[Docs] [txt|pdf] [Tracker] [WG] [Email] [Diff1] [Diff2] [Nits]

Versions: (draft-huston-multi6-architectures) 00 01 02 03 04 RFC 4177

Individual Submission                                          G. Huston
Internet-Draft                                                     APNIC
Expires: April 24, 2005                                 October 24, 2004

           Architectural Approaches to Multi-Homing for IPv6

Status of this Memo

   This document is an Internet-Draft and is subject to all provisions
   of section 3 of RFC 3667.  By submitting this Internet-Draft, each
   author represents that any applicable patent or other IPR claims of
   which he or she is aware have been or will be disclosed, and any of
   which he or she become aware will be disclosed, in accordance with
   RFC 3668.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups.  Note that
   other groups may also distribute working documents as

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   The list of current Internet-Drafts can be accessed at

   The list of Internet-Draft Shadow Directories can be accessed at

   This Internet-Draft will expire on April 24, 2005.

Copyright Notice

   Copyright (C) The Internet Society (2004).


   This memo provides an analysis of the architectural aspects of
   multi-homing support for the IPv6 protocol suite.  The purpose of
   this analysis is to provide a taxonomy for classification of various
   proposed approaches to multi-homing.  It is also an objective of this
   exercise to identify common aspects of this domain of study, and also
   to provide a framework that can allow exploration of some of the
   further implications of various architectural extensions that are
   intended to support multi-homing.

Huston                   Expires April 24, 2005                 [Page 1]

Internet-Draft            Multi6 Architecture               October 2004

Document Revision Notes

   [RFC Editor: Please remove this section prior to publication.]

   The following changes have been made to the draft:


      Minor nits
         Added null IANA Considerations, section.  Minor grammatical
         correction to the abstract, Sections 4.2, 5,and 5.2 .

      Section 5.3.3 Layering Identity
         Additional text regarding requirement for additional
         information to be passed between transpot, intenret and
         identity protocol elements.


      Section 4.2: Multi-homing: Mobility
         New section added based on contribution from Marcelo Bagnulo.

      Section 5.2: Persistent, Opportunistic and Ephemeral Identities
         Additional text added about considerations if id/locator split
         in persistent identities and the requirements of multi-homing.

      Appendix A:Notes on Various approaches
         This section removed, to be replaced by a new WG document.


      Notes: IPv6
         Added text outlining the MIPv6 return routeability tests and
         the implications of this approach with multi-homing.

      Section 2: The Multi-Homing Space
         Added text on session initiation by the local host in the
         circumstance of degraded path connectivity.

      Section 4.2: Multi-homing: Identity Considerations
         Added text on session initiation by the local host in the
         circumstance of degraded path connectivity.  Also added text to
         cover the case that the remote host may need to discover the
         need to perform a locator switch for the multi-homed host in
         ways other than direct notification from the local end.

Huston                   Expires April 24, 2005                 [Page 2]

      Section 4.4: Multi-homing: Modified Protocol Element
         Change "single endpoint-to-endpoint session" to "single
         endpoint to single endpoint communication".

      Section 4.5: MModified Site-Exit and Host Behaviors
         Change NAT analysis reference to the multi6-threats draft.

      Section 5.1 Endpoint Identity Structure
         Added a qualification about unstructured identities and their
         utility as a resolution key.

      Section 5.3.2 Locator Selection
         Added this section which describes the considerations of
         traffic engineering in the context of locator selection.

      Section 5.3.3 Layering Identity
         Added qualification about use of transport (session) identities
         for UDP.

      Section 6.1 Establishing Session State
         Qualified the use of "transport" to be "identity protocol
         element", indicating that this may be transport, IP of a wedge
         layer, and edited the text to reflect multi-homing capabilities
         in the protocol stack.  Added text on locator discovery and
         selection in the functional decomposition of session


      Section 2: The Multi-Homing Space
         Added text to include consideration of session initiation in
         the face of changes to the connectivity topology, and a note
         about the potential to consider traffic engineering across
         multiple paths.

      Section 3: Functional Goals and Considerations
         Changed 'requirements' to 'goals'.

      Section 5.1 Endpoint Identity Structure
         Added consideration of disambiguating locators and identities
         when identities are drawn from the same address space as
         locators.  Added text about identities drawn from PA space and
         the problems this raises.  Also added text about disambiguating
         DNS FQDN pseudo-anycast from DNS-based multi-homing with
         equivalent locator sets.

Huston                   Expires April 24, 2005                 [Page 3]

Internet-Draft            Multi6 Architecture               October 2004

      Section 5.2 Persistent, Opportunistic and Ephemeral Identities
         New section added to the draft considering the implications of
         these three approaches to identity.

      Section 5.3.1 Triggering Locator Switches
         Added section on ICMP triggers.

      Section 5.3.3 Layering Identity
         New section added, considering the implications of placing
         endpoint identity functionality in the transport or intenetwork
         protocol elements, or as a wedge element, conceptually placed
         between these two elements.

      Section 6. Functional Decomposition of Multi-Homing Approaches
         New section added.

Huston                   Expires April 24, 2005                 [Page 4]

