RFC 7527

Internet Engineering Task Force (IETF)                          R. Asati
Request for Comments: 7527                                      H. Singh
Updates: 4429, 4861, 4862                                      W. Beebee
Category: Standards Track                                   C. Pignataro
ISSN: 2070-1721                                      Cisco Systems, Inc.
                                                                 E. Dart
                                   Lawrence Berkeley National Laboratory
                                                               W. George
                                                       Time Warner Cable
                                                              April 2015

                  Enhanced Duplicate Address Detection


   IPv6 Loopback Suppression and Duplicate Address Detection (DAD) are
   discussed in Appendix A of RFC 4862.  That specification mentions a
   hardware-assisted mechanism to detect looped back DAD messages.  If
   hardware cannot suppress looped back DAD messages, a software
   solution is required.  Several service provider communities have
   expressed a need for automated detection of looped back Neighbor
   Discovery (ND) messages used by DAD.  This document includes
   mitigation techniques and outlines the Enhanced DAD algorithm to
   automate the detection of looped back IPv6 ND messages used by DAD.
   For network loopback tests, the Enhanced DAD algorithm allows IPv6 to
   self-heal after a loopback is placed and removed.  Further, for
   certain access networks, this document automates resolving a specific
   duplicate address conflict.  This document updates RFCs 4429, 4861,
   and 4862.

Status of This Memo

   This is an Internet Standards Track document.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Further information on
   Internet Standards is available in Section 2 of RFC 5741.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at

Asati, et al.                Standards Track                    [Page 1]

RFC 7527                      Enhanced DAD                    April 2015

Copyright Notice

   Copyright (c) 2015 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (http://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   2
     1.1.  Requirements Language . . . . . . . . . . . . . . . . . .   3
     1.2.  Terminology . . . . . . . . . . . . . . . . . . . . . . .   3
   2.  Problem Statement . . . . . . . . . . . . . . . . . . . . . .   4
   3.  Operational Mitigation Options  . . . . . . . . . . . . . . .   4
     3.1.  Disable DAD on an Interface . . . . . . . . . . . . . . .   4
     3.2.  Dynamic Disable/Enable of DAD Using Layer 2 Protocol  . .   5
     3.3.  Operational Considerations  . . . . . . . . . . . . . . .   5
   4.  The Enhanced DAD Algorithm  . . . . . . . . . . . . . . . . .   6
     4.1.  Processing Rules for Senders  . . . . . . . . . . . . . .   6
     4.2.  Processing Rules for Receivers  . . . . . . . . . . . . .   7
     4.3.  Changes to RFC 4861 . . . . . . . . . . . . . . . . . . .   7
   5.  Action to Perform on Detecting a Genuine Duplicate  . . . . .   7
   6.  Security Considerations . . . . . . . . . . . . . . . . . . .   8
   7.  Normative References  . . . . . . . . . . . . . . . . . . . .   8
   Acknowledgements  . . . . . . . . . . . . . . . . . . . . . . . .   9
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .  10

1.  Introduction

   IPv6 Loopback Suppression and Duplicate Address Detection (DAD) are
   discussed in Appendix A of [RFC4862].  That specification mentions a
   hardware-assisted mechanism to detect looped back DAD messages.  If
   hardware cannot suppress looped back DAD messages, a software
   solution is required.  One specific DAD message is the Neighbor
   Solicitation (NS), specified in [RFC4861].  The NS is issued by the
   network interface of an IPv6 node for DAD.  Another message involved
   in DAD is the Neighbor Advertisement (NA).  The Enhanced DAD
   algorithm specified in this document focuses on detecting an NS
   looped back to the transmitting interface during the DAD operation.
   Detecting a looped back NA does not solve the looped back DAD

Asati, et al.                Standards Track                    [Page 2]

RFC 7527                      Enhanced DAD                    April 2015

   problem.  Detection of any other looped back ND messages during the
   DAD operation is outside the scope of this document.  This document
   also includes a section on mitigation that discusses means already
   available to mitigate the DAD loopback problem.  This document
   updates RFCs 4429, 4861, and 4862.  It updates RFCs 4429 and 4862 to
   use the Enhanced DAD algorithm to detect looped back DAD probes, and
   it updates RFC 4861 as described in Section 4.3 below.

