DHCPv6 only requires link-local multicast not site or global. From draft-ietf-dhc-dhcpv6-23:
7.1. Multicast Addresses
DHCP makes use of the following multicast addresses:
All_DHCP_Agents (FF02::1:2) This link-scoped multicast address
is used by clients to communicate with the
on-link agent(s) when they do not know the
link-local address(es) for those agents. All
agents (servers and relays) are members of this
multicast group.
All_DHCP_Servers (FF05::1:3) This site-scoped multicast address is
used by clients or relays to communicate with
server(s), either because they want to send
messages to all servers or because they do not
know the server(s) unicast address(es). Note
that in order for a client to use this address,
it must have an address of sufficient scope to
be reachable by the server(s). All servers
within the site are members of this multicast
group.
The All_DHCP_Servers is not a requirement to make the protocol work. It is usable in two cases:
- For Information-Request messages to avoid the need for relays. But, the All_DHCP_Agents is usable with relays.
- For auto-configuration of relays - if a relay doesn't have a configured DHCP server to forward packets to, it can forward them to the ALL_DHCP_Servers address.
So, DHCPv6 MUST have link-local multicast.
- Bernie Volz
-----Original Message-----
From: Bob Hinden [mailto:[EMAIL PROTECTED]]
Sent: Monday, March 18, 2002 12:36 PM
To: Thomas Narten
Cc: [EMAIL PROTECTED]
Subject: Re: Deployability/Useability of Multicast [was: Re: PPP and
Global Addresses ]
Thomas,
>1) Do people think that the existance of multicast within a site
> (e.g., home, enterprise, etc.) can be assumed?
>
> Note: I'm assuming link-local multicast will always work, the
> question is whether wider scope (beyond neighboring routers)
> multicast is available.
As much as I would like to see it otherwise, except for link (and maybe
subnet) scoped multicast, I don't think we can assume multicast across a
wider scope. I suspect it will be more the exception than the norm.
>2) If the answer is "can't assume multicast is present", what should
> be done about it. e.g.,
>
> - What are the barriers that are preventing multicast from being an
> assumed part of the infrastructure? What is needed to remove
> those barriers?
Don't we have a whole working group dealing with this (e.g., mboned)?
It's a complicated mixture of routing protocol scaling, lack of agreement
of inter-domain routing protocols, lack of interoperability between vendors
of multicast routing protocols, few compelling applications that need
multicast that can't be meet with unicast, etc. There are a few places
where it is used (i.e., financial industry), but as far as I can tell it's
not widely adopted much beyond that.
> - Which infrastructure protocols does IPv6 rely on that assume the
> existance of multicast, and how are they limited if multicast is
> not available?
Off the top of my head, DHCPv6 and Router Renumbering.
Bob
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