Internet-Draft            Multi6 Architecture               October 2004

Table of Contents

   1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . .  6
   2.  The Multi-Homing Space . . . . . . . . . . . . . . . . . . . .  6
   3.  Functional Goals and Considerations  . . . . . . . . . . . . .  8
   4.  Approaches to Multi-Homing . . . . . . . . . . . . . . . . . .  9
     4.1   Multi-Homing: Routing  . . . . . . . . . . . . . . . . . . 10
     4.2   Multi-Homing: Mobility . . . . . . . . . . . . . . . . . . 11
     4.3   Multi-homing: Identity Considerations  . . . . . . . . . . 13
     4.4   Multi-homing: Identity Protocol Element  . . . . . . . . . 15
     4.5   Multi-homing: Modified Protocol Element  . . . . . . . . . 16
     4.6   Modified Site-Exit and Host Behaviors  . . . . . . . . . . 17
   5.  Approaches to Endpoint Identity  . . . . . . . . . . . . . . . 18
     5.1   Endpoint Identity Structure  . . . . . . . . . . . . . . . 19
     5.2   Persistent, Opportunistic and Ephemeral Identities . . . . 22
     5.3   Common Issues for Multi-Homing Approaches  . . . . . . . . 25
       5.3.1   Triggering Locator Switches  . . . . . . . . . . . . . 25
       5.3.2   Locator Selection  . . . . . . . . . . . . . . . . . . 26
       5.3.3   Layering Identity  . . . . . . . . . . . . . . . . . . 27
       5.3.4   Session Startup and Maintenance  . . . . . . . . . . . 29
       5.3.5   Dynamic Capability Negotiation . . . . . . . . . . . . 31
       5.3.6   Identity Uniqueness and Stability  . . . . . . . . . . 31
   6.  Functional Decomposition of Multi-Homing Approaches  . . . . . 32
     6.1   Establishing Session State . . . . . . . . . . . . . . . . 32
     6.2   Rehoming Triggers  . . . . . . . . . . . . . . . . . . . . 33
     6.3   Rehoming Locator Pair Selection  . . . . . . . . . . . . . 33
     6.4   Locator Change . . . . . . . . . . . . . . . . . . . . . . 33
     6.5   Removal of Session State . . . . . . . . . . . . . . . . . 33
   7.  IANA Considerations  . . . . . . . . . . . . . . . . . . . . . 34
   8.  Security Considerations  . . . . . . . . . . . . . . . . . . . 34
   9.  Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 34
   10.   Informative References . . . . . . . . . . . . . . . . . . . 34
       Author's Address . . . . . . . . . . . . . . . . . . . . . . . 35
       Intellectual Property and Copyright Statements . . . . . . . . 36

Huston                   Expires April 24, 2005                 [Page 5]

Internet-Draft            Multi6 Architecture               October 2004

1.  Introduction

   The objective of this analysis is to allow various technical
   proposals relating to the support of multi-homing environment in IPv6
   to be placed within an architectural taxonomy.  This is intended to
   allow these proposals to be classified and compared in a structured
   fashion.  It is also an objective of this exercise to identify common
   aspects across all proposals within this domain of study, and also to
   provide a framework that can allow exploration of some of the further
   implications of various architectural extensions that are intended to
   support multi-homing.  The scope of this study is limited to the IPv6
   protocol suite architecture, although reference is made to IPv4
   approaches as required.

2.  The Multi-Homing Space

   A simple formulation of the multi-homing environment is indicated in
   Figure 1.

Huston                   Expires April 24, 2005                 [Page 6]

Internet-Draft            Multi6 Architecture               October 2004

                           | host |
                           |  R   |
                    + - - - - - - - - - - - +
                    | Internet Connectivity |
                    + - - - - - - - - - - - +
                         /            \
                   +---------+    +---------+
                   | ISP A   |    |  ISP B  |
                   +---------+    +---------+
                       | Path A        | Path B
         + - - - - - - - - - - - - - - - - - - - - +
         | multi-      |               |           |
           homed   +------+         +------+
         | site    | site |         | site |       |
                   | exit |         | exit |
         |         |router|         |router|       |
                   |  A   |         |  B   |
         |         +------+         +------+       |
                      |                |
         |         local site connectivity         |
         |           +-----------+                 |
         |           |   host    |                 |
         + - - - - - - - - - - - - - - - - - - - - +

   The Multi-Homed Domain

                                Figure 1

   The environment of multi-homing is one that is intended to provide
   sufficient support to local hosts so as to allow local hosts to
   exchange IP packets with remote hosts, such that this exchange of
   packets is to be seamlessly supported across dynamic changes in
   connectivity.  Session resilience implies that if a local
   multi-homed-aware host establishes an application session with the
   remote host using "Path A", and this path fails, the application
   session should be mapped across to "Path B" without requiring any
   application-visible re-establishment of the session.  In other words,
   the application session should not be required to be explicitly aware
   of underlying path changes at the level of packet forwarding paths
   chosen by the network.

Huston                   Expires April 24, 2005                 [Page 7]

Internet-Draft            Multi6 Architecture               October 2004

   In addition to this objective of session resilience across network
   reachability changes, there are also considerations of providing
   mechanisms to support site visibility in the face of dynamic changes
   in external reachability.  Sustained site visibility implies that
   external attempts to initiate a communication with hosts within the
   site will succeed as long as there is at least one viable path
   between the external host and the multi-homed site.  It also implies
   that local attempts to initiate a communication with remote hosts
   should take into account the current connectivity state in
   undertaking locator selection and setting up initial locator sets.

   In addition there is the potential consideration of being able to
   distribute traffic load across a number of network paths according to
   some pre-determined objective, as a form of traffic engineering.

   This simple multi-homing scenario also includes "site-exit' routers,
   where the local site interfaces to the upstream Internet transit
   providers.  The nature of the interactions between the external
   routing system and the site-exit routers, and interactions between
   the site-exit routers and the local multi-homed host, and the
   interactions between local connectivity forwarding and the local host
   and site exit routers are not defined a priori in this scenario, as
   they form part of the framework of interaction between the various
   multi-homing components.

   The major characteristic of this scenario is that the address space
   used by, and advertised as reachable by, ISP A is distinct from the
   address space used by ISP B.

   This simple scenario is intended to illustrate the basic multi-homing
   environment.  Variations of this scenario include additional external
   providers of transit connectivity to the local site, complex site
   requirements and constraints, where the site may not interface
   uniformly to all external transit providers, sequential rather than
   simultaneous external transit reachability, communication with remote
   multi-homed hosts, multi-way communications, use of host addresses in
   a referential context (third party referrals) and the imposition of
   policy constraints on path selection.  However, the basic scenario is
   sufficient to illustrate the major architectural aspects of support
   for multi-homing, so this scenario will be used as the reference
   model for this analysis.

3.  Functional Goals and Considerations

   RFC 3582 [1] documents some goals that a multi-homing approach should
   attempt to address.  These goals include:
   o  redundancy

Huston                   Expires April 24, 2005                 [Page 8]

Internet-Draft            Multi6 Architecture               October 2004

   o  load sharing
   o  traffic engineering
   o  policy constraints
   o  simplicity of approach
   o  transport-layer survivability
   o  DNS compatibility
   o  packet filtering capability
   o  scalability
   o  legacy compatibility
   The reader is referred to [1] for a complete description of each of
   these goals.