1.1.  Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   document are to be interpreted as described in [RFC2119].

1.2.  Terminology

   o  DAD-failed state - Duplication Address Detection failure as
      specified in [RFC4862].  Note even Optimistic DAD as specified in
      [RFC4429] can fail due to a looped back DAD probe.  This document
      covers looped back detection for Optimistic DAD as well.

   o  Looped back message - also referred to as a reflected message.
      The message sent by the sender is received by the sender due to
      the network or an upper-layer protocol on the sender looping the
      message back.

   o  Loopback - A function in which the router's Layer 3 interface (or
      the circuit to which the router's interface is connected) is
      looped back or connected to itself.  Loopback causes packets sent
      by the interface to be received by the interface and results in
      interface unavailability for regular data traffic forwarding.  See
      more details in Section 9.1 of [RFC2328].  The Loopback function
      is commonly used in an interface context to gain information on
      the quality of the interface, by employing mechanisms such as
      ICMPv6 pings and bit-error tests.  In a circuit context, this
      function is used in wide-area environments including optical Dense
      Wavelength Division Multiplexing (DWDM) and Synchronous Optical
      Network / Synchronous Digital Hierarchy (SONET/SDH) for fault
      isolation (e.g., by placing a loopback at different geographic
      locations along the path of a wide-area circuit to help locate a
      circuit fault).  The Loopback function may be employed locally or

   o  NS(DAD) - shorthand notation to denote a Neighbor Solicitation
      (NS) (as specified in [RFC4861]) that has an unspecified IPv6
      source address and was issued during DAD.

Asati, et al.                Standards Track                    [Page 3]

RFC 7527                      Enhanced DAD                    April 2015

2.  Problem Statement

   Service providers have reported a problem with DAD that arises in a
   few scenarios.  In the first scenario, loopback testing for
   troubleshooting purposes is underway on a circuit connected to an
   IPv6-enabled interface on a router.  The interface issues an NS for
   the IPv6 link-local address DAD.  The NS is reflected back to the
   router interface due to the loopback condition of the circuit, and
   the router interface enters a DAD-failed state.  After the loopback
   condition is removed, IPv4 will return to operation without further
   manual intervention.  However, IPv6 will remain in DAD-failed state
   until manual intervention on the router restores IPv6 to operation.

   In the second scenario, two broadband modems are served by the same
   service provider and terminate to the same Layer 3 interface on an
   IPv6-enabled access concentrator.  In this case, the two modems'
   Ethernet interfaces are also connected to a common local network
   (collision domain).  The access concentrator serving the modems is
   the first-hop IPv6 router for the modems and issues a NS(DAD) message
   for the IPv6 link-local address of its Layer 3 interface.  The NS
   message reaches one modem first, and this modem sends the message to
   the local network, where the second modem receives the message and
   then forwards it back to the access concentrator.  The looped back NS
   message causes the network interface on the access concentrator to be
   in a DAD-failed state.  Such a network interface typically serves
   thousands of broadband modems, and all would have their IPv6
   connectivity affected until the DAD-failed state is cleared.
   Additionally, it may be difficult for the user of the access
   concentrator to determine the source of the looped back DAD message.
   Thus, in order to avoid IPv6 outages that can potentially affect
   multiple users, there is a need for automated detection of looped
   back NS messages during DAD operations by a node.

   Note: In both examples above, the IPv6 link-local address DAD
   operation fails due to a looped back DAD probe.  However, the problem
   of a looped back DAD probe exists for any IPv6 address type including
   global addresses.

3.  Operational Mitigation Options

   Two mitigation options are described below that do not require any
   change to existing implementations.

3.1.  Disable DAD on an Interface

   One can disable DAD on an interface so that there are no NS(DAD)
   messages issued.  While this mitigation may be the simplest, the
   mitigation has three drawbacks: 1) care is needed when making such

Asati, et al.                Standards Track                    [Page 4]

RFC 7527                      Enhanced DAD                    April 2015

   configuration changes on point-to-point interfaces, 2) this is a one-
   time manual configuration on each interface, and 3) genuine
   duplicates on the link will not be detected.

