RFC 5909






Internet Engineering Task Force (IETF)                       J-M. Combes
Request for Comments: 5909                         France Telecom Orange
Category: Informational                                      S. Krishnan
ISSN: 2070-1721                                                 Ericsson
                                                                G. Daley
                                                       Netstar Logicalis
                                                               July 2010


          Securing Neighbor Discovery Proxy: Problem Statement

Abstract



   Neighbor Discovery Proxies are used to provide an address presence on
   a link for nodes that are no longer present on the link.  They allow
   a node to receive packets directed at its address by allowing another
   device to perform Neighbor Discovery operations on its behalf.

   Neighbor Discovery Proxy is used in Mobile IPv6 and related protocols
   to provide reachability from nodes on the home network when a Mobile
   Node is not at home, by allowing the Home Agent to act as proxy.  It
   is also used as a mechanism to allow a global prefix to span multiple
   links, where proxies act as relays for Neighbor Discovery messages.

   Neighbor Discovery Proxy currently cannot be secured using Secure
   Neighbor Discovery (SEND).  Today, SEND assumes that a node
   advertising an address is the address owner and in possession of
   appropriate public and private keys for that node.  This document
   describes how existing practice for proxy Neighbor Discovery relates
   to SEND.

Status of This Memo



   This document is not an Internet Standards Track specification; it is
   published for informational purposes.

   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).  Not all documents
   approved by the IESG are a candidate for any level of Internet
   Standard; see 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
   http://www.rfc-editor.org/info/rfc5909.





Combes, et al.                Informational                     [Page 1]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


Copyright Notice



   Copyright (c) 2010 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.





































Combes, et al.                Informational                     [Page 2]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


Table of Contents



   1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . .  3
   2.  Scenarios  . . . . . . . . . . . . . . . . . . . . . . . . . .  4
     2.1.  IPv6 Mobile Nodes and Neighbor Discovery Proxy . . . . . .  4
     2.2.  IPv6 Fixed Nodes and Neighbor Discovery Proxy  . . . . . .  6
     2.3.  Bridge-Like ND Proxies . . . . . . . . . . . . . . . . . .  6
   3.  Proxy Neighbor Discovery and SEND  . . . . . . . . . . . . . .  9
     3.1.  CGA Signatures and Proxy Neighbor Discovery  . . . . . . .  9
     3.2.  Non-CGA Signatures and Proxy Neighbor Discovery  . . . . . 10
     3.3.  Securing Proxy DAD . . . . . . . . . . . . . . . . . . . . 11
     3.4.  Securing Router Advertisements . . . . . . . . . . . . . . 11
   4.  Potential Approaches to Securing Proxy ND  . . . . . . . . . . 12
     4.1.  Secured Proxy ND and Mobile IPv6 . . . . . . . . . . . . . 12
       4.1.1.  Mobile IPv6 and Router-Based Authorization . . . . . . 13
       4.1.2.  Mobile IPv6 and Per-Address Authorization  . . . . . . 13
       4.1.3.  Cryptographic-Based Solutions  . . . . . . . . . . . . 13
       4.1.4.  Solution Based on the 'Point-to-Point' Link Model  . . 14
     4.2.  Secured Proxy ND and Bridge-Like Proxies . . . . . . . . . 14
       4.2.1.  Authorization Delegation . . . . . . . . . . . . . . . 14
       4.2.2.  Unauthorized Routers and Proxies . . . . . . . . . . . 14
       4.2.3.  Multiple Proxy Spans . . . . . . . . . . . . . . . . . 15
       4.2.4.  Routing Infrastructure Delegation  . . . . . . . . . . 15
       4.2.5.  Local Delegation . . . . . . . . . . . . . . . . . . . 16
       4.2.6.  Host Delegation of Trust to Proxies  . . . . . . . . . 17
     4.3.  Proxying Unsecured Addresses . . . . . . . . . . . . . . . 17
   5.  Two or More Nodes Defending the Same Address . . . . . . . . . 18
   6.  Security Considerations  . . . . . . . . . . . . . . . . . . . 19
     6.1.  Router Trust Assumption  . . . . . . . . . . . . . . . . . 19
     6.2.  Certificate Transport  . . . . . . . . . . . . . . . . . . 19
     6.3.  Timekeeping  . . . . . . . . . . . . . . . . . . . . . . . 19
   7.  Acknowledgments  . . . . . . . . . . . . . . . . . . . . . . . 20
   8.  References . . . . . . . . . . . . . . . . . . . . . . . . . . 20
     8.1.  Normative References . . . . . . . . . . . . . . . . . . . 20
     8.2.  Informative References . . . . . . . . . . . . . . . . . . 21

1.  Introduction



   Neighbor Discovery Proxy is defined in IPv6 Neighbor Discovery
   [RFC4861].  It is used in networks where a prefix has to span
   multiple links [RFC4389] but also in Mobile IPv6 [RFC3775] (and so in
   Mobile-IPv6-based protocols like Network Mobility (NEMO) [RFC3963],
   Fast Handovers for Mobile IPv6 (FMIPv6) [RFC5568], or Hierarchical
   Mobile IPv6 (HMIPv6) [RFC5380]) and in the Internet Key Exchange
   Protocol (IKE) version 2 (IKEv2) [RFC4306].  It allows a device that
   is not physically present on a link to have another advertise its
   presence, and forward packets to the off-link device.