   In addition, [2] documents further considerations for IPv6
   multi-homing.  Again, the reader is referred to this document for the
   detailed enumeration of these considerations.  The general topic
   areas considered in this study include:
   o  interaction with routing systems,
   o  aspects of a split between end-point-identifier and forwarding
   o  changes to packets on the wire, and
   o  the interaction between names, endpoints and the DNS.

   In evaluating various approaches, further consideration also include:
   o  the role of helpers and agents in the approach,
   o  modifications to host behaviors,
   o  the required trust model to support the interactions, and
   o  the nature of potential vulnerabilities in the approach.

4.  Approaches to Multi-Homing

   There appear to be five generic forms of architectural approaches to
   this problem, namely:
   o  Routing
      Use the IPv4 multi-homing approach
   o  Mobility
      Use the IPv6 Mobility approach
   o  New Protocol Element
      Insertion of a new element in the protocol stack that manages a
      persistent identity for the session
   o  Modify a Protocol Element
      Modify the Transport or IP protocol stack element in the host in
      order to support dynamic forwarding locator change
   o  Modified Site-Exit Router / Local Host interaction
      Modify the site-exit router and local forwarding system to allow
      various behaviors including source-based forwarding, site-exit
      hand-offs, and address rewriting by site-exit routers

   These approaches will be described in detail in the following

Huston                   Expires April 24, 2005                 [Page 9]

Internet-Draft            Multi6 Architecture               October 2004


4.1  Multi-Homing: Routing

   The approach used in IPv4 for multi-homing support is to preserve the
   semantics of the IPv4 address as both an endpoint identifier and a
   forwarding locator.  For this to work in a multi-homing context it is
   necessary for the transit ISPs to announce the local site's address
   prefix as a distinct routing entry in the inter-domain routing
   system.  This approach could be used in an IPv6 context, and, as with
   IPv4, no modifications to the IPv6 architecture are required to
   support this approach.

   The local site's address prefix may be a more specific address prefix
   drawn from the address space advertised by one of the transit
   providers, or from some third party provider not current directly
   connected to the local site.  Alternatively the address space may be
   a distinct address block obtained by direct assignment from a
   Regional Internet Registry as Provider Independent space.  Each host
   within the local site is uniquely addressed from the site's address

   All transit providers for the site accept a prefix advertisement from
   the multi-homed site, and advertise this prefix globally in the
   inter-domain routing table.  When connectivity between the local site
   and an individual transit provider is lost, normal operation of the
   routing protocol will ensure that the routing advertisement
   corresponding to this particular path will be withdrawn from the
   routing system, and those remote domain domains who had selected this
   path as the best available will select another candidate path as the
   best path.  Upon restoration of the path, the path is re-advertised
   in the inter-domain routing system.  Remote domains will undertake a
   further selection of the best path based on this re-advertised
   reachability information.  Neither the local or the remote host need
   to have multiple addresses, nor undertake any form of address
   selection.  The path chosen for forward and reverse direction path
   flows is a decision made by the routing system.

   This approach generally meets all the goals for multi-homing
   approaches with one notable exception: scalability.  Each site that
   multi-homes in this fashion adds a further entry in the global
   inter-domain routing table.  Within the constraints of current
   routing and forwarding technologies it is not clearly evident that
   this approach can scale to encompass a population of multi-homed
   sites of the order of 10**7 such sites.  The implication here is that
   this would add a similar number of unique prefixes into the
   inter-domain routing environment, which in turn would add to the
   storage and computational load imposed on inter-domain routing

Huston                   Expires April 24, 2005                [Page 10]

Internet-Draft            Multi6 Architecture               October 2004

   elements within the network.  This scale of additional load is not
   supportable within the current capabilities of the IPv4 global
   Internet, nor is it clear at present that the routing capabilities of
   the entire network could be expanded to manage this load in a
   cost-effective fashion, within the bounds of the current inter-
   domain routing protocol architecture.

4.2  Multi-Homing: Mobility

   Preserving established communications through movement is similar to
   preserving established communications through outages in multihomed
   sites as both scenarios require the capability of dynamically
   changing the locators used during the communication while
   maintaining, unchanged, the endpoint identifier used by ULP.  Since
   MIPv6 protocol [3] already provides the required support to preserve
   established communications through movement, it seems worthwhile to
   explore whether it could also be used to provide session
   survivability in multihomed environments.

   MIPv6 uses a preferred IP address, the Home Address (HoA), as a
   stable identifier for the mobile node (MN).  This identifier is then
   dynamically mapped to a valid locator (Care-of Address, or CoA) that
   corresponds to the current attachment point within the network
   topology.  When the MN is at the Home Network, the HoA is used both
   as locator and as identifier.  When the MN is not at the Home
   Network, the HoA is used as an identifier, and the CoA is used as
   locator.  A relaying agent (Home Agent) placed in the Home Network is
   used to forward packets addressed to the HoA to the current location,
   specified by the CoA.  After each movement, the MN must inform its
   Home Agent of the new CoA, and optionally inform those entities with
   which it has established communications with (Correspondent Nodes, or
   CNs).  The mapping information between the HoA and the current CoA is
   conveyed using Binding Update (BU) messages.  When the BU message is
   exchanged between the MN and the Home Agent, it is possible to assume
   the existence of a pre-established Security Association that can be
   used to protect the binding information.  However, when the BU
   message is exchanged between the MN and the CN, it is not possible to
   assume the existence of such Security Association.  In this case, it
   is necessary to adopt an alternative mechanism to protect the binding
   information contained in the message.  The selected mechanism is
   called Return Routability procedure and the background for its design
   is detailed in [4].  The goal of the mechanism is to allow the CN to
   verify that the MN that is claiming that a HoA is currently located
   at a CoA is entitled to make such claim, which essentially means that
   the HoA was assigned to the MN, and that the MN is currently located
   at the CoA.  In order to verify these updates the CN sends two
   different secrets, one to the claimed HoA and another one to the
   claimed CoA.  If the MN receives both secrets, this means that the

Huston                   Expires April 24, 2005                [Page 11]

Internet-Draft            Multi6 Architecture               October 2004

   Home Agent located at the Home Network has a trust relationship with
   the MN and it has forwarded the secret sent to the HoA, and that the
   MN is receiving packets sent to the CoA.  So, by including
   authorization information derived from both secrets within the BU
   message, the MN will be able to prove to the CN that the claimed
   binding between the HoA and the CoA is valid.

