Internet Engineering Task Force P. Savola
Internet Draft CSC/FUNET
Expiration Date: August 2004
February 2004
IPv6 Multicast Deployment Issues
draft-savola-v6ops-multicast-issues-03.txt
Status of this Memo
This document is an Internet-Draft and is subject to all provisions
of Section 10 of RFC2026.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF), its areas, and its working groups. Note that
other groups may also distribute working documents as Internet-
Drafts.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt.
To view the list Internet-Draft Shadow Directories, see
http://www.ietf.org/shadow.html.
Abstract
There are many issues concerning the deployment and implementation,
and to a lesser degree, specification of IPv6 multicast. This memo
describes known problems, trying to raise awareness. Currently,
global IPv6 interdomain multicast is impossible except using SSM:
there is no way to convey information about multicast sources between
PIM-SM RPs; the situation is analyzed, and a technique called
Embedded RP is offered as a solution. The deploability of SSM and
Embedded RP is also considered. In addition, an issue regarding
link-local multicast-blocking Ethernet switches is brought up.
Finally, the requirement for functionality like MLD snooping is
noted.
Savola [Expires August 2004] [Page 1]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
Table of Contents
1. Introduction ............................................... 2
2. Issues with Multiple PIM Domains and Any Source Multicast .. 3
2.1. Changing the Multicast Usage Model ..................... 3
2.2. Implementing MSDP for IPv6 Interdomain Multicast ....... 4
2.3. Implementing Another Multicast Routing Protocol ........ 4
2.4. Embedding the RP Address in an IPv6 Multicast Address .. 4
2.5. Site-local Group Scoping ............................... 5
3. Issues with SSM and Embedded RP ............................ 5
3.1. SSM Deployment ......................................... 5
3.2. RP Failover with Embedded RP .......................... 6
4. Neighbor Discovery Using Multicast ......................... 6
5. Functionality Like MLD Snooping ............................ 7
6. Security Considerations .................................... 7
7. Acknowledgements ........................................... 8
8. References ................................................. 8
8.1. Normative References ................................... 8
8.2. Informative References ................................. 8
Author's Address ............................................... 9
Intellectual Property Statement ................................ 9
Full Copyright Statement ....................................... 10
1. Introduction
There are many issues concerning the deployment and implementation,
and to a lesser degree, specification of IPv6 multicast. This memo
describes known problems, trying to raise awareness.
Currently, global IPv6 interdomain multicast is impossible except
using SSM: there is no way to convey information about multicast
sources between PIM-SM RPs; site-scoped multicast problematic. A few
possible solutions, such as Embedded RP [EMBEDRP] are outlined or
referred to. These are discussed in section 2. Deployment issues
with both SSM and Embedded RP are discussed separately in section 3.
In addition, an issue regarding link-local multicast -blocking
Ethernet switches is brought up. Finally, the requirement for
functionality like MLD snooping is noted. These are discussed in
sections 4 and 5, respectively.
[MULTIGAP] analyses the more generic set of issues with multicast;
this memo focuses on critical issues regarding IPv6.
Savola [Expires August 2004] [Page 2]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
2. Issues with Multiple PIM Domains and Any Source Multicast
For both administrative and technical reasons, there must be multiple
Protocol-Independent Multicast - Sparse Mode (PIM-SM) [PIM-SM]
domains in the Internet, which means there will be multiple PIM-SM
Rendezvous Points (RPs) -- and communication mechanisms between these
RPs will become critical.
These issues only come up with classical Any Source Multicast;
Source-Specific Multicast [SSM] does not require RPs and is not
affected, as the multicast "channel" is identified by the combination
<source address, group address> and can be communicated out-of-band.
In IPv4, notification of multicast sources between these PIM RPs is
done with Multicast Source Discovery Protocol (MSDP) [MSDP]. Many
consider this a sub-optimal, but unfortunately necessary, solution;
when it was specified, it was only meant as a "stop-gap" measure.
Below, some issues and solutions or work-arounds are described.
2.1. Changing the Multicast Usage Model
As "Any Source Multicast" -model has been found to be complex and a
bit problematic, there may be an incentive to move to SSM for good
(at least for most cases). Below two paragraphs are adapted from
[PIMSO]:
The most serious criticism of the SSM architecture is that it does
not support shared trees which may be useful for supporting many-to-
many applications. In the short-term this is not a serious concern
since the multicast application space is likely to be dominated by
one-to-many applications. Some other classes of multicast
applications that are likely to emerge in the future are few-to-few
(e.g. private chat rooms, whiteboards), few-to-many (e.g., video
conferencing, distance learning) and many-to-many (e.g., large chat
rooms, multi-user games). The first two classes can be easily handled
using a few one-to-many source-based trees.