Combes, et al.                Informational                     [Page 3]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   Neighbor Discovery Proxy relies upon another device, the proxy, to
   monitor for Neighbor Solicitations (NSs), and answer with Neighbor
   Advertisements (NAs).  These proxy Neighbor Advertisements direct
   data traffic through the proxy.  Proxied traffic is then forwarded to
   the end destination.

2.  Scenarios



   This section describes the different scenarios where the interaction
   between Secure Neighbor Discovery (SEND) and ND Proxy raises issues.

2.1.  IPv6 Mobile Nodes and Neighbor Discovery Proxy



   The goal of IPv6 mobility is to allow nodes to remain reachable while
   moving around in the IPv6 Internet.  The following text is focused on
   Mobile IPv6 but the issue raised by the interaction between SEND and
   ND Proxy may be the same with Mobile IPv6 based protocols (e.g.,
   NEMO, HMIPv6).

   For Mobile IPv6 Mobile Nodes (MNs), it is necessary to keep existing
   sessions going or to allow new sessions even when one leaves the home
   network.

   In order to continue existing sessions, when nodes are present on the
   home link, the Proxy (i.e., the Home Agent in Mobile IPv6) sends an
   unsolicited NA to the all-nodes multicast address on the home link as
   specified [RFC3775].

   For new sessions, the Proxy, which listens to the MN's address
   responds with a Neighbor Advertisement that originates at its own
   IPv6 address and has the proxy's address as the Target Link-Layer
   Address, but contains the absent mobile in the Target Address field
   of the Neighbor Advertisement.  In this case, SEND cannot be applied
   because the address in the Target Address field is not the same as
   the one in the Source Address field of the IP header.

   As seen in Figure 1, solicitors send a multicast solicitation to the
   solicited nodes multicast address (based on the unicast address) of
   the absent node (a mobile node that is away from the home link).












Combes, et al.                Informational                     [Page 4]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


            Absent Mobile       Proxy         Solicitor

                                        NS:SL3=S,DL3=Sol(A),TA=A
                               +-----+     SL2=s,DL2=sol(a),SLL=s
                               |     |<================
                               |     |
                               |     |================>
                               +-----+  NA:SL3=P,DL3=S,TA=A,
                                           SL2=p,DL2=s,TLL=p

   Legend:
      SL3: Source      IPv6 Address         NS: Neighbor Solicitation
      DL3: Destination IPv6 Address         NA: Neighbor Advertisement
      SL2: Source Link-Layer Address        RS: Router Solicitation
      DL2: Destination Link-Layer Address   RA: Router Advertisement
      TA:  Target Address
      SLL/TLL:  Source/Target Link-Layer Address Option

                                 Figure 1

   While at home, if the MN has configured Cryptographically Generated
   Addresses (CGAs) [RFC3972], it can secure establishment by its on-
   link neighbors of Neighbor Cache Entries (NCEs) for its CGAs by using
   SEND [RFC3971].  SEND security requires a node sending Neighbor
   Advertisements for a given address to be in possession of the public/
   private key pair that generated the address.

   When an MN moves away from the home link, a proxy has to undertake
   Neighbor Discovery signaling on behalf of the MN.  In Mobile IPv6,
   the role of the proxy is undertaken by the Home Agent.  While the
   Home Agent has a security association with the MN, it does not have
   access to the public/private key pair used to generate the MN's CGA.
   Thus, the Home Agent acting as an ND proxy cannot use SEND for the
   address it is proxying [RFC3971].

   When an MN moves from the home network to a visited network, the
   proxy will have to override the MN's existing Neighbor Cache Entries
   that are flagged as secure [RFC3971].  This is needed for the Home
   Agent to intercept traffic sent on-link to the MN that would
   otherwise be sent to the MN's link-layer address.

   With the current SEND specification, any solicitation or
   advertisement sent by the proxy will be unsecure and thus will not be
   able to update the MN's NCE for the home address because it is
   flagged as secured.  These existing Neighbor Cache Entries will only
   time-out after Neighbor Unreachability Detection [RFC4861] concludes
   the Home Address is unreachable at the link layer recorded in the
   NCE.



Combes, et al.                Informational                     [Page 5]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   Where secured proxy services are not able to be provided, a proxy's
   advertisement may be overridden by a rogue proxy without the
   receiving host realizing that an attack has occurred.  This is
   identical to what happens in a network where SEND is not deployed.

2.2.  IPv6 Fixed Nodes and Neighbor Discovery Proxy



   This scenario is a sub-case of the previous one.  In this scenario,
   the IPv6 node will never be on the link where the ND messages are
   proxied.  For example, an IPv6 node gains remote access to a network
   protected by a security gateway that runs IKEv2 [RFC4306].  When a
   node needs an IP address in the network protected by a security
   gateway, the security gateway assigns an address dynamically using
   Configuration Payload during IKEv2 exchanges.  The security gateway
   then needs to receive packets sent to this address; one way to do so
   would be to proxy ND messages.