   The lifetime of the binding that is created in the CN using
   authorization information obtained through the Return Routability
   procedure is limited to 7 minutes, in order to prevent time-shifted
   attacks [4].  In a time-shifted attack, an attacker located along the
   path between the CN and the MN forges the Return Routability packet
   exchange.  The result of such attack is that the CN will forward all
   the traffic addressed to the HoA to the CoA selected by the attacker.
   The attacker can then leave the position along the path but the
   effects of the attack will remain until the binding is deleted,
   shifting in time the effect of the attack.  By limiting the lifetime
   of the binding in the CN, the effect of this attack is reduced to 7
   minutes, because after that period a new Return Routability procedure
   is needed to extend the binding lifetime.  It should be noted that
   the Return Routability procedure is vulnerable to Man-In-The-Middle
   attacks, since an attacker located along the path between the CN and
   the MN can forge the periodic Return Routability packet exchange.

   The possible application of the MIPv6 protocol to the multihoming
   problem would be to use BU messages to convey information about
   alternative addresses to be used after the outage.

   In this scenario, the multihomed host adopts the MN role and the host
   outside the multihomed site adopts the CN role.  When a communication
   is established between the multihomed host and the external host, the
   address used for initiating the communication is used as a HoA.  The
   communication continues using this address as long as no outage
   occurs.  If an outage occurs and the HoA becomes unreachable, an
   alternative address of the multihomed node is used as a CoA.  In this
   case, the multihomed node sends a BU message to the external host,
   informing about the new CoA to be used for the HoA, so that the
   established communication can be preserved using the alternative
   address.  However, such BU message has to be validated using
   authorization information obtained through the Return Routability
   procedure, which implies that the binding lifetime will be limited to
   a fixed period of no more than 7 minutes.  The result is that the
   binding between the HoA and the new CoA will expire after this
   interval has elapsed, and then the HoA will be used for the
   communication.  Since the HoA is unreachable because of the outage,
   the communication will be interrupted.  It should be noted that it is
   not possible to acquire new authorization information by performing a
   new Return Routability procedure, because it requires communication

Huston                   Expires April 24, 2005                [Page 12]

Internet-Draft            Multi6 Architecture               October 2004

   through the HoA, which is no longer reachable.  Consequently, a
   mechanism based on the MIPv6 BU messages to convey information about
   alternative addresses will preserve communications only for 7

   The aspect of MIPv6 which appears to present issues in the context of
   multi-homing is the return routeability mechanism.  In MIPv6 identity
   validity is periodically tested by return routeability of the
   identity address.  This regular use of a distinguished locator as the
   identity token cannot support return reachability in the multi-homing
   context in the event of extended path failure of the path that is
   associated with the identity locator.

4.3  Multi-homing: Identity Considerations

   The intent of multi-homing in the IPv6 domain is to achieve a
   comparable functional outcome for multi-homed sites without an
   associated additional load being imposed on the routing system.  The
   overall intent of IPv6 is to provide a scalable protocol framework to
   support the deployment of communications services for an extended
   period of time, and this implies that the scaling properties of the
   deployment environment remain tractable within projections of size of
   deployment and underlying technology capabilities.  Within the
   inter-domain routing space, the basic approach used in IPv4 and IPv6
   is to attempt to align address deployment with network topology, so
   that address aggregation can be used to create a structured hierarchy
   of the routing space.

   Within this constraint of topological-based address deployment and
   provider aggregatable addressing architectures, the local site that
   is connected to multiple providers is delegated addresses from each
   of these providers' address blocks.  In the example network in Figure
   1, the local multi-homed host will conceivably be addressed in two
   ways: one using transit provider A's address prefix and the other
   using transit provider B's address prefix.

   If remote host R is to initiate a communication with the local
   multi-homed host, it would normally query the DNS for an address for
   the local host.  In this context the DNS would return 2 addresses
   (One using the A prefix and the other using the B prefix).  The
   remote host would select one of these addresses and send a packet to
   this destination address.  This would direct the packet to the local
   host along a path through A or B, depending on the selected address.
   If the path between the local site and the transit provider fails,
   then the address prefix announced by the transit provider to the
   inter-domain routing system will continue to be the provider's
   address prefix.  The remote host will not see any change in routing,
   yet packets sent to the local host will now fail to be delivered.

Huston                   Expires April 24, 2005                [Page 13]

Internet-Draft            Multi6 Architecture               October 2004

   The question posed by the multi-homing problem is: "If the remote
   host is aware of multi-homing, how could it switch over to using the
   equivalent address for the local multi-homed host that transits the
   other provider?"

   If the local multi-homed host wishes to initiate a session with
   remote host R, it needs to send a packet to R with a valid source and
   destination address.  While the destination address is that of R,
   what source address should the local host use? There are two
   implications for this choice.  Firstly the remote host will, by
   default use this source address as the destination address in its
   response, and hence this choice of source address will direct the
   reverse path from R to the local host.  Secondly, the ISPs A and B
   may be using some form of reverse unicast address filtering on source
   addresses of packets passed to the ISP, as a means of prevention of
   source address spoofing.  This implies that if the multi-homed
   address selects a source address from address prefix A, and the local
   routing to R selects a best path via ISP B, then ISP B's ingress
   filters will discard the packet.

   Within this addressing structure there is no form of routing-based
   repair of certain network failures.  If the link between the local
   site and ISP A fails, there is no change in the route advertisements
   made by ISP A to its external routing peers.  Even though the multi-
   homed site continues to be reachable via ISP B, packets directed to
   the site using ISP A's prefix will be discarded by ISP A as the
   destination is unreachable.  The implication here is that if the
   local host wishes to maintain a session across such events it needs
   to communicate to remote host R that it is possible to switch to
   using a destination address for the multi-homed host that is based on
   ISP B' address prefix.  In the event that the local host wishes to
   initiate a session at this point, then it may need to use an initial
   source locator that reflects the situation that the only viable
   destination address to use the one that is based on ISP B's address
   prefix.  It may be the case that the local host is not always aware
   of this return routeability constraint, or it may not be able to
   communicate this information directly to R, in which case R needs to
   discover or be passed this information in other ways.

