Mobile IPv6 Location Privacy: IP Address Location Privacy and Mobile IPv6: Problem Statement

Network Working Group                                          R. Koodli
Request for Comments: 4882                        Nokia Siemens Networks
Category: Informational                                         May 2007


     IP Address Location Privacy and Mobile IPv6: Problem Statement

Status of This Memo

   This memo provides information for the Internet community.  It does
   not specify an Internet standard of any kind.  Distribution of this
   memo is unlimited.

Copyright Notice

   Copyright (C) The IETF Trust (2007).

Abstract

   In this document, we discuss location privacy as applicable to Mobile
   IPv6.  We document the concerns arising from revealing a Home Address
   to an onlooker and from disclosing a Care-of Address to a
   correspondent.

Table of Contents

   1. Introduction ....................................................2
   2. Definitions .....................................................3
   3. Problem Definition ..............................................4
      3.1. Disclosing the Care-of Address to the Correspondent Node ...4
      3.2. Revealing the Home Address to Onlookers ....................4
      3.3. Problem Scope ..............................................4
   4. Problem Illustration ............................................5
   5. Conclusion ......................................................7
   6. Security Considerations .........................................7
   7. Acknowledgments .................................................8
   8. References ......................................................8
      8.1. Normative References .......................................8
      8.2. Informative References .....................................8
   Appendix A. Background ............................................10

1.  Introduction

   The problems of location privacy, and privacy when using IP for
   communication, have become important.  IP privacy is broadly
   concerned with protecting user communication from unwittingly
   revealing information that could be used to analyze and gather
   sensitive user data.  Examples include gathering data at certain
   vantage points, collecting information related to specific traffic,
   and monitoring (perhaps) certain populations of users for activity
   during specific times of the day, etc.  In this document, we refer to
   this as the "profiling" problem.

   Location privacy is concerned with the problem of revealing roaming,
   which we define here as the process of a Mobile Node (MN) moving from
   one network to another with or without ongoing sessions.  A constant
   identifier with global scope can reveal roaming.  Examples are a
   device identifier such as an IP address, and a user identifier such
   as a SIP [RFC3261] URI [RFC3986].  Often, a binding between these two
   identifiers is available, e.g., through DNS [RFC1035].  Traffic
   analysis of such IP and Upper Layer Protocol identifiers on a single
   network can indicate device and user roaming.  Roaming could also be
   inferred by means of profiling constant fields in IP communication
   across multiple network movements.  For example, an Interface
   Identifier (IID) [RFC2462] in the IPv6 address that remains unchanged
   across networks could suggest roaming.  The Security Parameter Index
   (SPI) in the IPsec [RFC4301] header is another field that may be
   subject to such profiling and inference.  Inferring roaming in this
   way typically requires traffic analysis across multiple networks, or
   colluding attackers, or both.  When location privacy is compromised,
   it could lead to more targeted profiling of user communication.

   As can be seen, the location privacy problem spans multiple protocol
   layers.  Nevertheless, we can examine problems encountered by nodes
   using a particular protocol layer.  Roaming is particularly important
   to Mobile IP, which defines a global identifier (Home Address) that
   can reveal device roaming, and in conjunction with a corresponding
   user identifier (such as a SIP URI), can also reveal user roaming.
   Furthermore, a user may not wish to reveal roaming to
   correspondent(s), which translates to the use of a Care-of Address.
   As with a Home Address, the Care-of Address can also reveal the
   topological location of the Mobile Node.

   This document scopes the problem of location privacy for the Mobile
   IP protocol.  The primary goal is to prevent attackers on the path
   between the Mobile Node (MN) and the Correspondent Node (CN) from
   detecting roaming due to the disclosure of the Home Address.  The
   attackers are assumed to be able to observe, modify, and inject
   traffic at one point between the MN and the CN.  The attackers are
   assumed not to be able to observe at multiple points and correlate
   observations to detect roaming.  For this reason, MAC addresses,
   IIDs, and other fields that can be profiled to detect roaming are
   only in scope to the extent that they can be used by an attacker at
   one point.  Upper layer protocol identifiers that expose roaming are
   out of scope except that a solution to the problem described here
   needs to be usable as a building block in solutions to those
   problems.

   This document also considers the problem from the exposure of a
   Care-of Address to the CN.

   This document is only concerned with IP address location privacy in
   the context of Mobile IPv6.  It does not address the overall privacy
   problem.  For instance, it does not address privacy issues related to
   MAC addresses or the relationship of IP and MAC addresses [HADDAD],
   or the Upper Layer Protocol addresses.  Solutions to the problem
   specified here should provide protection against roaming disclosure
   due to using Mobile IPv6 addresses from a visited network.