2.3.  Bridge-Like ND Proxies



   The Neighbor Discovery (ND) Proxy specification [RFC4389] defines an
   alternative method to classic bridging.  Just as with classic
   bridging, multiple link-layer segments are bridged into a single
   segment, but with the help of proxying at the IP layer rather than
   link-layer bridging.  In this case, the proxy forwards messages while
   modifying their source and destination MAC addresses, and it rewrites
   their solicited and override flags and Link-Layer Address Options.

   This rewriting is incompatible with SEND signed messages for a number
   of reasons:

   o  Rewriting elements within the message will break the digital
      signature.

   o  The source IP address of each packet is the packet's origin, not
      the proxy's address.  The proxy is unable to generate another
      signature for this address, as it doesn't have the CGA private key
      [RFC3971].

   Thus, proxy modification of SEND solicitations may require sharing of
   credentials between the proxied node and the proxying node or
   creation of new options with proxying capabilities.

   While bridge-like ND proxies aim to provide as little interference
   with ND mechanisms as possible, SEND has been designed to prevent
   modification or spoofing of advertisements by devices on the link.






Combes, et al.                Informational                     [Page 6]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   Of particular note is the fact that ND Proxy performs a different
   kind of proxy Neighbor Discovery to Mobile IPv6 [RFC3775] [RFC4389].
   RFC 3775 (Mobile IPv6) specifies that the Home Agent as proxy sends
   Neighbor Advertisements from its own address with the Target Address
   set to the absent Mobile Node's address.  The Home Agent's own link-
   layer address is placed in the Target Link-Layer Address Option
   [RFC3775].  On the other hand, ND Proxy resends messages containing
   their original address, even after modification (i.e., the IP source
   address remains the same) [RFC4389].  Figure 2 describes packet
   formats for proxy Neighbor solicitation and advertisement as
   specified by RFC 4389.

            Advertiser          Proxy         Solicitor

     NS:SL3=S,DL3=Sol(A),TA=A,          NS:SL3=S,DL3=Sol(A),TA=A,
        SL2=p,DL2=sol(a),SLL=p +-----+      SL2=s,DL2=sol(a),SLL=s
            <==================|     |<================
                               |     |
            ==================>|     |================>
     NA:SL3=A,DL3=S,TA=A,      +-----+  NA:SL3=A,DL3=S,TA=A
        SL2=a,DL2=p,TLL=a                  SL2=p,DL2=s,TLL=p

   Legend:
      SL3: Source      IPv6 Address         NS: Neighbor Solicitation
      DL3: Destination IPv6 Address         NA: Neighbor Advertisement
      SL2: Source Link-Layer Address
      DL2: Destination Link-Layer Address
      TA:  Target Address
      SLL/TLL:  Source/Target Link-Layer Address Option

                                 Figure 2

   In order to use the same security procedures for both ND Proxy and
   Mobile IPv6, changes may be needed to the proxying procedures in
   [RFC4389], as well as changes to SEND.

   An additional (and undocumented) requirement for bridge-like proxying
   is the operation of router discovery.  Router discovery packets may
   similarly modify Neighbor Cache state, and require protection from
   SEND.

   In Figure 3, the router discovery messages propagate without
   modification to the router address, but elements within the message
   change.  This is consistent with the description of Neighbor
   Discovery above.






Combes, et al.                Informational                     [Page 7]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


            Advertiser          Proxy         Solicitor

     RS:SL3=S,DL3=AllR,                 RS:SL3=S,DL3=AllR,
        SL2=p,DL2=allr,SLL=p   +-----+     SL2=s,DL2=allr,SLL=s
            <==================|     |<================
                               |     |
            ==================>|     |================>
     RA:SL3=A,DL3=S,           +-----+  RA:SL3=A,DL3=S,
        SL2=a,DL2=p,SLL=a                 SL2=p,DL2=s,SLL=p

   Legend:
      SL3: Source      IPv6 Address         RS: Router Solicitation
      DL3: Destination IPv6 Address         RA: Router Advertisement
      SL2: Source Link-Layer Address
      DL2: Destination Link-Layer Address
      TA:  Target Address
      SLL/TLL:  Source/Target Link-Layer Address Option

                                 Figure 3

   Once again, these messages may not be signed with a CGA signature by
   the proxy, because it does not own the source address.

   Additionally, Authorization Delegation Discovery messages need to be
   exchanged for bridge-like ND proxies to prove their authority to
   forward.  Unless the proxy receives explicit authority to act as a
   router, or the router knows of its presence, no authorization may be
   made.  This explicit authorization requirement may be at odds with
   the zero configuration goal of ND proxying [RFC4389].

   An alternative (alluded to in an appendix of ND Proxy [RFC4389])
   suggests that the proxy send Router Advertisements (RAs) from its own
   address.  As described by ND Proxy, this is insufficient for
   providing proxied Neighbor Advertisement service, but may be matched
   with Neighbor solicitation and advertisement services using the
   proxy's source address in the same way as Mobile IPv6 [RFC4389]
   [RFC3775].  This means that all router and Neighbor advertisements
   would come from the proxied address, but may contain a target address
   that allows proxied Neighbor presence to be established with peers on
   other segments.  Router discovery in this case has the identity of
   the original (non-proxy) router completely obscured in router
   discovery messages.