   In an aggregated routing environment multiple transit paths to a host
   imply multiple address prefixes for the host, where each possible
   transit path is identified by an address for the host.  The
   implication of this constraint on multi-homing is that paths being
   passed to the local multi-homed site via transit provider ISP A must
   use a forwarding-level destination IP address drawn from ISP A's
   advertised address prefix set that maps to the multi-homed host.
   Equally, packets being passed via the transit of ISP B must use a
   destination address drawn from ISP B's address prefix set.  The

Huston                   Expires April 24, 2005                [Page 14]

Internet-Draft            Multi6 Architecture               October 2004

   further implication here is that path selection (ISP A vs ISP B
   transit for incoming packets) is an outcome of the process of
   selecting an address for the destination host.

   The architectural consideration here is that in the conventional IP
   protocol architecture the assumption is made that the transport-layer
   endpoint identity is the same identity used by the internet-layer
   forwarding layer, namely the IP address.

   If multiple forwarding paths are to be supported for a single
   transport session, and path selection is to be decoupled from the
   functions of transport session initiation and maintenance, then the
   corollary of this requirement in architectural terms appears to be
   that some changes are required in the protocol architecture to
   decouple the concepts of identification of the endpoint and
   identification of the location and associated path selection for the
   endpoint.  This is a fundamental change in the semantics of an IP
   address in the context of the role of the endpoint address within the
   end-to-end architectural model [5].  This change in the protocol
   architecture would permit a transport session to use an invariant
   endpoint identity value to initiate and maintain a session, while
   allowing the forwarding layer to dynamically change paths and
   associated endpoint locator identities without impacting on the
   operation of the session, nor would such a decoupled concept of
   identities and locators add any incremental load to the inter-domain
   routing system.

   Some generic approaches to this form of separation of endpoint
   identity and locator value are described in the following sections.

4.4  Multi-homing: Identity Protocol Element

   One approach to this objective is to add a new element into the model
   of the protocol stack.

   The presentation to the upper level protocol stack element (ULP)
   would use endpoint identifiers to uniquely identify both the local
   stack and the remote stack.  This will provide the ULP with stable
   identifiers for the duration of the ULP session.

   The presentation to the lower level protocol stack element (LLP)
   would be of the form of a locator.  This implies that the protocol
   stack element would need to maintain a mapping of endpoint identifier
   values to locator values.  In a multi-homing context one of the
   essential characteristics of this mapping is that it needs to be
   dynamic, in that environmental triggers should be able to trigger a
   change in mappings, which in turn would correspond to a change in the
   paths (forward and/or reverse) used by the endpoints to traverse the

Huston                   Expires April 24, 2005                [Page 15]

Internet-Draft            Multi6 Architecture               October 2004

   network.  In this way the ULP session is defined by a peering of
   endpoint identifiers that remain constant throughout the lifetime of
   the ULP session, while the locators may change to maintain end-to-end
   reachability for the session.

   The operation of the new protocol stack element (termed here the
   "endpoint identity protocol stack element", or "EIP") is to establish
   a synchronized state with its remote counterpart.  This would allow
   the stack elements to exchange a set of locators that may be used
   within the context of the session.  A change in the local binding
   between the current endpoint identity value and a locator will cause
   a change in the source locator value used in the forwarding level
   packet header.  The actions of the remote EIP upon receipt of this
   packet with the new locator is to firstly recognize this locator as
   part of an existing session, and, upon some trigger condition, to
   change its session view of the mapping of the remote endpoint
   identity to the corresponding locator, and use this locator as the
   destination locator in subsequent packets passed to the LLP.

   From the perspective of the IP protocol architecture there are two
   possible locations to insert the EIP into the protocol stack.

   One possible location is at the upper level of the transport
   protocol.  Here the application program interface (API) of the
   application level protocols would interface to the EIP element, and
   use endpoint identifiers to refer to the remote entity.  The EIP
   would pass locators to the API of the transport layer.

   The second approach is to insert the EIP between the transport and
   internet protocol stack elements, so that the transport layer would
   function using endpoint identifiers, and maintain a transport session
   using these endpoint identifiers.  The IP or internetwork layer would
   function using locators, and the mapping from endpoint identifier to
   locator is undertaken within the EIP stack element.

4.5  Multi-homing: Modified Protocol Element

   As an alternative to insertion of a new protocol stack element into
   the protocol architecture, an alternative approach is to modify an
   existing protocol stack element to include the functionality
   performed by the EIP element.  This modification could be undertaken
   within the transport protocol stack element, or within the
   internetworking stack element.  The functional outcome from these
   modifications would be to create a mechanism to support the use of
   multiple locators within the context of a single endpoint to single
   endpoint communication.

   Within the transport layer, this functionality can be achieved, for

Huston                   Expires April 24, 2005                [Page 16]

Internet-Draft            Multi6 Architecture               October 2004

   example, by the binding of a set of locators to a single session, and
   then communicating this locator set to the remote transport entity.
   This would allow the local transport entity to switch the mapping to
   a different locator for either the local endpoint or the remote
   endpoint while maintaining the integrity of the ULP session.

   Within the IP level this functionality could be supported by a form
   of dynamic rewriting of the packet header as it is processed by the
   protocol element.  Incoming packets with the source and destination
   locators in the packet header are mapped to packets with the
   equivalent endpoint identifiers in both fields, and the reverse
   mapping is performed to outgoing packets passed from the transport
   layer.  Mechanisms that support direct rewriting of the packet header
   are potential candidates in this approach, as are various forms of
   packet header transformations of encapsulation, where the original
   endpoint identifier packet header is preserved in the packet and an
   outer level locator packet header is wrapped around the packet as it
   is passed through the internetworking protocol stack element.

   In all these scenarios, there are common issues of what state is
   kept, by which part of the protocol stack, how state is maintained
   with additions, removals of locator bindings, and does only one piece
   of code have to be aware of the endpoint / locator split or do
   multiple protocol elements have to be modified? For example, if the
   functionality is added at the internetworking (IP) layer, there is no
   context of an active transport session, so that removal of identity /
   locator state information for terminated sessions needs to be
   triggered by some additional mechanism from the transport layer to
   the internetworking layer.

4.6  Modified Site-Exit and Host Behaviors

   The above approaches all assume that the hosts are explicitly aware
   of the multi-homed environment and use modified protocol behavior to
   support multi-homing functionality.  A further approach to this
   objective is to split this functionality across a number of network
   elements and potentially perform packet header rewriting from a
   persistent endpoint identity value to a locator value at a remote

   One possible approach proposes the use of site-exit routers to
   perform some form of packet header manipulation as packets are passed
   out from the local multi-homed site to a particular transit provider.
   The local site routing system will select the best path to a
   destination host based on the remote hosts's locator value.  The
   local host will write its endpoint identity as the source address of
   the packet.  When the packet reaches a site-exit router, the
   site-exit router will rewrite the source field of the packet to a

Huston                   Expires April 24, 2005                [Page 17]

Internet-Draft            Multi6 Architecture               October 2004

   corresponding locator that selects a reverse path through the same
   transit ISP when the locator is used as a destination locator by the
   remote host.  In order to preserve session integrity there is a need
   for a corresponding reverse transformation to be undertaken on
   incoming packets, where the destination locator has to be mapped back
   to the host's endpoint identifier.  There are a number of
   considerations whether this is best performed at the site exit router
   on packet ingress to the site, or by the local host.

   Packet header rewriting by remote network elements has a large number
   of associated security considerations, and any packet rewriting
   mechanism has to provide proper protection against the attacks
   described in [6], in particular against redirection attacks.

   An alternative for packet header rewriting on site exit is for the
   host to undertake the endpoint-to-locator mapping, using one of the
   approaches outlined above.  The consideration here is that there is
   some significant deployment of unicast reverse path filtering in
   Internet environments as a counter-measure to source address
   spoofing.  Using the example in Figure 1, if a host selects a locator
   drawn from the ISP B address prefix, and local routing directs that
   packet to site-exit router A, then if the packet is passed to ISP A,
   the this would be discarded by such filters.  Various approaches have
   been proposed to modify the behavior of the site forwarding
   environment all with the end effect that packets using a source
   locator drawn from the ISP B address prefix are passed to site-exit
   router B.  These approaches include forms of source address routing
   and site-exit router hand-over mechanisms, as well as augmentation of
   the routing information between site-exit routers and local
   multi-homed hosts, so that the choice of locator by the local host
   for the remote host is consistent with the current local routing
   state for the local site to reach the remote host.

5.  Approaches to Endpoint Identity

   Both the approach of the addition of an identity protocol element and
   the approach of modification of an existing protocol element assume
   some form of exchange of information that allows both parties to the
   communication to be aware of the other party's endpoint identity and
   the associated mapping to locators.  There are a number of choices in
   terms of the way in which this information exchange can be

   The first such possible approach is termed here a 'conventional'
   approach, where the mode of operation is in terms of encapsulating
   the protocol data unit (PDU) passed from the ULP with additional data
   elements that specifically refer to the function of the endpoint
   identity protocol stack element.  The compound data element is passed

Huston                   Expires April 24, 2005                [Page 18]

Internet-Draft            Multi6 Architecture               October 2004

   to the LLP as its PDU.  The corresponding actions on receipt of a PDU
   from a LLP is to extract the fields of the data unit that correspond
   to the EIP function, and pass the reminder of the PDU to the ULP.
   The EIP operates in an "in-band" mode, communicating with its remote
   peer entity through additional information wrapped around the ULP
   PDU.  This is equivalent to generic tunneling approaches where the
   outer encapsulation of the transmitted packet contains location
   address information, while the next level packet header contains
   information that is to be exposed and used at the location endpoints,
   being, in this case, identity information.

   Another approach is to allow the EIP to communicate using a separate
   communications channel, where the EIP generates dedicated messages
   that are directed to its peer EIP, and passes these PDUs to the LLP
   independently of the PDUs that are passed to the EIP from the ULP.
   This allows the EIP to exchange information and synchronize state
   with the remote EIP semi-independently of the ULP protocol exchange.
   As a part of the EIP function is to transform the ULP PDU to include
   locator information there is an associated requirement to ensure that
   the EIP peering state remains synchronized to the exchange of ULP
   PDUs, so that the remote EIP can correctly recognize the locator to
   endpoint mapping for each active session.

   Another potential approach here is to allow the endpoint to locator
   mappings to be held at a third party point.  This model is already
   used for supporting the name to IP address mappings performed by the
   Domain Name system, where the mapping is obtained by reference to a
   third party, namely a DNS resolver.  A similar form of third party
   mapping between endpoints and a locator set could be supported
   through the use of the DNS, or a similar third party referential
   mechanism.  Rather than have each party exchange endpoint to locator
   mappings, this approach would see this mapping being obtained as a
   result of a lookup for a DNS Endpoint to Locator set map contained as
   DNS Resource Records, for example.

5.1  Endpoint Identity Structure

   The previous section has used the term "endpoint identity" without
   examining what form this identity may take.  There are a number of
   salient considerations regarding the structure and form of this
   identity that should be enumerated within an architectural overview
   of this space.

   One possible form of an identity is the use of identity tokens lifted
   from the underlying protocol's "address space".  In other words an
   endpoint identity is a special case instance of an IPv6 protocol
   address.  There are a number of advantages in using this form of
   endpoint identity, observing that the suite of IP protocols and

Huston                   Expires April 24, 2005                [Page 19]

Internet-Draft            Multi6 Architecture               October 2004

   associated applications already manipulate IP addresses.  The
   essential difference in a domain that distinguishes between endpoint
   identity and locator is that the endpoint identity parts of the
   protocol would operate on those addresses that assume the role of
   endpoint identities, and the endpoint identity / locator mapping
   function would undertake a mapping from an endpoint "address" to a
   set of potential locator "addresses", and also undertake a reverse
   mapping from a locator "address" to the distinguished endpoint
   identifier "address".  The address space is hierarchically
   structured, permitting a suitably efficient mapping to be performed
   in both directions, and the underlying semantics of addresses in the
   context of public networking includes the necessary considerations of
   global uniqueness of endpoint identity token values.

   It is possible to take this approach further and allow the endpoint
   identifier to also be a valid locator.  This would imply the
   existence of a 'distinguished' or 'home' locator, and other locators
   could be dynamically mapped to this initial locator peering as
   required.  The drawback of this approach is that the endpoint
   identifier is now based on one of the transit provider's address
   prefixes, and a change of transit provider would necessarily require
   a change of endpoint identifier values within the multi-homed site.

   An alternative approach for address-formatted identifiers is to use
   distinguished identity address values which are not part of the
   global unicast locator space, allowing applications and protocol
   elements to distinguish between endpoint identity values and locators
   based on address prefix value.

   It is also possible to allow the endpoint identity and locator space
   to overlap, and distinguish between the two identity realms by the
   context of usage rather than by a prefix comparison.  However, this
   reuse of the locator token space as identity tokens has the potential
   to create the anomalous situation where a particular locator value is
   used as an identity value by a different endpoint.  It is not clear
   that the identity and locator contexts can be clearly disambiguated
   in every case, which is a major drawback to this particular approach.

   If identity values are to be drawn from the protocol's address space
   it would appear that the basic choice is to either draw these
   identity values from a different part of the address space, or use a
   distinguished or home address as both a locator and an identity.
   This latter option, that of using a locator as the basis of an
   endpoint identity on a locator, when coupled with a
   provider-aggregated address distribution architecture leads to the
   outcome of a multi-homed site using a provider-based address prefix
   as a common identity prefix.  As with locator addresses in the
   context of a single-homed network, a change of provider connectivity

Huston                   Expires April 24, 2005                [Page 20]

Internet-Draft            Multi6 Architecture               October 2004

   implies a consequent renumbering of identity across the multi-homed
   site.  If avoiding such forced renumbering is a goal here, there
   would be a preference in drawing identity tokens from a pool that is
   not aligned with network topology.  This may point to a preference
   from this sector to use of identity token values that are not drawn
   from the locator address space.

   It is also feasible to use the fully qualified domain name (FQDN) as
   an endpoint identity, undertaking a similar mapping as described
   above, using the FQDN as the lookup "key".  The implication here is
   that there is no default 'address' that is to be associated with the
   endpoint identifier, as the FQDN can be used in the context of
   session establishment, and a DNS query used to establish a set of
   initial locators.  Of course it is also the case that there may not
   necessarily be a unique endpoint associated with a FQDN, and in such
   cases if there were multiple locator addresses associated with the
   FQDN via DNS RRs, shifting between locators may imply directing the
   packet to a different endpoint where there is no knowledge of the
   active session on the original endpoint.

   The syntactic properties of these two different identity realms have
   obvious considerations in terms of the manner in which these
   identities may be used within PDUs.

   It is also an option to consider a new structured identity space
   which is not generated through the reuse of IPv6 address values nor
   drawn from the FQDN.  Given that the address space would need to be
   structured in such a fashion that permits it to be used as a lookup
   key to obtain the corresponding locator set, the obvious question in
   such an option is what additional or altered characteristics would be
   used in such an endpoint identity space that would distinguish it
   from either of the above approaches?

   Instead of structured tokens that double as lookup keys to obtain
   mappings from endpoint identities to locator sets, the alternative is
   to use an unstructured token space, where individual token values are
   drawn opportunistically for use within a multi-homed session context.
   If such unstructured tokens are used in a limited context then the
   semantics of the endpoint identity are subtly changed.  The endpoint
   identity is not a persistent alias or reference to the identity of
   the endpoint, but a means to allow the identity protocol element to
   confirm that two locators are part of the same mapped locator set for
   a remote endpoint.  In this context the unstructured opportunistic
   endpoint identifier values are used in determining locator
   equivalence rather than in some form of lookup function.

Huston                   Expires April 24, 2005                [Page 21]

Internet-Draft            Multi6 Architecture               October 2004

5.2  Persistent, Opportunistic and Ephemeral Identities

   The consideration in the previous section highlights one of the major
   aspects of variance in the method of supporting a split between
   identity and location information.

   One form uses a persistent identity field, by which it is inferred
   that the same identity value is used in all contexts where this form
   of identity is required, in support of concurrent sessions, and in
   support of sequential sessions.  This form of identity is intended to
   remain constant over time and over changes in the underlying
   connectivity.  It may also be the case that this identity is
   completely distinct from network topology, so that the same identity
   is used irrespective of the current connectivity and locator
   addressing used by the site and the host.  In this case the identity
   is persistent, and the identity value can be used as a reference to
   the endpoint stack.  This supports multi-party referrals, where if
   parties A and B establish a communication, B can pass A's identity to
   a third party C, who can then use this identity value to be the
   active party in establishing communication to A.

   If persistent identifiers are to be used to initiate a session, then
   it follows that the identity is used as a lookup key to establish a
   set of locators that are associated with the identified endpoint.  It
   is desirable that this lookup function be deterministic, reliable,
   robust, efficient and trustable.  The implication of this is that
   such identities must be uniquely assigned, and experience in identity
   systems points to a strong preference for a structured identity token
   space that has an internal hierarchy of token components.  These
   identity properties have significant commonality with those of
   unicast addresses and domain names.  The further implication here is
   that persistent structured identities also rely on the adoption of
   well-ordered distribution and management mechanisms to preserve their
   integrity and utility.  Such mechanisms generally imply a significant
   overhead in terms of administrative tasks.