   A service provider router, such as an access concentrator, or network
   core router, SHOULD support the DAD deactivation per interface.

3.2.  Dynamic Disable/Enable of DAD Using Layer 2 Protocol

   Some Layer 2 protocols include provisions to detect the existence of
   a loopback on an interface circuit, usually by comparing protocol
   data sent and received.  For example, the Point-to-Point Protocol
   (PPP) uses a magic number (Section 6.4 of [RFC1661]) to detect a
   loopback on an interface.

   When a Layer 2 protocol detects that a loopback is present on an
   interface circuit, the device MUST temporarily disable DAD on the
   interface.  When the protocol detects that a loopback is no longer
   present (or the interface state has changed), the device MUST
   (re-)enable DAD on that interface.

   This mitigation has several benefits.  It leverages the Layer 2
   protocol's built-in hardware loopback detection capability, if
   available.  Being a hardware solution, it scales better than the
   software solution proposed in this document.  This mitigation also
   scales better since it relies on an event-driven model that requires
   no additional state or timer.  This may be significant on devices
   with hundreds or thousands of interfaces that may be in loopback for
   long periods of time (e.g., awaiting turn-up).

   Detecting looped back DAD messages using a Layer 2 protocol SHOULD be
   enabled by default, and it MUST be a configurable option if the Layer
   2 technology provides means for detecting loopback messages on an
   interface circuit.

3.3.  Operational Considerations

   The mitigation options discussed above do not require the devices on
   both ends of the circuit to support the mitigation functionality
   simultaneously and do not propose any capability negotiation.  They
   are effective for unidirectional circuit or interface loopback (i.e.
   the loopback is placed in one direction on the circuit, rendering the
   other direction nonoperational), but they may not be effective for a
   bidirectional loopback (i.e., the loopback is placed in both
   directions of the circuit interface, so as to identify the faulty
   segment).  This is because, unless both ends followed a mitigation

Asati, et al.                Standards Track                    [Page 5]

RFC 7527                      Enhanced DAD                    April 2015

   option specified in this document, the noncompliant device would
   follow current behavior and disable IPv6 on that interface due to DAD
   until manual intervention restores it.

4.  The Enhanced DAD Algorithm

   The Enhanced DAD algorithm covers detection of a looped back NS(DAD)
   message.  This document proposes use of a random number in the Nonce
   Option specified in SEcure Neighbor Discovery (SEND) [RFC3971].  Note
   [RFC3971] does not provide a recommendation for pseudorandom
   functions.  Pseudorandom functions are covered in [RFC4086].  Since a
   nonce is used only once, the NS(DAD) for each IPv6 address of an
   interface uses a different nonce.  Additional details of the
   algorithm are included in Section 4.1.

   If there is a collision because two nodes used the same Target
   Address in their NS(DAD) and generated the same random nonce, then
   the algorithm will incorrectly detect a looped back NS(DAD) when a
   genuine address collision has occurred.  Since each looped back
   NS(DAD) event is logged to system management, the administrator of
   the network will have access to the information necessary to
   intervene manually.  Also, because the nodes will have detected what
   appear to be looped back NS(DAD) messages, they will continue to
   probe, and it is unlikely that they will choose the same nonce the
   second time (assuming quality random number generators).

   The algorithm is capable of detecting any ND solicitation (NS and
   Router Solicitation) or advertisement (NA and Router Advertisement)
   that is looped back.  However, there may be increased implementation
   complexity and memory usage for the sender node to store a nonce and
   nonce-related state for all ND messages.  Therefore, this document
   does not recommend using the algorithm outside of the DAD operation
   by an interface on a node.

4.1.  Processing Rules for Senders

   If a node has been configured to use the Enhanced DAD algorithm, when
   sending an NS(DAD) for a tentative or optimistic interface address,
   the sender MUST generate a random nonce associated with the interface
   address, MUST store the nonce internally, and MUST include the nonce
   in the Nonce option included in the NS(DAD).  If the interface does
   not receive any DAD failure indications within RetransTimer
   milliseconds (see [RFC4861]) after having sent DupAddrDetectTransmits
   Neighbor Solicitations, the interface moves the Target Address to the
   assigned state.