   The resultant proxy messages would have no identifying information
   indicating their origin, which means that proxying between multiple
   links would require state to be stored on outstanding solicitations
   (effectively a ND only NAT).  This level of state storage may be
   undesirable.



Combes, et al.                Informational                     [Page 8]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   Mobile IPv6 does not experience this issue when supplying its own
   address, since ND messages are never forwarded on to the absent node
   (the Home Agent having sufficient information to respond itself).

   Authorization from a router may still be required for Router
   Advertisement, and will be discussed in Section 4.2.

3.  Proxy Neighbor Discovery and SEND



   There are currently no existing secured Neighbor Discovery procedures
   for proxied addresses, and all Neighbor Advertisements from SEND
   nodes are required to have equal source and target addresses, and be
   signed by the transmitter (Section 7.4 of [RFC3971]).

   Signatures over SEND messages are required to be applied on the CGA
   source address of the message, and there is no way of indicating that
   a message is proxied.

   Even if the message is able to be transmitted from the original
   owner, differences in link-layer addressing and options require
   modification by a proxy.  If a message is signed with a CGA-based
   signature, the proxy is unable to regenerate a signature over the
   changed message as it lacks the keying material.

   Therefore, a router wishing to provide proxy Neighbor Advertisement
   service cannot use existing SEND procedures on those messages.

   A host may wish to establish a session with a device that is not on-
   link but is proxied.  As a SEND host, it prefers to create Neighbor
   Cache Entries using secured procedures.  Since SEND signatures cannot
   be applied to an existing proxy Neighbor Advertisement, it must
   accept non-SEND advertisements in order to receive proxy Neighbor
   Advertisements.

   Neighbor Cache spoofing of another node therefore becomes trivial, as
   any address may be proxy-advertised to the SEND node, and overridden
   only if the node is there to protect itself.  When a node is present
   to defend itself, it may also be difficult for the solicitor
   determine the difference between a proxy-spoofing attack, and a
   situation where a proxied device returns to a link and overrides
   other proxy advertisers [RFC4861].

3.1.  CGA Signatures and Proxy Neighbor Discovery



   SEND defines one public-key and signature format for use with
   Cryptographically Generated Addresses (CGAs) [RFC3972].  CGAs are
   intended to tie address ownership to a particular public/private key
   pair.



Combes, et al.                Informational                     [Page 9]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   In SEND as defined today, Neighbor Discovery messages (including the
   IP Addresses from the IPv6 header) are signed with the same key used
   to generate the CGA.  This means that message recipients have proof
   that the signer of the message owned the address.

   When a proxy replaces the message's source IPv6 address with its own
   CGA, the existing CGA option and RSA signature option would need to
   be replaced with ones that correspond to the CGA of the proxy.  To be
   valid according to the SEND specification, the Target Address of the
   Neighbor Advertisement message would need to be replaced also to be
   equal to the Source Address [RFC3971].

   Additional authorization information may be needed to prove that the
   proxy is indeed allowed to advertise for the target address, as is
   described in Section 4.

3.2.  Non-CGA Signatures and Proxy Neighbor Discovery



   Where a proxy retains the original source address in a proxied
   message, existing security checks for SEND will fail, since fields
   within the message will be changed.  In order to achieve secured
   proxy Neighbor Discovery in this case, extended authorization
   mechanisms may be needed for SEND.

   SEND provides mechanisms for extension of SEND to non-CGA-based
   authorization.  Messages are available for Authorization Delegation
   Discovery, which is able to carry arbitrary PKIX/X.509 certificates
   [RFC5280].

   There is, however, no specification of keying information option
   formats analogous to the SEND CGA Option [RFC3971].  The existing
   option allows a host to verify message integrity by specifying a key
   and algorithm for digital signature, without providing authorization
   via mechanisms other than CGA ownership.

   The digital signature in SEND is transported in the RSA Signature
   Option.  As currently specified, the signature operation is performed
   over a CGA Message type, and allows for CGA verification.  Updating
   the signature function to support non-CGA operations may be
   necessary.

   Within SEND, more advanced functions such as routing may be
   authorized by certificate path verification using Authorization
   Delegation Discovery.

   With non-CGA signatures and authentication, certificate contents for
   authorization may need to be determined, as outlined in Section 4.




Combes, et al.                Informational                    [Page 10]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   While SEND provides for extensions to new non-CGA methods, existing
   SEND hosts may silently discard messages with unverifiable RSA
   signature options (Section 5.2.2 of [RFC3971]), if configured only to
   accept SEND messages.  In cases where unsecured Neighbor Cache
   Entries are still accepted, messages from new algorithms will be
   treated as unsecured.

3.3.  Securing Proxy DAD



   Initiation of proxy Neighbor Discovery also requires Duplicate
   Address Detection (DAD) checks of the address [RFC4862].  These DAD
   checks need to be performed by sending Neighbor Solicitations, from
   the unspecified source address, with the target being the proxied
   address.

   In existing SEND procedures, the address that is used for CGA tests
   on DAD NS is the target address.  A Proxy that originates this
   message while the proxied address owner is absent is unable to
   generate a CGA-based signature for this address and must undertake
   DAD with an unsecured NS.  It may be possible that the proxy can
   ensure that responding NAs are secured though.

   Where bridge-like ND proxy operations are being performed, DAD NSs
   may be copied from the original source, without modification
   (considering they have an unspecified source address and contain no
   link-layer address options) [RFC4389].

   If non-CGA-based signatures are available, then the signature over
   the DAD NS doesn't need to have a CGA relationship to the Target
   Address, but authorization for address configuration needs to be
   shown using certificates.

   In case there is a DAD collision between two SEND nodes on different
   interfaces of the proxy, it is possible that the proxy may not have
   the authority to modify the NA defending the address.  In this case,
   the proxy still needs to modify the NA and pass it onto the other
   interfaces even if it will fail SEND verification on the receiving
   node.

3.4.  Securing Router Advertisements



   While Router Solicitations are protected in the same manner as
   Neighbor Solicitations, the security for Router Advertisements is
   mainly based on the use of certificates.  Even though the mechanism
   for securing RAs is different, the problems that arise due to the
   modification of the L2 addresses are exactly the same: the proxy
   needs to have the right security material (e.g., certificate) to sign
   the RA messages after modification.



Combes, et al.                Informational                    [Page 11]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


4.  Potential Approaches to Securing Proxy ND



   SEND nodes already have the concept of delegated authority through
   requiring external authorization of routers to perform their routing
   and advertisement roles.  The authorization of these routers takes
   the form of delegation certificates.

   Proxy Neighbor Discovery requires a delegation of authority (on
   behalf of the absent address owner) to the proxier.  Without this
   authority, other devices on the link have no reason to trust an
   advertiser.

   For bridge-like proxies, it is assumed that there is no preexisting
   trust between the host owning the address and the proxy.  Therefore,
   authority may necessarily be dynamic or based on topological roles
   within the network [RFC4389].

   Existing trust relationships lend themselves to providing authority
   for proxying in two alternative ways.

   First, the SEND router authorization mechanisms described above
   provide delegation from the organization responsible for routing in
   an address domain to the certified routers.  It may be argued that
   routers so certified may be trusted to provide service for nodes that
   form part of a link's address range, but are themselves absent.
   Devices which are proxies could either be granted the right to proxy
   by the network's router, or be implicitly allowed to proxy by virtue
   of being an authorized router.

   Second, where the proxied address is itself a CGA, the holder of the
   public and private keys is seen to be authoritative about the
   address's use.  If this address owner was able to sign the proxier's
   address and public key information, it would be possible to identify
   that the proxy is known and trusted by the CGA address owner for
   proxy service.  This method requires that the proxied address know or
   learn the proxy's address and public key, and that the certificate
   signed by the proxied node's is passed to the proxy, either while
   they share the same link, or at a later stage.

   In both methods, the original address owner's advertisements need to
   override the proxy if it suddenly returns, and therefore timing and
   replay protection from such messages need to be carefully considered.

4.1.  Secured Proxy ND and Mobile IPv6



   Mobile IPv6 has a security association between the Mobile Node and
   Home Agent.  The Mobile Node sends a Binding Update to the Home
   Agent, to indicate that it is not at home.  This implies that the



Combes, et al.                Informational                    [Page 12]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   Mobile Node wishes the Home Agent to begin proxy Neighbor Discovery
   operations for its home address(es).

4.1.1.  Mobile IPv6 and Router-Based Authorization



   A secured Proxy Neighbor Advertisements proposal based on existing
   router trust would require no explicit authorization signaling
   between HA and MN to allow proxying.  Hosts on the home link will
   believe proxied advertisements solely because they come from a
   trusted router.

   Where the home agent operates as a router without explicit trust to
   route from the advertising routing infrastructure (such as in a home,
   with a router managed by an ISP), more explicit proxying
   authorization may be required, as described in Section 4.2.

4.1.2.  Mobile IPv6 and Per-Address Authorization



   Where proxy Neighbor Discovery is delegated by the MN to the home
   agent, the MN needs to learn the public key for the Home Agent, so
   that it can generate a certificate authorizing the public/private key
   pair to be used in proxying.  It may conceivably do this using
   Certificate Path Solicitations either over a home tunnel, when it is
   away from home, or during router discovery while still at home
   [RFC3971] [RFC3775].

   When sending its Binding Update to the HA, the MN would need to
   provide a certificate containing the subject's (i.e., proxy HA's)
   public key and address, the issuer's (i.e., MN's) CGA and public key,
   and timestamps indicating when the authority began and when it ends.
   This certificate would need to be transmitted at binding time.
   Messaging or such an exchange mechanism would have to be developed.

4.1.3.  Cryptographic-Based Solutions



   Specific cryptographic algorithms may help to allow trust between
   entities of a same group.

   This is the case, for example, with ring signature algorithms.  These
   algorithms generate a signature using the private key of any member
   from the same group, but to verify the signature the public keys of
   all group members are required.  Applied to SEND, the addresses are
   cryptographically generated using multiple public keys, and the
   Neighbor Discovery messages are signed with an RSA ring signature
   [RING].  (Note that the cryptographic algorithms that are the
   foundation for [RING] and other similar solutions are not widely
   accepted in the security community; additional research is needed
   before a Standards Track protocol could be developed.)



Combes, et al.                Informational                    [Page 13]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


4.1.4.  Solution Based on the 'Point-to-Point' Link Model



   Another approach is to use the 'Point-to-Point' link model.

   In this model, one prefix is provided per MN, and only an MN and the
   HA are on a same link.  The consequence is the HA no longer needs to
   act as ND Proxy.

   One way to design such a solution is to use virtual interfaces, on
   the MN and the HA, and a virtual link between them.  Addresses
   generated on the virtual interfaces will only be advertised on the
   virtual link.  For Mobile IPv6, this results in a virtual Home
   Network where the MN will never come back.

4.2.  Secured Proxy ND and Bridge-Like Proxies



   In link-extension environments, the role of a proxy is more
   explicitly separated from that of a router.  In SEND, routers may
   expect to be authorized by the routing infrastructure to advertise
   and may provide this authority to hosts in order to allow them to
   change forwarding state.

   Proxies are not part of the traditional infrastructure of the
   Internet, and hosts or routers may not have an explicit reason to
   trust them, except that they can forward packets to regions where
   otherwise those hosts or routers could not reach.

4.2.1.  Authorization Delegation



   If a proxy can convince a device that it should be trusted to perform
   proxying function, it may require that device to vouch for its
   operation in dealing with other devices.  It may do this by receiving
   a certificate, signed by the originating device that the proxy is
   believed capable of proxying under certain circumstances.

   This allows nodes receiving proxied Neighbor Discovery packets to
   quickly check if the proxy is authorized for the operation.  There
   are several bases for such trust, and requirements in proxied
   environments, which are discussed below.

4.2.2.  Unauthorized Routers and Proxies



   Routers may be advertising on networks without any explicit
   authorization, and SEND hosts will register these routers if there
   are no other options [RFC3971].  While proxies may similarly attempt
   to advertise without authority, this provides no security for the
   routing infrastructure.  Any device can be setup as a SEND proxy/
   router so long as it signs its own ND messages from its CGA.



Combes, et al.                Informational                    [Page 14]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   This may not help in the case that a proxy attempts to update
   Neighbor Cache Entries for a SEND node that moves between links,
   since the SEND node's authority to advertise its own CGA address
   would not be superseded by a proxy with no credentials.

4.2.3.  Multiple Proxy Spans



   Proxies may have multiple levels of nesting, which allow the network
   to connect between non-adjacent segments.

   In this case, authority delegated at one point will have to be
   redelegated (possibly in a diluted form) to proxies further away from
   the origin of the trust.

       Trust        Proxy A            Proxy B     Distant
       Origin - T                                  Node - D

        +-----+                                    +-----+
        |     |                                    |     |
        +-----+     +-----+            +-----+     +-----+
           |        |     |            |     |        |
        ------------|     |------------|     |----------
                    |     |            |     |
                    +-----+            +-----+
          ==========>     ==============>    ==========>
          Deleg(A,T)    Deleg(B,Deleg(A,T))   Advertise(D, Deleg(B,
                                                    Deleg(A,T))

                                 Figure 4

   As shown in Figure 4, the Proxy A needs to redelegate authority to
   proxy for T to Proxy B; this allows it to proxy advertisements that
   target T back to D.

4.2.4.  Routing Infrastructure Delegation



   Where it is possible for the proxy to pre-establish trust with the
   routing infrastructure, or at least to the local router, it may be
   possible to authorize proxying as a function of routing within the
   subnet.  The router or CA may then be able to certify proxying for
   only a subset of the prefixes for which it is itself certified.

   If a router or CA provides certification for a particular prefix, it
   may be able to indicate that only proxying is supported, so that
   Neighbor Cache Entries of routers connected to Internet
   infrastructure are never overridden by the proxy, if the router is
   present on a segment.




Combes, et al.                Informational                    [Page 15]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   Hosts understanding such certificates may allow authorized proxies
   and routers to override the host when assuming proxy roles, if the
   host is absent.

   Proxy certificate signing could be done either dynamically (requiring
   exchanges of identity and authorization information) or statically
   when the network is set up.

4.2.5.  Local Delegation



   Where no trust tie exists between the authority that provides the
   routing infrastructure and the provider of bridging and proxying
   services, it may still be possible for SEND hosts to trust the
   bridging provider to authorize proxying operations.

   SEND itself requires that routers be able to show authorization, but
   doesn't require routers to have a single trusted root.

   A local bridging/proxying authority trust delegation may be possible.
   It would be possible for this authority to pass out local-use
   certificates, allowing proxying on a specific subnet or subnets, with
   a separate authorization chain to those subnets for the routers with
   Internet access.

   This would require little modification to SEND, other than the
   addition of router-based proxy authority (as in Section 4.2.4), and
   proxies would in effect be treated as routers by SEND hosts
   [RFC3971].  Distribution of keying and trust material for the initial
   bootstrap of proxies would not be provided though (and may be
   static).

   Within small domains, key management and distribution may be a
   tractable problem, so long as these operations are simple enough to
   perform.

   Since these domains may be small, it may be necessary to provide
   certificate chains for trust anchors that weren't requested in
   Certificate Path Solicitations, if the proxy doesn't have a trust
   chain to any requested trust anchor.

   This is akin to 'suggesting' an appropriate trusted root.  It may
   allow for user action in allowing trust extension when visiting
   domains without ties to a global keying infrastructure.  In this
   case, the trust chain would have to start with a self-signed
   certificate from the original CA.






Combes, et al.                Informational                    [Page 16]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


4.2.6.  Host Delegation of Trust to Proxies



   Unlike Mobile IPv6, for bridge-like proxied networks, there is no
   existing security association upon which to transport proxying
   authorization credentials.

   Thus, proxies need to convince Neighbors to delegate proxy authority
   to them, in order to proxy-advertise to nodes on different segments.
   It will be difficult without additional information to distinguish
   between legitimate proxies and devices that have no need or right to
   proxy (and may want to make two network segments appear connected).

   When proxy advertising, proxies must not only identify that proxying
   needs to occur, but provide proof that they are allowed to do so, so
   that SEND Neighbor Cache Entries may be updated.  Unless the
   authorization to update such entries is tied to address ownership
   proofs from the proxied host or the verifiable routing
   infrastructure, spoofing may occur.

   When a host received a proxied Neighbor advertisement, it would be
   necessary to check authorization in the same way that authorization
   delegation discovery is performed in SEND.

   Otherwise, certificate transport will be required to exchange
   authorization between proxied nodes and proxies.

   Proxies would have to be able to delegate this authorization to
   downstream proxies, as described in Section 4.2.3.

4.3.  Proxying Unsecured Addresses



   Where the original Neighbor Discovery message is unsecured, there is
   an argument for not providing secured proxy service for that node.

   In both the Mobile IPv6 and extended networks cases, the node may
   arrive back at the network and require other hosts to map their
   existing Neighbor Cache Entry to the node's link-layer address.  The
   re-arriving node's overriding of link-layer address mappings will
   occur without SEND in this case.

   It is notable that without SEND protection any node may spoof the
   arrival, and effectively steal service across an extended network.
   This is the same as in the non-proxy case, and is not made
   significantly worse by the proxy's presence (although the identity of
   the attacker may be masked if source addresses are being replaced).






Combes, et al.                Informational                    [Page 17]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   If signatures over the proxied messages were to be used, re-arrival
   and override of the Neighbor Cache Entries would have to be allowed,
   so the signatures would indicate that at least the proxy wasn't
   spoofing (even if the original sender was).

   For non-SEND routers, though, it may be possible for secured proxies
   to send signed router advertisement messages, in order to ensure that
   routers aren't spoofed, and subsequently switched to different parts
   of the extended network.

   This has problems in that the origin is again unsecured, and any node
   on the network could spoof router advertisement for an unsecured
   address.  These spoofed messages may become almost indistinguishable
   (except for the non-CGA origin address) from unspoofed messages from
   SEND routers.

   Given these complexities, the simplest method is to allow unsecured
   devices to be spoofed from any port on the network, as is the case
   today.

5.  Two or More Nodes Defending the Same Address



   All the previous sections of this document focused on the case where
   two nodes defend the same address (i.e., the node and the proxy).
   However, there are also cases where two or more nodes are defending
   the same address.  This is at least the case for:

   o  Nodes having the same address, as the Mobile Access Gateway's
      (MAG's) ingress link-local address in Proxy Mobile IPv6 (PMIPv6)
      [RFC5213].

   o  Nodes having a common anycast address [RFC4291].

   The problem statement, described previously in this document, applies
   for these cases, and the issues are the same from a signaling point
   of view.

   Multicast addresses are not mentioned here because Neighbor Discovery
   Protocol is not used for them.

   In the first case, [RFC5213] assumes that the security material used
   by SEND (i.e., public-private key pair) is shared between all the
   MAGs.  For the second case, there is no solution today.  But, in the
   same way, it should be possible to assume that the nodes having a
   common anycast address could also share the security material.






Combes, et al.                Informational                    [Page 18]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   It is important to notice that when many nodes defending the same
   address are not in the same administrative domain (e.g., MAGs in
   different administrative domains but in the same PMIPv6 domain
   [RFC5213]), sharing the security material used by SEND may raise a
   security issue.

6.  Security Considerations



6.1.  Router Trust Assumption



   Router-based authorization for Secured Proxy ND may occur without the
   knowledge or consent of a device.  It is susceptible to the 'Good
   Router Goes Bad' attack described in [RFC3756].

6.2.  Certificate Transport



   Certificate delegation relies upon transfer of the new credentials to
   the proxying HA in order to undertake ND proxy on its behalf.  Since
   the binding cannot come into effect until DAD has taken place, the
   delegation of the proxying authority necessarily predates the return
   of the Binding Ack, as described in [RFC3775].  In the case above
   described, the home tunnel that comes into creation as part of the
   binding process may be required for transport of Certificate Path
   Solicitations or Advertisements [RFC3971].  This constitutes a
   potential chicken-and-egg problem.  Either modifications to initial
   home binding semantics or certificate transport are required.  This
   may be trivial if certificates are sent in the clear between the MN's
   Care-of Address (CoA) and the HA without being tunneled.

6.3.  Timekeeping



   All of the presented methods rely on accurate timekeeping on the
   receiver nodes of Neighbor Discovery Timestamp Options.

   For router-authorized proxy ND, a Neighbor may not know that a
   particular ND message is replayed from the time when the proxied host
   was still on-link, since the message's timestamp falls within the
   valid timing window.  Where the router advertises its secured proxy
   NA, a subsequent replay of the old message will override the NCE
   created by the proxy.