The issue of many-to-many multicasting service on top of a SSM
architecture is an open issue at this point. However, some feel that
even many-to-many applications should be handled with multiple one-
to-many instead of shared trees.
In any case, even though SSM would avoid mentioned problems, it is
far from being generally implemented, much less deployed, yet.
Savola [Expires August 2004] [Page 3]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
2.2. Implementing MSDP for IPv6 Interdomain Multicast
One could argue that currently, the easiest stop-gap solution (to a
stop-gap solution) would be to specify IPv6 TLV's for MSDP. This
would be fairly straightforward, and existing implementations would
probably be relatively easily modifiable.
There has been some resistance to this, as MSDP was not supposed to
last this long in the first place.
2.3. Implementing Another Multicast Routing Protocol
One possibility might be to specify and/or implement a different
multicast routing protocol. In fact, Border Gateway Multicast
Protocol (BGMP) [BGMP] has been specified for a few years; however,
it seems quite complex and there have been no implementations.
Lacking deployment experience and specification analysis, it is
difficult to say which problems it might solve (and possibly, which
new ones to introduce).
In conclusion, looking for a solution in BGMP may not be realistic in
this time frame.
2.4. Embedding the RP Address in an IPv6 Multicast Address
One way to work around these problems would be to allocate and assign
multicast addresses in such a fashion that the address of the RP
could be automatically calculated from the multicast address.
At the first glance, this appears to be an impossible problem: the
address of the RP, as well as the multicast address, are both 128
bits long; in the general case, embedding one in the other is
impossible.
However, making some assumptions about multicast addressing, this is
can be done -- a proposed solution is presented in a different memo
[EMBEDRP]. Additionally, this requires a PIM-SM implementation of
the Embedded RP group-to-RP mapping mechanism which takes this
encoding to the account.
One should note that MSDP is also used in "Anycast RP" [ANYCASTRP]
-technique, for sharing the state information between different RP's
in one PIM domain; unless other proposals, such as [ANYPIMRP], are
deployed, or MSDP for IPv6 implemented, Anycast-RP technique cannot
be used.
However, a "cold failover" variant of anycast-RP (for long-term
redundancy only) would still be possible. In this mechanism, multiple
Savola [Expires August 2004] [Page 4]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
routers would be configured with the RP address, but only one would
be active at the time: if the RP goes down, another takes its place.
The multicast state stored in the RP would be lost, though, unless it
is copied by some out-of-band mechanism (e.g. placing the backup RP
absolutely on-the-path and have it snoop all the relevant packets).
2.5. Site-local Group Scoping
Site-local groups must be their own PIM domains to prevent site-local
data leaking to other sites. A more complex possibility would be to
implement something resembling "BSR border" feature which would
filter out all site-local components in PIM packets: if this is not
done very carefully, site-local information will leak to the global
network. This is operationally difficult, and PIM working group has
come to consensus that a scope boundary MUST also be a a site
boundary for certain critical PIM messages (e.g. C-RP and Bootstrap).
Especially if site-local multicast is used (and the site also wants
to engage in global multicast), there will be a huge number of
domains and communication required between them. This will increase
the need for a global multicast solution.
3. Issues with SSM and Embedded RP
This section briefly describes some challenges with SSM deployment,
and gaps caused by the introduction of Embedded RP.
3.1. SSM Deployment
To be deployed, SSM requires changes to routers, MLD-snooping
Ethernet switches, host systems, APIs, and the multicast usage
models.
Introducing SSM support in the routers has been straightforward.
IGMP-snooping Ethernet switches have been a more difficult issue,
though [SSMSNOOP]; some which perform IGMPv2 snooping discard IGMPv3
reports or queries, or multicast transmissions associated to them.
If MLDv1 snooping had been implemented, this would likely have
affected that as well.
Host systems require MLDv2 support. The situation has become
slightly better with respect to MLDv2 support for end systems. The
multicast source filtering API specification has also been completed;
its deployment is likely roughly equal (or slightly worse) than
MLDv2. The API is required for creating SSM applications.
Savola [Expires August 2004] [Page 5]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
Now the most difficult problem, multicast usage models, remains a
problem. SSM is an excellent fit for one-to-many distribution
topologies, and porting such applications to use SSM would likely be
rather simple. However, a significant number of current applications
are many-to-many (e.g., conferencing applications) which cannot be
converted to SSM without significant effort, including, for example,
out-of-band source discovery. For such applications to be usable for
IPv6 at least in a short to medium term, Any Source Multicast -like
techniques seem to be required.
3.2. RP Failover with Embedded RP
Embedded RP provides a means for ASM multicast without inter-domain
MSDP. However, to continue providing failover mechanisms for RPs, a
form of state sharing, called Anycast-RP, should still be supported.
MSDP could still be used for that; an alternative is doing the same
in PIM itself [PIMANYRP].
However, one should note that as Embedded RP does not require MSDP
peerings between the RPs, it's possible to deploy more RPs in a PIM
domain. For example, the scalability and redundancy could be
achieved by co-locating RP functionality in the DRs: each major
source, which "owns" a group, could have its own DR act as the RP.
This has about the same redundancy characteristics as using SSM -- so
there may not be an actually very urgent need for Anycast-RP if
operational methods to include fate-sharing of the groups is
followed.
4. Neighbor Discovery Using Multicast
Neighbor Discovery [NDISC] uses link-local multicast in Ethernet
media, not broadcast as does ARP with IPv4. This may cause some
operational problems with some equipment.
The author has seen one brand of managed Ethernet switches, and heard
reports of a few unmanaged switches, that do not forward IPv6 link-
layer multicast packets to other ports at all. In essence, native
IPv6 is impossible with this equipment. Investigation is still going
on whether these issues can be worked around.
However, it seems likely this may be a problem with some switches
that build multicast forwarding state based on Layer 3 information
(and do not support IPv6); state using Layer 2 information would work
just fine [MLDSNOOP].
For the deployment of IPv6, it would be important to find out how
this can be fixed (e.g. how exactly this breaks specifications) and
how one can identify which equipment could case problems like these
Savola [Expires August 2004] [Page 6]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
(and whether there are workarounds).
One workaround might be to implement a toggle in the nodes that would
use link-layer broadcast instead of multicast as a fallback solution.
However, this would have to be used in all the systems in the same
segment one wishes to communicate with.
5. Functionality Like MLD Snooping
On Ethernet, multicast frames are forwarded to every port, even
without subscribers (or IPv6 support).
Especially if multicast traffic is relatively heavy (e.g. video
streaming), it becomes particularly important to have some feature
like Multicast Listener Discovery (MLD) snooping implemented in the
equipment, most importantly Ethernet switches [MLDSNOOP].
In addition, there have been some misunderstandings wrt. which
multicast addresses (in particular, link-locals) MLD reports
(utilized in the snooping) should be generated for. If all
implementations do not generate enough MLD reports, the introduction
of MLD snooping could cause them being blocked out. To clarify, a
MLD report MUST be generated for every group except all-nodes
(ff02::1 -- which is forwarded to all ports); this also includes all
the other link-local groups.
Looking at the actual problem from a higher view, it is not clear
that MLD snooping is the right long-term solution. It makes the
switches complex, requires the processing of datagrams above the
link-layer, and should be discouraged [MULTIGAP]: the whole idea of
L2-only devices having be able to peek into L3 datagrams seems like a
severe layering violation -- and often the devices aren't upgradeable
in any way. Better mechanisms could be having routers tell switches
which multicasts to forward where (e.g. [CGMP]) or by using some
other mechanisms [GARP].
6. Security Considerations
Only deployment/implementation issues are considered, and these do
not have any particular security considerations; security
considerations for each technology are covered in the respective
specifications.
One fairly obvious issue raised in this memo is that if there is no
adminsitrative PIM domain border between site-local multicast
domains, the site-local traffic could very easily flow into other
sites and the Internet.
Savola [Expires August 2004] [Page 7]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
7. Acknowledgements
Early discussions with Stig Venaas, Jerome Durand, Tim Chown et al.
led to the writing of this draft. Brian Haberman offered extensive
comments along the way. "Itojun" Hagino brought up the need for MLD
snooping in a presentation. Bill Nickless pointed out issues in the
gap analysis and provided a pointer to GARP/GMRP; HÂ…vard Eidnes made
a case for a protocol like CGMP. Leonard Giuliano pointed out a more
complete analysis of SSM with different kind of applications.
8. References
8.1. Normative References
[ANYCASTRP] Kim, D. et al, "Anycast RP mechanism using PIM and
MSDP", RFC 3446, January 2003.
[EMBEDRP] Savola, P., Haberman, B., "Embedding the RP Address
in an IPv6 Multicast Address", work-in-progress,
draft-ietf-mboned-embeddedrp-01.txt, Feb 2004.
[MSDP] Fenner, B., Meyer, D., "Multicast Source Discovery
Protocol", RFC 3618, Oct 2003.
[NDISC] Narten, T., Nordmark, E., Simpson W., "Neighbor Discovery
for IP Version 6 (IPv6)", RFC2461, December 1998.
[PIM-SM] Fenner, B. et al, "Protocol Independent Multicast -
Sparse Mode (PIM-SM): Protocol Specification (Revised)",
work-in-progress, draft-ietf-pim-sm-v2-new-08.txt,
October 2003.
[SSM] Holbrook, H. et al, "Source-Specific Multicast for IP",
work-in-progress, draft-ietf-ssm-arch-04.txt,
Oct 2003.
8.2. Informative References
[ANYPIMRP] Farinacci, D., Cai, Y., "Anycast-RP using PIM",
work-in-progress, draft-ietf-pim-anycast-rp-00.txt,
November 2003.
[BGMP] Thaler, D., "Border Gateway Multicast Protocol (BGMP)",
work-in-progress, draft-ietf-bgmp-spec-06.txt,
January 2004.
[BSR] Fenner, B., et al., "Bootstrap Router (BSR) Mechanism for
PIM Sparse Mode", work-in-progress, draft-ietf-pim-sm-
Savola [Expires August 2004] [Page 8]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
bsr-03.txt, February 2003.
[CGMP] Cisco, "Cisco Group Management Protocol", e.g.
http://www.cisco.com/en/US/tech/tk648/tk363/tk105/
tech_protocol_home.html
[GARP] Tobagi, F., et al, "Study of IEEE 802.1p GARP/GMRP Timer
Values", (for introduction to GARP/GMRP, see section 2),
Sep 1997.
[MLDSNOOP] Christensen, M., Solensky, F., "IGMP and MLD snooping
switches", work-in-progress, draft-ietf-magma-
snoop-10.txt, October 2003.
[MULTIGAP] Meyer, D., Nickless, B., "Internet Multicast Gap
Analysis [...]", work-in-progress,
draft-ietf-mboned-iesg-gap-analysis-00.txt, July 2002.
[PIMSO] Bhattacharyya, S. et al, "Deployment of PIM-SO at Sprint
(PIM-SO)", work-in-progress,
draft-bhattach-diot-pimso-00.txt (expired), March 2000.
[SSMSNOOP] Thaler, D., "Operational Problems with IGMP snooping
switches", presentation in MAGMA WG at IETF56,
http://www.ietf.org/proceedings/03mar/148.htm,
March 2003.
Author's Address
Pekka Savola
CSC/FUNET
Espoo, Finland
EMail: psavola@funet.fi
Intellectual Property Statement
The IETF takes no position regarding the validity or scope of any
intellectual property or other rights that might be claimed to
pertain to the implementation or use of the technology described in
this document or the extent to which any license under such rights
might or might not be available; neither does it represent that it
has made any effort to identify any such rights. Information on the
IETF's procedures with respect to rights in standards-track and
standards-related documentation can be found in BCP-11. Copies of
claims of rights made available for publication and any assurances of
licenses to be made available, or the result of an attempt made to
obtain a general license or permission for the use of such
proprietary rights by implementors or users of this specification can
Savola [Expires August 2004] [Page 9]
Internet Draft draft-savola-v6ops-multicast-issues-03.txt February 2004
be obtained from the IETF Secretariat.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights which may cover technology that may be required to practice
this standard. Please address the information to the IETF Executive
Director.
Full Copyright Statement
Copyright (C) The Internet Society (2004). All Rights Reserved.
This document and translations of it may be copied and furnished to
others, and derivative works that comment on or otherwise explain it
or assist in its implementation may be prepared, copied, published
and distributed, in whole or in part, without restriction of any
kind, provided that the above copyright notice and this paragraph are
included on all such copies and derivative works. However, this
document itself may not be modified in any way, such as by removing
the copyright notice or references to the Internet Society or other
Internet organizations, except as needed for the purpose of
developing Internet standards in which case the procedures for
copyrights defined in the Internet Standards process must be
followed, or as required to translate it into languages other than
English.
The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assignees.
This document and the information contained herein is provided on an
"AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.
Savola [Expires August 2004] [Page 10]