Asati, et al.                Standards Track                    [Page 6]

RFC 7527                      Enhanced DAD                    April 2015

   If any probe is looped back within RetransTimer milliseconds after
   having sent DupAddrDetectTransmits NS(DAD) messages, the interface
   continues with another MAX_MULTICAST_SOLICIT number of NS(DAD)
   messages transmitted RetransTimer milliseconds apart.  Section 2 of
   [RFC3971] defines a single-use nonce, so each Enhanced DAD probe uses
   a different nonce.  If no probe is looped back within RetransTimer
   milliseconds after MAX_MULTICAST_SOLICIT NS(DAD) messages are sent,
   the probing stops.  The probing MAY be stopped via manual
   intervention.  When probing is stopped, the interface moves the
   Target Address to the assigned state.

4.2.  Processing Rules for Receivers

   If the node has been configured to use the Enhanced DAD algorithm and
   an interface on the node receives any NS(DAD) message where the
   Target Address matches the interface address (in tentative or
   optimistic state), the receiver compares the nonce included in the
   message, with any stored nonce on the receiving interface.  If a
   match is found, the node SHOULD log a system management message,
   SHOULD update any statistics counter, and MUST drop the received
   message.  If the received NS(DAD) message includes a nonce and no
   match is found with any stored nonce, the node SHOULD log a system
   management message for a DAD-failed state and SHOULD update any
   statistics counter.

4.3.  Changes to RFC 4861

   The following text is appended to the Source Address definition in
   Section 4.3 of [RFC4861]:

   If a node has been configured to use the Enhanced DAD algorithm, an
   NS with an unspecified source address adds the Nonce option to the
   message and implements the state machine of the Enhanced DAD

   The following text is appended to the RetransTimer variable
   description in Section 6.3.2 of [RFC4861]:

   The RetransTimer MAY be overridden by a link-specific document if a
   node supports the Enhanced DAD algorithm.

5.  Action to Perform on Detecting a Genuine Duplicate

   As described in the paragraphs above, the nonce can also serve to
   detect genuine duplicates even when the network has potential for
   looping back ND messages.  When a genuine duplicate is detected, the
   node follows the manual intervention specified in Section 5.4.5 of
   [RFC4862].  However, in certain cases, if the genuine duplicate

Asati, et al.                Standards Track                    [Page 7]

RFC 7527                      Enhanced DAD                    April 2015

   matches the tentative or optimistic IPv6 address of a network
   interface of the access concentrator, additional automated action is

   Some networks follow a trust model where a trusted router serves
   untrusted IPv6 host nodes.  Operators of such networks have a desire
   to take automated action if a network interface of the trusted router
   has a tentative or optimistic address duplicated by a host.  One
   example of a type of access network is cable broadband deployment
   where the access concentrator is the first-hop IPv6 router to
   multiple broadband modems and supports proxying of DAD messages.  The
   network interface on the access concentrator initiates DAD for an
   IPv6 address and detects a genuine duplicate due to receiving an
   NS(DAD) or an NA message.  On detecting such a duplicate, the access
   concentrator SHOULD log a system management message, drop the
   received ND message, and block the modem on whose Layer 2 service
   identifier the duplicate NS(DAD) or NA message was received.  Any
   other network that follows the same trust model MAY use the automated
   action proposed in this section.

6.  Security Considerations

   This document does not improve or reduce the security posture of
   [RFC4862].  The nonce can be exploited by a rogue deliberately
   changing the nonce to fail the looped back detection specified by the
   Enhanced DAD algorithm.  SEND is recommended to circumvent this
   exploit.  Additionally, the nonce does not protect against the DoS
   caused by a rogue node replying by a fake NA to all DAD probes.  SEND
   is recommended to circumvent this exploit also.  Disabling DAD has an
   obvious security issue before a remote node on the link can issue
   reflected NS(DAD) messages.  Again, SEND is recommended for this
   exploit.  Source Address Validation Improvement (SAVI) [RFC6620] also
   protects against various attacks by on-link rogues.

