configurable by some TBD DHCPv6 option.
I'm confused, is the DHCPv6 part connected in any way to the zone
transfer part? I thought the DHCP part that Simon was talking about
was to auto-populate the (master's) zone via an existing dhcp option
(and he's doing it via DHCPv4, not v6). I don't understand where DHCP
comes in with the ability to do a zone transfer to a slave.
CER SHOULD permit NS records for zones to be configured
manually, and via DHCPv6 option, and SHOULD support
situation where no NS record points at CER itself.
Are you saying that the CER would use the ISP's DHCP to get the
addresses of the authoritative slave servers? What if my ISP isn't
my DSP? I think we need to be careful with any such assumption.
Mike
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet
