Hi Med,
>
>Dear Tomek,
>
>Thank you for this clarification.
>
>Providing the IP address in the option is not flexible enough and especially
>it
>
>may not be recommended to achieve load balancing. Otherwise the DHCPv6 server
>should act as a load balancer, which is not currently our preference.
>
>To illustrate more the usage with the domain name option: The scenario I
>mentioned (is similar to what currently done in some VoIP networks for
>distributing customers among SBC nodes or P-CSCFs): you provide a generic
>FQDN
>of the AFTR by the DHCP server together with a suffix in the dns serach list
>option. When the B4 receives these two options, it will form a the FQDN to
>use
>to resolve its AFTR. Then a query is sent to the DNS system (it can be
>dedicated to the service), and based on the load considerations, the
>requesting
>
>client is redirected to the appropriate AFTR nodes (i.e., an IP address is
>returned).
>
>With the sole IP address option we can not have such engineering flexibility
>for the provisioning of the AFTR.
>
>Having the two options allow for more flexibility for the engineering. IMHO,
>the IETF should not impose engineering choice to SPs. It will be up to each
>service provider to decide whether an IP address or a FQDN is more convenient
>in his deployment context.
I think this point should have been emphasized in the draft before sending it
to
IESG.
That would have made some difference in IESG.
Regards,
Behcet
_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires
