Internet DRAFT - draft-vogt-dna-relocation
draft-vogt-dna-relocation
Network Working Group C. Vogt
Internet-Draft R. Bless
Expires: January 19, 2006 M. Doll
Universitaet Karlsruhe (TH)
G. Daley
Monash University
July 18, 2005
Analysis of IPv6 Relocation Delays
draft-vogt-dna-relocation-01.txt
Status of this Memo
By submitting this Internet-Draft, each author represents that any
applicable patent or other IPR claims of which he or she is aware
have been or will be disclosed, and any of which he or she becomes
aware will be disclosed, in accordance with Section 6 of BCP 79.
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.
The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.
This Internet-Draft will expire on January 19, 2006.
Copyright Notice
Copyright (C) The Internet Society (2005).
Abstract
As mobile nodes move to new links, they need to resume their
communications in a timely fashion. To communicate, IPv6 nodes need
to complete router discovery, address auto-configuration, and other
tasks. Unfortunately, this depends on prior completion of the
Multicast Listener Discovery (MLD) protocol, introducing a mandatory
Vogt, et al. Expires January 19, 2006 [Page 1]
Internet-Draft IPv6 Relocation Delays July 2005
delay of up to one second. This document discusses where this can be
problematic and proposes solutions that can alleviate the problem.
Vogt, et al. Expires January 19, 2006 [Page 2]
Internet-Draft IPv6 Relocation Delays July 2005
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4
2. History . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3. Problem Statement . . . . . . . . . . . . . . . . . . . . . . 4
3.1 Situation A . . . . . . . . . . . . . . . . . . . . . . . 5
3.2 Situation B . . . . . . . . . . . . . . . . . . . . . . . 6
3.3 Situation C . . . . . . . . . . . . . . . . . . . . . . . 6
3.4 Situation D . . . . . . . . . . . . . . . . . . . . . . . 6
3.5 Situation E . . . . . . . . . . . . . . . . . . . . . . . 7
4. Possible Solutions . . . . . . . . . . . . . . . . . . . . . . 7
4.1 Eliminating the Delay for MLD Reports . . . . . . . . . . 7
4.2 Using Optimistic Addresses Before the Initial NS . . . . . 8
4.3 Increasing Robustness . . . . . . . . . . . . . . . . . . 8
5. Brief Analysis . . . . . . . . . . . . . . . . . . . . . . . . 9
6. Security Considerations . . . . . . . . . . . . . . . . . . . 10
7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . 12
A. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 13
Intellectual Property and Copyright Statements . . . . . . . . 14
Vogt, et al. Expires January 19, 2006 [Page 3]
Internet-Draft IPv6 Relocation Delays July 2005
1. Introduction
Mobile nodes require fast router discovery and auto-configuration of
global IPv6 addresses. New mechanisms defined in [1], [3], and [4]
attend to this.
IPv6 Neighbor Discovery [1] accelerates router discovery in that it
allows mobile nodes to forego the recommended backoff for the first
Router Solicitation (RS) after interface (re-) initialization.
Optimistic Duplicate Address Detection (ODAD) [3] allows early use of
new IPv6 addresses once an initial Neighbor Solicitation (NS) has
been sent. Nodes who support the Tentative Source Link-Layer Address
Option (TSLLAO) [4] can save additional address-resolution round
trips during router discovery and address auto-configuration.
Still, mobile nodes must send a MLD Report [5][6] at some stage
during router discovery or address auto-configuration. MLD Reports
are subject to a random backoff between 0 and
MAX_RTR_SOLICITATION_DELAY (1 second) time RFC 2462bis [2]. There is
currently no acceleration for this, defeating the benefits of the
afore-mentioned optimizations in many situations.
This document identifies such situations and suggests two approaches
to improve them: relaxing the delay for MLD Reports under certain
conditions or permitting use of optimistic addresses prior to the
initial NS. The document is considered a basis for further mailing
list discussion. It is supposed to aid the revision process of
existing DNA Working Group documents rather than to evolve itself
towards a separate document.
2. History
This problem was first raised and discussed on the IPv6 mailing list
[9]. The outcome was to not incorporate a delay relaxation for the
MLD Report into RFC 2461bis and RFC 2462bis, as it was unclear then
whether this could negatively impact other on-link nodes (both mobile
and stationary). Instead, the consensus was to discuss this in
greater detail in the DNA Working Group.
The problem appeared again in a similar context on the DNA mailing
list [10].
3. Problem Statement
This section identifies five situations, A through E, in which IPv6
relocation is substantially delayed in spite of optimizations defined
by RFC 2461bis, ODAD, and TSLLAO.
Vogt, et al. Expires January 19, 2006 [Page 4]
Internet-Draft IPv6 Relocation Delays July 2005
In all situations, after changing links, the mobile node avoids using
its configured global unicast addresses during router discovery,
since it does not know before reception of a Router Advertisement
(RA) whether or not it has changed IPv6 attachment. Also, in order
to avoid Neighbor Cache pollution of on-link neighbors, the mobile
node must handle its configured link-local unicast addresses as if
those where tentative.
It is also assumed that the mobile node implements ODAD.
Nonetheless, the identified issues are essentially the same when the
mobile node uses non-optimized RFC 2462bis.
3.1 Situation A
Routers on the new IP link support the TSLLAO for RSs. The mobile
node solicits a unicast RA by sending an RS with an TSLLAO from the
unspecified address.
The TSLLAO allows the router to unicast the RA without performing
address resolution. What follows is the message exchange from the
mobile node's perspective:
1. MN sends an RS from the unspecified address with an TSLLAO.
2. MN receives an IPv6-multicast RA by link-layer unicast.
3. MN sends an MLD Report for an optimistic address.
4. MN sends an NS for the optimistic address, initiating ODAD.
The RS (step 1) may be sent immediately, even though this is the
first message transmitted after (re-) enabling the interface [1].
From a standard's perspective, it is debatable whether or not the MLD
Report must be delayed. RFC 2462bis says in section 5.4.2:
Even if the Neighbor Solicitation is not going to be the first
message to be sent, the node SHOULD delay joining the solicited-
node multicast address by a random delay between 0 and
MAX_RTR_SOLICITATION_DELAY if the address being checked is
configured by a router advertisement message sent to a multicast
address.
The RA (step 2) is an IPv6 multicast, yet a link-layer unicast.
Since there is only a single recipient in this case, omitting the
delay for the MLD Report would be feasible. But the mobile node must
inspect the link-layer destination address in order to determine
whether the RA was a multicast or a unicast. This may not always be
possible.
Vogt, et al. Expires January 19, 2006 [Page 5]
Internet-Draft IPv6 Relocation Delays July 2005
3.2 Situation B
Routers on the new IP link support the TSLLAO for RSs. The mobile
node solicits a unicast RA by sending an RS with an TSLLAO from an
optimistic address. Overall, it sends and receives the following
messages in sequence:
1. MN sends an MLD Report for an optimistic address.
2. MN sends an NS for the optimistic address, initiating ODAD.
3. MN sends an RS with an TSLLAO from the optimistic address.
4. MN receives a unicast RA.
The MLD Report (step 1) is delayed for up to 1 second
(MAX_RTR_SOLICITATION_DELAY) because it is the first message
transmitted after (re-) enabling the interface [2].
3.3 Situation C
Routers on the new IP link do not support the TSLLAO for RSs. The
mobile node solicits an RA. Routers unicast solicited RAs.
Soliciting a unicast RA requires the mobile node to send an RS from
an optimistic address (without a SLLAO). This is the complete
message exchange:
1. MN sends an MLD Report for an optimistic address.
2. MN sends an NS for the optimistic address, initiating ODAD.
3. MN sends an RS from the optimistic address.
4. MN receives an NS for the optimistic address from the router.
5. MN sends a Neighbor Advertisement (NA) for the optimistic address
to the router.
6. MN receives a unicast RA.
The MLD Report (step 1) is delayed for up to 1 second
(MAX_RTR_SOLICITATION_DELAY) because it is the first message
transmitted after (re-) enabling the interface [2].
3.4 Situation D
Routers on the new IP link do not support the TSLLAO for RSs, but do
send unsolicited multicast RAs every 30ms to 70ms according to rules
in [7]. The mobile node sends and receives the following messages in
sequence:
1. MN receives a multicast RA.
2. MN sends an MLD Report for an optimistic address.
Vogt, et al. Expires January 19, 2006 [Page 6]
Internet-Draft IPv6 Relocation Delays July 2005
3. MN sends an initial NS for the optimistic address (ODAD).
The MLD Report (step 2) is delayed for two reasons. First, it is the
first message transmitted after (re-) enabling the interface. This
calls for a delay of up to 1 second (MAX_RTR_SOLICITATION_DELAY) [2].
Second, the MLD Report is sent in response to a multicast RA. This
also calls for a delay of up to 1 seconds
(MAX_RTR_SOLICITATION_DELAY) [2].
3.5 Situation E
Routers on the new IP link do not support the TSLLAO for RSs. The
mobile node solicits a multicast RA.
The mobile node can send the RS from the unspecified address,
eliminating the requirement to initiate ODAD at this point. Overall,
it sends an receives the following messages:
1. MN sends an RS from the unspecified address.
2. MN receives a multicast RA.
3. MN sends an MLD Report for an optimistic address.
4. MN sends an NS for the optimistic address, initiating ODAD.
The RS (step 1) may be sent immediately, even though this is the
first message transmitted after (re-) enabling the interface [1].
The multicast RA (step 2) is delayed for up to 3 seconds (maximum of
MAX_RA_DELAY_TIME and MIN_DELAY_BETWEEN_RAS) [1].
In addition, the MLD Report (step 3) is delayed for up to 1 second
(MAX_RTR_SOLICITATION_DELAY) because it is sent in response to a
multicast RA [2].
4. Possible Solutions
The following approaches can be used to eliminate the high IPv6
relocation delays identified in Section 3.
4.1 Eliminating the Delay for MLD Reports
The random backoff for MLD Reports is "RECOMMENDED" in RFC 2461bis
and RFC 2462bis to resolve contention amongst multiple neighbors when
booting up simultaneously. It has little benefit when applied by a
mobile node after IPv6 relocation.
It hence seems feasible to eliminate the backoff for MLD Reports
after IPv6 relocation. In particular, both of the following two
conditions would have to be met:
Vogt, et al. Expires January 19, 2006 [Page 7]
Internet-Draft IPv6 Relocation Delays July 2005
o The mobile node has received a trigger from its local link layer
indicating that an interface, which was previously operational,
has gone down and come up again.
o The MLD Report is either the first message transmitted on the new
link or it is sent in response to a unicast RA indicating IPv6
relocation.
4.2 Using Optimistic Addresses Before the Initial NS
ODAD allows (limited) use of optimistic addresses after an initial NS
has been sent. This NS must be preceeded by a MLD Report for the
corresponding solicited-nodes multicast address. The random backoff
for the MLD Report foils the benefits of ODAD.
Fortunately, it appears feasible to allow (limited) use of an
optimistic address even before the initial NS is sent. The delay for
the MLD Report can so be moved to a non-critical phase.
Note that mobile nodes would still have to send a MLD Report prior to
sending a RS from an optimistic address in situation C: Since the RS
does not include a SLLAO, the router will have to invoke address
resolution for the optimistic address. Sending the MLD Report
ensures that the mobile node can receive the NS in the presence of
snooping switches. Note that RFC 2461bis currently does not require
transmission of a MLD Report prior to the RS in case of interface
(re-) initialization.
4.3 Increasing Robustness
The optimum desynchonization delays for signaling messages such as
the MLD Report are highly application- and environment-specific. RFC
2461bis and RFC 2462bis are tailored towards stationary nodes, so the
delays of up to MAX_RTR_SOLICITATION_DELAY (1 second) are acceptable
and make sense. However, for mobile nodes, such delays will badly
impact real-time applications like VoIP. It may be possible in rare
deployment scenarios to hard-code application- and environment-
specific behavior into the nodes. But this approach is infeasible in
general, simply because the applications and environments are unknown
a priori.
One solution to this problem is to differentiate between messages for
which loss is recoverable and messages for which it is not. E.g.,
loss of an RS can be detected by the mobile node by lack of a
corresponding RA. In this (unlikely) case, the mobile node loses
some time, but will eventually retransmit the RS. On the other hand,
loss of a MLD Report or NS sent during ODAD cannot be detected,
because there is no expected response. This might lead to an address
duplicate, which is currently not recoverable.
Vogt, et al. Expires January 19, 2006 [Page 8]
Internet-Draft IPv6 Relocation Delays July 2005
In this sense, a mobile node should retransmit an MLD Report and RS
after an appropriate time period if it fails to receive a
corresponding RA. For the purpose of ODAD, the mobile node should
retransmit MLD Reports and NSs in background mode two, three, or more
times. An optimistic address would then be considered unique only
after none of these transmissions resulted in a response. The mobile
node may use plenty of desynchronization delay without negatively
affecting applications because all transmission occur in background
mode.
Specifically, mobile nodes SHOULD retransmit multiple NSs during
ODAD. Each NS SHOULD be preceded by a retransmitted MLD Report.
Thus, robustness to loss of (undelayed) MLD Report can be increased.
5. Brief Analysis
The conditions defined in Section 4.1 should be conservative enough
to limit the proposed optimization to non-critical situations. In
particular, the random backoff for MLD Reports should be retained for
responses to multicast RAs as well as when the node boots up. This
assumes that a reliable indication for these conditions can be
implemented.
Generally, collisions of MLD and IPv6 Neighbor Discovery messages are
more likely to occur in a mobile environment than in a stationary
because mobile nodes use these messages for movement detection and
IPv6 relocation. Delaying the MLD Report does not mitigate this,
however.
Retransmitted MLD Reports and RSs, as described in Section 4.3, may
repair router-discovery failure in situation C due to a collided
(undelayed) MLD Report. Similarly, retransmitting MLD Reports and
NSs, during ODAD, is likely to resolve collisions and ensure correct
operation of ODAD in situations A through E.
ODAD may still fail when neighbors located at different sides of a
snooping switch simultaneously attempt to auto-configure the same
IPv6 address. If one of the nodes' MLD Report collides, but its NS
does not, that node will not receive the competing node's NS. The
result is that the former "wins" this race condition and the latter
"loses" it.
More serious is a situation in which both MLD Reports get lost. The
MLD state will eventually be repaired by the retransmitted MLD Report
of both the mobile node and the neighbor. But there will be two
nodes on the link using the same address. The IPv6 protocol suite
currently does not recover from configured duplicate addresses.
Vogt, et al. Expires January 19, 2006 [Page 9]
Internet-Draft IPv6 Relocation Delays July 2005
Transmission of MLD Reports can lead to undesired signaling overhead.
Furthermore, with respect to router discovery and address auto-
configuration, MLD Reports have a benefit only in the presence of
snooping switches [8]. The crux is that a mobile node usually does
not know a new link's topology at attachment time, so it seems that
omission of MLD Reports would become feasible only with appropriate
support from the link layer.
6. Security Considerations
The optimizations proposed in this document affect desynchronization
delays and retransmission policies applied by mobile nodes.
Malicious nodes can always choose their own delays and policies, of
course. As a consequence, the optimizations proposed in this
document do not imply security concerns in addition to those which
already exist in the standard IPv6 protocol suite [1][2], in ODAD
[3], or in TSLLAO [4].
Vogt, et al. Expires January 19, 2006 [Page 10]
Internet-Draft IPv6 Relocation Delays July 2005
7. References
[1] Narten, T., "Neighbor Discovery for IP version 6 (IPv6)",
draft-ietf-ipv6-2461bis-03 (work in progress), May 2005.
[2] Thomson, S., "IPv6 Stateless Address Autoconfiguration",
draft-ietf-ipv6-rfc2462bis-08 (work in progress), May 2005.
[3] Moore, N., "Optimistic Duplicate Address Detection for IPv6",
draft-ietf-ipv6-optimistic-dad-05 (work in progress),
February 2005.
[4] Daley, G., "Tentative Source Link-Layer Address Options for
IPv6 Neighbour Discovery", draft-daley-ipv6-tsllao-01 (work in
progress), February 2005.
[5] Deering, S., Fenner, W., and B. Haberman, "Multicast Listener
Discovery (MLD) for IPv6", RFC 2710, October 1999.
[6] Rolland, R., Luis, L., , S., Deering, S., Fenner, W., Kouvelas,
I., and B. Haberman, "Multicast Listener Discovery Version 2
(MLDv2) for IPv6", RFC 3810, June 2004.
[7] Johnson, D., Perkins, C., and J. Arkko, "Mobility Support in
IPv6", RFC 3775, June 2004.
[8] Christensen, M., Kimball, K., and F. Solensky, "Considerations
for IGMP and MLD Snooping Switches", draft-ietf-magma-snoop-12
(work in progress), February 2005.
[9] "Comment on RFCs 2461bis and 2462bis", Discussion on the IPv6
mailing list, May 27, 2005, <http://www1.ietf.org/mail-archive/
web/ipv6/current/msg05084.html>.
[10] "DAD and MLD Interaction", Discussion on the DNA mailing
list, June 13, 2005, <http://webcamserver.eng.monash.edu.au/
~warchive/dna/2005-06/msg00098.html>.
Vogt, et al. Expires January 19, 2006 [Page 11]
Internet-Draft IPv6 Relocation Delays July 2005
Authors' Addresses
Christian Vogt
Institute of Telematics
Universitaet Karlsruhe (TH)
P.O. Box 6980
76128 Karlsruhe
Germany
Email: chvogt@tm.uka.de
URI: http://www.tm.uka.de/~chvogt/
Roland Bless
Institute of Telematics
Universitaet Karlsruhe (TH)
P.O. Box 6980
76128 Karlsruhe
Germany
Email: bless@tm.uka.de
URI: http://www.tm.uka.de/~bless/
Mark Doll
Institute of Telematics
Universitaet Karlsruhe (TH)
P.O. Box 6980
76128 Karlsruhe
Germany
Email: doll@tm.uka.de
URI: http://www.tm.uka.de/~doll/
Gregory Daley
Centre for Telecommunications and Information Engineering
Department of Electrical and Computer Systems Engineering
Monash University
Clayton, Victoria 3800
Australia
Email: reg.daley@eng.monash.edu.au
Vogt, et al. Expires January 19, 2006 [Page 12]
Internet-Draft IPv6 Relocation Delays July 2005
Appendix A. Acknowledgments
Credits go to Gregory Daley for a valuable discussion on interactions
between MLD and router discovery plus address auto-configuration.
Thanks also to Jari Arkko, who thoroughly reviewed version 00 of this
draft. Last, but not least, the authors appreciate the contributions
of folks to the initial thread on the IPv6 mailing list [9].
Vogt, et al. Expires January 19, 2006 [Page 13]
Internet-Draft IPv6 Relocation Delays July 2005
Intellectual Property Statement
The IETF takes no position regarding the validity or scope of any
Intellectual Property Rights 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; nor does it represent that it has
made any independent effort to identify any such rights. Information
on the procedures with respect to rights in RFC documents can be
found in BCP 78 and BCP 79.
Copies of IPR disclosures made to the IETF Secretariat 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 implementers or users of this
specification can be obtained from the IETF on-line IPR repository at
http://www.ietf.org/ipr.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights that may cover technology that may be required to implement
this standard. Please address the information to the IETF at
ietf-ipr@ietf.org.
Disclaimer of Validity
This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
ENGINEERING TASK FORCE DISCLAIM 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.
Copyright Statement
Copyright (C) The Internet Society (2005). This document is subject
to the rights, licenses and restrictions contained in BCP 78, and
except as set forth therein, the authors retain all their rights.
Acknowledgment
Funding for the RFC Editor function is currently provided by the
Internet Society.
Vogt, et al. Expires January 19, 2006 [Page 14]