   Creating the NCE in this way, without reference to accurate
   subsequent timing, may only be done once.  Otherwise, the receiver
   will notice that the timestamp of the advertisement is old or doesn't
   match.






Combes, et al.                Informational                    [Page 19]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   One way of creating a sequence of replayable messages that have
   timestamps likely to be accepted is to pretend to do an unsecured DAD
   on the address each second while the MN is at home.  The attacker
   saves each DAD defense in a sequence.  The granularity of SEND
   timestamp matching is around one second, so the attacker has a set of
   SEND NAs to advertise, starting at a particular timestamp, and valid
   for as many seconds as the original NA gathering occurred.

   This sequence may then be played against any host that doesn't have a
   timestamp history for that MN, by tracking the number of seconds
   elapsed since the initial transmission of the replayed NA to that
   victim, and replaying the appropriate cached NA.

   Where certificate-based authorization of ND proxy is in use, the
   origination/starting timestamp of the delegated authority may be used
   to override a replayed (non-proxy) SEND NA, while also ensuring that
   the Proxy NA's timestamp (provided by the proxy) is fresh.  A
   returning MN would advertise a more recent timestamp than the
   delegated authority and thus override it.  This method is therefore
   not subject to the above attack, since the proxy advertisement's
   certificate will have a timestamp greater than any replayed messages,
   preventing it from being overridden.

7.  Acknowledgments



   James Kempf and Dave Thaler particularly contributed to work on this
   document.  Contributions to discussion on this topic helped to
   develop this document.  The authors would also like to thank Jari
   Arkko, Vijay Devarapalli, Mohan Parthasarathy, Marcelo Bagnulo,
   Julien Laganier, Tony Cheneau, Michaela Vanderveen, Sean Shen, and
   Sheng Jiang for their comments and suggestions.

   Jean-Michel Combes is partly funded by MobiSEND, a research project
   supported by the French 'National Research Agency' (ANR).

8.  References



8.1.  Normative References



   [RFC3775]  Johnson, D., Perkins, C., and J. Arkko, "Mobility Support
              in IPv6", RFC 3775, June 2004.

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

   [RFC3972]  Aura, T., "Cryptographically Generated Addresses (CGA)",
              RFC 3972, March 2005.




Combes, et al.                Informational                    [Page 20]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


   [RFC4291]  Hinden, R. and S. Deering, "IP Version 6 Addressing
              Architecture", RFC 4291, February 2006.

   [RFC4306]  Kaufman, C., "Internet Key Exchange (IKEv2) Protocol",
              RFC 4306, December 2005.

   [RFC4389]  Thaler, D., Talwar, M., and C. Patel, "Neighbor Discovery
              Proxies (ND Proxy)", RFC 4389, April 2006.

   [RFC4861]  Narten, T., Nordmark, E., Simpson, W., and H. Soliman,
              "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861,
              September 2007.

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

8.2.  Informative References



   [RFC3756]  Nikander, P., Kempf, J., and E. Nordmark, "IPv6 Neighbor
              Discovery (ND) Trust Models and Threats", RFC 3756,
              May 2004.

   [RFC3963]  Devarapalli, V., Wakikawa, R., Petrescu, A., and P.
              Thubert, "Network Mobility (NEMO) Basic Support Protocol",
              RFC 3963, January 2005.

   [RFC5213]  Gundavelli, S., Leung, K., Devarapalli, V., Chowdhury, K.,
              and B. Patil, "Proxy Mobile IPv6", RFC 5213, August 2008.

   [RFC5280]  Cooper, D., Santesson, S., Farrell, S., Boeyen, S.,
              Housley, R., and W. Polk, "Internet X.509 Public Key
              Infrastructure Certificate and Certificate Revocation List
              (CRL) Profile", RFC 5280, May 2008.

   [RFC5380]  Soliman, H., Castelluccia, C., ElMalki, K., and L.
              Bellier, "Hierarchical Mobile IPv6 (HMIPv6) Mobility
              Management", RFC 5380, October 2008.

   [RFC5568]  Koodli, R., "Mobile IPv6 Fast Handovers", RFC 5568,
              July 2009.

   [RING]     Kempf, J. and C. Gentry, "Secure IPv6 Address Proxying
              using Multi-Key Cryptographically Generated Addresses
              (MCGAs)", Work in Progress, August 2005.







Combes, et al.                Informational                    [Page 21]

RFC 5909            SEND ND Proxy: Problem Statement           July 2010


Authors' Addresses



   Jean-Michel Combes
   France Telecom Orange
   38 rue du General Leclerc
   92794 Issy-les-Moulineaux Cedex 9
   France

   EMail: jeanmichel.combes@orange-ftgroup.com


   Suresh Krishnan
   Ericsson
   8400 Decarie Blvd.
   Town of Mount Royal
   QC Canada

   EMail: Suresh.Krishnan@ericsson.com


   Greg Daley
   Netstar Logicalis
   Level 6/616 St Kilda Road
   Melbourne, Victoria  3004
   Australia

   Phone: +61 401 772 770
   EMail: hoskuld@hotmail.com























Combes, et al.                Informational                    [Page 22]