7.  Normative References

   [RFC1661]  Simpson, W., Ed., "The Point-to-Point Protocol (PPP)", STD
              51, RFC 1661, July 1994,

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997,

   [RFC2328]  Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998,

Asati, et al.                Standards Track                    [Page 8]

RFC 7527                      Enhanced DAD                    April 2015

   [RFC3971]  Arkko, J., Ed., Kempf, J., Zill, B., and P. Nikander,
              "SEcure Neighbor Discovery (SEND)", RFC 3971, March 2005,

   [RFC4086]  Eastlake 3rd, D., Schiller, J., and S. Crocker,
              "Randomness Requirements for Security", BCP 106, RFC 4086,
              June 2005, <http://www.rfc-editor.org/info/rfc4086>.

   [RFC4429]  Moore, N., "Optimistic Duplicate Address Detection (DAD)
              for IPv6", RFC 4429, April 2006,

   [RFC4861]  Narten, T., Nordmark, E., Simpson, W., and H. Soliman,
              "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861,
              September 2007, <http://www.rfc-editor.org/info/rfc4861>.

   [RFC4862]  Thomson, S., Narten, T., and T. Jinmei, "IPv6 Stateless
              Address Autoconfiguration", RFC 4862, September 2007,

   [RFC6620]  Nordmark, E., Bagnulo, M., and E. Levy-Abegnoli, "FCFS
              SAVI: First-Come, First-Served Source Address Validation
              Improvement for Locally Assigned IPv6 Addresses", RFC
              6620, May 2012, <http://www.rfc-editor.org/info/rfc6620>.


   Thanks (in alphabetical order by first name) to Adrian Farrel, Benoit
   Claise, Bernie Volz, Brian Haberman, Dmitry Anipko, Eric Levy-
   Abegnoli, Eric Vyncke, Erik Nordmark, Fred Templin, Hilarie Orman,
   Jouni Korhonen, Michael Sinatra, Ole Troan, Pascal Thubert, Ray
   Hunter, Suresh Krishnan, Tassos Chatzithomaoglou, and Tim Chown for
   their guidance and review of the document.  Thanks to Thomas Narten
   for encouraging this work.  Thanks to Steinar Haug and Scott Beuker
   for describing some of the use cases.

Asati, et al.                Standards Track                    [Page 9]

RFC 7527                      Enhanced DAD                    April 2015

Authors' Addresses

   Rajiv Asati
   Cisco Systems, Inc.
   7025 Kit Creek road
   Research Triangle Park, NC  27709-4987
   United States

   EMail: rajiva@cisco.com
   URI:   http://www.cisco.com/

   Hemant Singh
   Cisco Systems, Inc.
   1414 Massachusetts Ave.
   Boxborough, MA  01719
   United States

   Phone: +1 978 936 1622
   EMail: shemant@cisco.com
   URI:   http://www.cisco.com/

   Wes Beebee
   Cisco Systems, Inc.
   1414 Massachusetts Ave.
   Boxborough, MA  01719
   United States

   Phone: +1 978 936 2030
   EMail: wbeebee@cisco.com
   URI:   http://www.cisco.com/

   Carlos Pignataro
   Cisco Systems, Inc.
   7200-12 Kit Creek Road
   Research Triangle Park, NC  27709
   United States

   EMail: cpignata@cisco.com
   URI:   http://www.cisco.com/

Asati, et al.                Standards Track                   [Page 10]

RFC 7527                      Enhanced DAD                    April 2015

   Eli Dart
   Lawrence Berkeley National Laboratory
   1 Cyclotron Road, Berkeley, CA 94720
   United States

   EMail: dart@es.net
   URI:   http://www.es.net/

   Wesley George
   Time Warner Cable
   13820 Sunrise Valley Drive
   Herndon, VA  20171
   United States

   EMail: wesley.george@twcable.com

Asati, et al.                Standards Track                   [Page 11]