   This document assumes that the reader is familiar with the basic
   operation of Mobile IPv6 [RFC3775] and the associated terminology
   defined therein.  For convenience, we provide some definitions below.

2.  Definitions

   o  Mobile Node (MN): A Mobile IPv6 Mobile Node that freely roams
      around networks

   o  Correspondent Node (CN): A Mobile IPv6 that node corresponds with
      an MN

   o  Home Network: The network where the MN is normally present when it
      is not roaming

   o  Visited Network: A network that an MN uses to access the Internet
      when it is roaming

   o  Home Agent: A router on the MN's home network that provides
      forwarding support when the MN is roaming

   o  Home Address (HoA): The MN's unicast IP address valid on its home
      network

   o  Care-of Address (CoA): The MN's unicast IP address valid on the
      visited network
   o  Reverse Tunneling or Bidirectional Tunneling: A mechanism used for
      packet forwarding between the MN and its Home Agent

   o  Route Optimization: A mechanism that allows direct routing of
      packets between a roaming MN and its CN, without having to
      traverse the home network

3.  Problem Definition

3.1.  Disclosing the Care-of Address to the Correspondent Node

   When a Mobile IP MN roams from its home network to a visited network
   or from one visited network to another, use of Care-of Address in
   communication with a correspondent reveals that the MN has roamed.
   This assumes that the correspondent is able to associate the Care-of
   Address to the Home Address, for instance, by inspecting the Binding
   Cache Entry.  The Home Address itself is assumed to have been
   obtained by whatever means (e.g., through DNS lookup).

3.2.  Revealing the Home Address to Onlookers

   When a Mobile IP MN roams from its home network to a visited network
   or from one visited network to another, use of a Home Address in
   communication reveals to an onlooker that the MN has roamed.  When a
   binding of a Home Address to a user identifier (such as a SIP URI) is
   available, the Home Address can be used to also determine that the
   user has roamed.  This problem is independent of whether the MN uses
   a Care-of Address to communicate directly with the correspondent
   (i.e., uses route optimization), or the MN communicates via the Home
   Agent (i.e., uses reverse tunneling).  Location privacy can be
   compromised when an onlooker is present on the MN - CN path (when
   route optimization is used).  It may also be compromised when the
   onlooker is present on the MN - HA path (when bidirectional tunneling
   without encryption is used; see below).

3.3.  Problem Scope

   With existing Mobile IPv6 solutions, there is some protection against
   location privacy.  If a Mobile Node uses reverse tunneling with
   Encapsulating Security Payload (ESP) encryption, then the Home
   Address is not revealed on the MN - HA path.  So, eavesdroppers on
   the MN - HA path cannot determine roaming.  They could, however,
   still profile fields in the ESP header; however, this problem is not
   specific to Mobile IPv6 location privacy.

   When an MN uses reverse tunneling (regardless of ESP encryption), the
   correspondent does not have access to the Care-of Address.  Hence, it
   cannot determine that the MN has roamed.

   Hence, the location privacy problem is particularly applicable when
   Mobile IPv6 route optimization is used or when reverse tunneling is
   used without protecting the inner IP packet containing the Home
   Address.

4.  Problem Illustration

   This section is intended to provide an illustration of the problem
   defined in the previous section.

   Consider a Mobile Node at its home network.  Whenever it is involved
   in IP communication, its correspondents can see an IP address valid
   on the home network.  Elaborating further, the users involved in
   peer-to-peer communication are likely to see a user-friendly
   identifier such as a SIP URI, and the communication endpoints in the
   IP stack will see IP addresses.  Users uninterested in or unaware of
   IP communication details will not see any difference when the MN
   acquires a new IP address.  Of course, any user can "tcpdump" or
   "ethereal" a session, capture IP packets, and map the MN's IP address
   to an approximate geo-location.  This mapping may reveal the home
   location of a user, but a correspondent cannot ascertain whether the
   user has actually roamed or not.  Assessing the physical location
   based on IP addresses has some similarities to assessing the
   geographical location based on the area code of a telephone number.
   The granularity of the physical area corresponding to an IP address
   can vary depending on how sophisticated the available tools are, how
   often an ISP conducts its network re-numbering, etc.  Also, an IP
   address cannot guarantee that a peer is at a certain geographic area
   due to technologies such as VPN and tunneling.

   When the MN roams to another network, the location privacy problem
   consists of two parts: revealing information to its correspondents
   and to onlookers.

   With its correspondents, the MN can either communicate directly or
   reverse-tunnel its packets through the Home Agent.  Using reverse
   tunneling does not reveal the Care-of Address of the MN, although
   end-to-end delay may vary depending on the particular scenario.  With
   those correspondents with which it can disclose its Care-of Address
   "on the wire", the MN has the option of using route-optimized
   communication.  The transport protocol still sees the Home Address
   with route optimization.  Unless the correspondent runs some packet
   capturing utility, the user cannot see which mode (reverse tunneling
   or route optimization) is being used, but knows that it is
   communicating with the same peer whose URI it knows.  This is similar
   to conversing with a roaming cellphone user whose phone number, like
   the URI, remains unchanged.

   Regardless of whether the MN uses route optimization or reverse
   tunneling (without ESP encryption), its Home Address is revealed in
   data packets.  When equipped with an ability to inspect packets "on
   the wire", an onlooker on the MN - HA path can determine that the MN
   has roamed and could possibly also determine that the user has
   roamed.  This could compromise the location privacy even if the MN
   took steps to hide its roaming information from a correspondent.

   The above description is valid regardless of whether a Home Address
   is statically allocated or is dynamically allocated.  In either case,
   the mapping of IP address to a geo-location will most likely yield
   results with the same level of granularity.  With the freely
   available tools on the Internet, this granularity is the physical
   address of the ISP or the organization that registers ownership of a
   prefix chunk.  Since an ISP or an organization is not, rightly,
   required to provide a blueprint of its subnets, the granularity
   remains fairly coarse for a mobile wireless network.  However,
   sophisticated attackers might be able to conduct site mapping and
   obtain more fine-grained subnet information.

   A compromise in location privacy could lead to more targeted
   profiling of user data.  An eavesdropper may specifically track the
   traffic containing the Home Address, and monitor the movement of the
   Mobile Node with a changing Care-of Address.  The profiling problem
   is not specific to Mobile IPv6, but could be triggered by a
   compromise in location privacy due to revealing the Home Address.  A
   correspondent may take advantage of the knowledge that a user has
   roamed when the Care-of Address is revealed, and modulate actions
   based on such knowledge.  Such information could cause concern to a
   mobile user, especially when the correspondent turns out be
   untrustworthy.  For these reasons, appropriate security measures on
   the management interfaces on the MN to guard against the disclosure
   or misuse of location information should be considered.

   Applying existing techniques to thwart profiling may have
   implications to Mobile IPv6 signaling performance.  For instance,
   changing the Care-of Address often would cause additional Return
   Routability [RFC3775] and binding management signaling.  And,
   changing the Home Address often has implications on IPsec security
   association management.  Careful consideration should be given to the
   signaling cost introduced by changing either the Care-of Address or
   the Home Address.

   When roaming, an MN may treat its home network nodes as any other
   correspondents.  Reverse tunneling is perhaps sufficient for home
   network communication, since route-optimized communication will
   traverse the identical path.  Hence, an MN can avoid revealing its
   Care-of Address to its home network correspondents simply by using
   reverse tunneling.  The Proxy Neighbor Advertisements [RFC2461] from
   the Home Agent could serve as hints to the home network nodes that
   the Mobile Node is away.  However, they will not be able to know the
   Mobile Node's current point of attachment unless the MN uses route
   optimization with them.

5.  Conclusion

   In this document, we have discussed the location privacy problem as
   applicable to Mobile IPv6.  The problem can be summarized as follows:
   disclosing the Care-of Address to a correspondent and revealing the
   Home Address to an onlooker can compromise the location privacy of a
   Mobile Node, and hence that of a user.  We have seen that
   bidirectional tunneling allows an MN to protect its Care-of Address
   to the CN.  And, ESP encryption of an inner IP packet allows the MN
   to protect its Home Address from the onlookers on the MN - HA path.
   However, with route optimization, the MN will reveal its Care-of
   Address to the CN.  Moreover, route optimization causes the Home
   Address to be revealed to onlookers in the data packets as well as in
   Mobile IPv6 signaling messages.  The solutions to this problem are
   expected to be protocol specifications that use the existing Mobile
   IPv6 functional entities, namely, the Mobile Node, its Home Agent,
   and the Correspondent Node.

6.  Security Considerations

   This document discusses the location privacy problem specific to
   Mobile IPv6.  Any solution must be able to protect (e.g., using
   encryption) the Home Address from disclosure to onlookers in data
   packets when using route optimization or reverse tunneling.  In
   addition, solutions must consider protecting the Mobile IPv6
   signaling messages from disclosing the Home Address along the MN - HA
   and MN - CN paths.

   Disclosing the Care-of Address is inevitable if an MN wishes to use
   route optimization.  Regardless of whether the Care-of Address is an
   on-link address of the MN on the visited link or that of a
   cooperating proxy, mere presence of a Binding Cache Entry is
   sufficient for a CN to ascertain roaming.  Hence, an MN concerned
   with location privacy should exercise prudence in determining whether
   to use route optimization with, especially previously unacquainted,
   correspondents.

   The solutions should also consider the use of temporary addresses and
   their implications on Mobile IPv6 signaling as discussed in Section
   4, "Problem Illustration".  Use of IP addresses with privacy
   extensions [RFC3041] could be especially useful for Care-of Addresses
   if appropriate trade-offs with Return Routability signaling are taken
   into account.

7.  Acknowledgments

   James Kempf, Qiu Ying, Sam Xia, and Lakshminath Dondeti are
   acknowledged for their review and feedback.  Thanks to Jari Arkko and
   Kilian Weniger for the last-call review and for suggesting
   improvements and text.  Thanks to Sam Hartman for providing text to
   improve the problem scope.

8.  References

8.1.  Normative References

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

8.2.  Informative References

   [HADDAD]   Haddad, W., et al., "Privacy for Mobile and Multi-homed
              Nodes: Problem Statement" Work in Progress, June 2006.

   [RFC1035]  Mockapetris, P., "Domain names - implementation and
              specification", STD 13, RFC 1035, November 1987.

   [RFC3986]  Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform
              Resource Identifier (URI): Generic Syntax", STD 66, RFC
              3986, January 2005.

   [RFC2461]  Narten, T., Nordmark, E., and W. Simpson, "Neighbor
              Discovery for IP Version 6 (IPv6)", RFC 2461, December
              1998.

   [RFC2462]  Thomson, S. and T. Narten, "IPv6 Stateless Address
              Autoconfiguration", RFC 2462, December 1998.


   [RFC3041]  Narten, T. and R. Draves, "Privacy Extensions for
              Stateless Address Autoconfiguration in IPv6", RFC 3041,
              January 2001.

   [RFC3261]  Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston,
              A., Peterson, J., Sparks, R., Handley, M., and E.
              Schooler, "SIP: Session Initiation Protocol", RFC 3261,
              June 2002.

   [RFC3825]  Polk, J., Schnizlein, J., and M. Linsner, "Dynamic Host
              Configuration Protocol Option for Coordinate-based
              Location Configuration Information", RFC 3825, July 2004.

   [RFC4301]  Kent, S. and K. Seo, "Security Architecture for the
              Internet Protocol", RFC 4301, December 2005.

Appendix A.  Background

   The location privacy topic is broad and often has different
   connotations.  It also spans multiple layers in the OSI reference
   model.  Besides, there are attributes beyond an IP address alone that
   can reveal hints about location.  For instance, even if a
   correspondent is communicating with the same endpoint it is used to,
   the "time of day" attribute can reveal a hint to the user.  Some
   roaming cellphone users may have noticed that their SMS messages
   carry a timestamp of their "home network" time zone (for location
   privacy or otherwise), which can reveal that the user is in a
   different time zone when messages are sent during a "normal" time of
   the day.  Furthermore, tools exist on the Internet that can map an IP
   address to the physical address of an ISP or the organization that
   owns the prefix chunk.  Taking this to another step, with built-in
   GPS receivers on IP hosts, applications can be devised to map geo-
   locations to IP network information.  Even without GPS receivers,
   geo-locations can also be obtained in environments where "Geopriv" is
   supported, for instance, as a DHCP option [RFC3825].  In summary, a
   user's physical location can be determined or guessed with some
   certainty and with varying levels of granularity by different means,
   even though IP addresses themselves do not inherently provide any
   geo-location information.  It is perhaps useful to bear this broad
   scope in mind as the problem of IP address location privacy in the
   presence of IP Mobility is addressed.

Author's Address

   Rajeev Koodli
   Nokia Siemens Networks
   313 Fairchild Drive
   Mountain View, CA 94043

   EMail: rajeev.koodli@nokia.com

Full Copyright Statement

   Copyright (C) The IETF Trust (2007).

   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.

   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, THE IETF TRUST 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.

Intellectual Property

   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.

Acknowledgement

   Funding for the RFC Editor function is currently provided by the
   Internet Society.