Updates to the IPv6 Multicast
Addressing ArchitectureFrance TelecomRennes35000Francemohamed.boucadair@orange.comCiscoUSAstig@cisco.com6man Working GroupThis document updates the IPv6 multicast addressing architecture by
defining the 17-20 reserved bits as generic flag bits. The document
provides also some clarifications related to the use of these flag
bits.This document updates RFC 3956, RFC 3306 and RFC 4291.The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119.This document updates the IPv6 multicast addressing architecture
by defining the 17-20 reserved bits as
generic flag bits (). The document provides
also some clarifications related to the use of these flag bits ().This document updates , , and .Bits 17-20 of a multicast address are defined in and as reserved
bits. This document defines these bits as generic flag bits so that they
apply to any multicast address. and
show the updated structure of the
addressing architecture. The first diagram shows the update of the base
IPv6 addressing architecture, and the second shows the update of
so-called Embedded-RP.Further specification documents may define a meaning for these flag
bits. Defining the bits 17-20 as flags for all IPv6 multicast addresses
allows addresses to be treated in a more uniform and generic way, and
allows for these bits to be defined in the future for different
purposes, irrespective of the specific type of multicast address.Some implementations and specification documents do not treat the
flag bits as separate bits but tend to use their combined value as a
4-bit integer. This practice is a hurdle for assigning a meaning to
the remaining flag bits. Below are listed some examples for
illustration purposes:the reading of may lead to
conclude that ff3x::/32 is the only allowed SSM IPv6 prefix
block. states only ff70::/12 applies to
Embedded-RP. Particularly, implementations should not treat the
fff0::/12 range as Embedded-RP.To avoid such confusion and to unambiguously associate a meaning
with the remaining flags, the following requirement is madeImplementations MUST treat flag bits as separate bits.This document changes Section 4 of
as follows:OLD:NEW:This document changes Section 6 of
as follows:OLD:These settings create an SSM range of FF3x::/32 (where 'x' is
any valid scope value). The source address field in the IPv6
header identifies the owner of the multicast address.NEW:If the flag bits are set to 0011, these settings create an SSM
range of ff3x::/32 (where 'x' is any valid scope value). The
source address field in the IPv6 header identifies the owner of
the multicast address. ff3x::/32 is not the only allowed SSM
prefix range. For example if the most significant flag bit is
set, then we would get the SSM range ffbx::/32.This document changes Section 2 of
as follows:OLD:NEW:This document changes Section 3 of
as follows:OLD:NEW:This document changes Section 4 of
as follows:OLD:It MUST be a multicast address with "flgs" set to 0111, that
is, to be of the prefix FF70::/12,NEW:It MUST be a multicast address with R-bit set to 1.It MUST have P-bit and T-bit both set to 1 when using the
embedding in this
document as it is a prefix-based address. This document changes Section 7.1 of
as follows:OLD:To avoid loops and inconsistencies, for addresses in the range
FF70::/12, the Embedded-RP mapping MUST be considered the longest
possible match and higher priority than any other mechanism.NEW:To avoid loops and inconsistencies, for addresses with R-bit
set to 1, the Embedded-RP mapping MUST be considered the longest
possible match and higher priority than any other mechanism. This document may require IANA updates. However, at this point it is
not clear exactly what these updates may be.Security considerations discussed in ,
and MUST
be taken into account.Many thanks to B. Haberman for the discussions prior to the
publication of this document.