RE: [dhcwg] Reserved anycast address for DHCPv6

Ralph:

I can't say for every kernel/stack, but for Unix (Solaris) there is the IOCTL call to read a flag that says whether the interface is broadcast/multicast capable. I think it is in the SIOCGIFFLAGS (Get interface flags) - one of the flag bits is true if broadcast capable.

Also, on Solaris you generally see (via IFCONFIG) something like:
e0:1: flags=2080841<UP,RUNNING,MULTICAST,ADDRCONF,IPv6>
          mtu 1500 index 2
        inet6 fec0::55:a00:20ff:fe8e:f3ad/64 
le0:2: flags=2080841<UP,RUNNING,MULTICAST,ADDRCONF,IPv6>
          mtu 1500 index 2
        inet6 3ff0::55:a00:20ff:fe8e:f3ad/64

The MULTICAST flag is the one of interest.

- Bernie

-----Original Message-----
From: Ralph Droms [mailto:rdroms@cisco.com]
Sent: Friday, April 05, 2002 5:09 PM
To: Bernie Volz (EUD)
Cc: dhcwg@ietf.org
Subject: RE: [dhcwg] Reserved anycast address for DHCPv6

Bernie - Can clients reliably differentiate between broadcast and NMBA
links?

- Ralph

On Fri, 5 Apr 2002, Bernie Volz (EUD) wrote:

> Ralph:
>
> >One potential problem is that clients now have two addresses.
>
> Is this really such a problem? If the client's interface is on a NBMA network, it uses the anycast address. If it is on a multicast capable network, it uses the Multicast address. So, yes, clients need to know about the two addresses but they only need to send packets to one.
>
> >Does anyone on the list know anything about the proposal for reserving a
> >site-scoped SLA for anycast addresses?
>
> Have no idea. But, if we submit the request to IANA won't they do the address assignment and follow the rules?
>
> - Bernie
>
> -----Original Message-----
> From: Ralph Droms [mailto:rdroms@cisco.com]
> Sent: Friday, April 05, 2002 2:40 PM
> To: dhcwg@ietf.org
> Subject: [dhcwg] Reserved anycast address for DHCPv6
>
>
> <draft-ietf-ipv6-dns-discovery-04.txt> includes the definition of three
> anycast addresses for use by DNS resolvers:
>
>     fec0:0:0:ffff::1
>     fec0:0:0:ffff::2
>     fec0:0:0:ffff::3
>
> I had a request in Minneapolis that we consider defining an anycast address
> for DHCPv6 that would allow the use of DHCPv6 on NBMA networks.  Seems to
> me like a reasonable suggestion.
>
> One potential problem is that clients now have two addresses: the
> All_DHCP_Agents multicast address and the new anycast address.  Which
> should the client choose to use?
>
> BTW, the DNS Discovery draft includes the following:
>
>    Note to readers: the above addresses are tentative, but the ffff
>    is intended to be consistent with a simultaneous proposal to
>    reserve the ffff SLA for use with IANA-assigned addresses such as
>    these.
>
> Does anyone on the list know anything about the proposal for reserving a
> site-scoped SLA for anycast addresses?
>
> - Ralph
>
>
> _______________________________________________
> dhcwg mailing list
> dhcwg@ietf.org
> https://www1.ietf.org/mailman/listinfo/dhcwg
>