   As noted in the previous section, an alternative form of identity is
   an unstructured identity space, where specific values are drawn from
   the space opportunistically.  In this case the uniqueness of any
   particular identity value is not assured in all cases.  The use of
   such identities as a lookup key to establish locators is also
   altered, as the unstructured nature of the space has implications
   relating to the efficiency of the lookup, and the authenticity of the
   lookup is weakened due to the inability to assure uniqueness of the
   identity key value.  A conservative approach to unstructured
   identities limits their scope of utility, such as per-session
   identity keys.  In this scenario the scope of the selected identity
   is limited to the parties who are communicating, and limited to the

Huston                   Expires April 24, 2005                [Page 22]

Internet-Draft            Multi6 Architecture               October 2004

   duration of the communication session.  The implication of this
   limitation is that the identity is a session-level binding point to
   allow multiple locators to be bound to the session, and the identity
   cannot be used as a reference to an endpoint beyond the context of
   the session.  Such opportunistic identities with explicitly limited
   scope do not require the adoption of any well-ordered mechanisms of
   token distribution and management.

   Another form of identity is an ephemeral form, where a session
   identity is a shared state between the endpoints, established without
   the exchange of particular token values that take the role of
   identity keys.  This could take the form of a defined locator set, or
   the form of a session key derived from some set of shared attributes
   of the session, as two examples here.  In this situation there is no
   form of reference or use of an identifier as a means of initiating a
   session.  The ephemeral identity value has a very limited role in
   terms of allowing each end to reliably determine the semantic
   equivalence of a set of locators within the context of membership of
   a particular session.

   The latter two forms of identity represents a approach to identity
   that minimizes management overhead, and provides mechanisms that are
   limited in scope to supporting session integrity.  This implies that
   support for identity functions in other contexts and at other levels
   of the protocol stack, such as within referrals, in the use of
   identities within an application's data payload, or as a key used to
   initiate a communication session with a remote endpoint would need to
   be supported by some other identity function.  Such per-session
   limited scope identities imply that the associated multi-homing
   approaches must use existing mechanisms for session startup, and the
   adoption of a session-based identity and associated locator switch
   agility becomes a negotiated session capability.  On the other hand,
   the use of a persistent identity as a session initiation key implies
   that identity is part of the established session state, and locator
   agility can be an associated attribute of the session, rather than a
   subsequent negotiated capability.  In a heterogeneous environment
   where such identity capability is not uniformly deployed this would
   imply that if a session cannot be established with a split identity
   locator binding, the application should be able to back off to a
   conventional session startup by mapping the identity to a specific
   locator value and initiating a session using such a value.  The
   reason why the application may want to be aware of this distinction
   is that if the application wishes to use self-referential mechanisms
   within the application payload, it would appear to be appropriate to
   use an identity-based self-reference only in the context of a session
   where the remote party was aware of the semantic properties of this
   referential tag.

Huston                   Expires April 24, 2005                [Page 23]

Internet-Draft            Multi6 Architecture               October 2004

   In terms of functionality and semantics opportunistic identities form
   a superset of ephemeral identities, although their implementation is
   significantly different.  Persistent identities support a superset of
   the functionality of opportunistic identities, and again the
   implementations will differ.

   In the context of support for multi-homing configurations, use of
   ephemeral identities that are used in the context of locator
   equivalence appears to represent a viable approach which allows a
   negotiated use of multiple locators within the context of
   communication between a pair of hosts in mosts contexts of
   multi-homing.  However ephemeral identities offer little more in
   terms of functionality.  They cannot be used in referential contexts,
   cannot be used to initiate communications, and provide limited means
   of support for various forms of mobility, and impose some constraints
   on the class of multi-homed scenarios that can be supported.
   Ephemeral identities are generated in the context of an established
   communication state, and the implication in terms of multi-homing is
   that a the two end points need to have discovered through existing
   mechanisms a viable pair of locators prior to generating an ephemeral
   identity binding.  The implication is that there is some form of
   static 'home' for the end points which is discivered by conventional
   referential lookup.

   The use of a persistent identity space that supports dynamic
   translation between an equivalent set of locators and one or more
   equivalent identity values offers the potential for greater
   flexibility in application, extending beyond the multi-homing
   configuration to various contexts of nomadism and mobility, as well
   as extending into service-specific functions, depending on how the
   mapping between identities and locators is managed.  However it
   remains an open question as to the nature of secure mapping
   mechanisms that would need to be used in the more general context of
   identity to locator mapping, and it is also an open question as how
   the mapping function would relate to viable endpoint-to-endpoint
   connectivity.  It is a common aspect of identity realms that the most
   critical aspect of the realm is the nature of the resolution of the
   identity into some other attribute space.

   It appears reasonable to observe that, within certain constraints,
   multi-homing does not generically require the overhead of the
   introduction of a fully distinct persistent identity space and the
   associated identity resolution functionality, and if the nature of
   the multi-homing space in this context is one of the use of a token
   to allow efficient detection of locator equivalence for session
   surviveability, then ephemeral identities appear to be an adequate
   mechanism for this role.

Huston                   Expires April 24, 2005                [Page 24]

Internet-Draft            Multi6 Architecture               October 2004

5.3  Common Issues for Multi-Homing Approaches

   The above overview encompasses a very wide range of potential
   approaches to multi-homing, and each particular approach necessarily
   has an associated set of considerations regarding its applicability.

   There are, however, a set of considerations that appear to be common
   across all approaches, and they are examined in further detail in
   this section.

5.3.1  Triggering Locator Switches

   Ultimately, regardless of the method of generation, a packet
   generated from a local multi-homed host to a remote host must have a
   source locator in the IP packet that is passed into the transit
   network.  In a multi-homed situation the local multi-homed host has a
   number of self-referential locators that are equivalent aliases in
   almost every respect.  The difference between locators is the
   inference that at the remote end the choice of locator may determine
   the path used to send a packet back to the local multi-homed host.
   The issue here is how does the local host make a selection of the
   "best" source locator to use?  Obviously the parameters of this
   selection include the objective to select a locator that represents a
   currently viable path from the remote host to the local multi-homed
   host.  Local routing information for the multi-homed host does not
   include this reverse path information.  Equally, the local host does
   not necessarily know of any additional policy constraints that apply
   to the remote host that may result in a remote host's preference to
   use one locator over another for the local host.  Considerations of
   unicast reverse path forwarding filters also indicate that the
   selection of a source locator should result in the packet being
   passed to a site-exit router that is connected to the associated ISP
   transit provider, and that the site-exit router passes the packet to
   the associated ISP.

   If the local multi-homed host is communicating with a remote
   multi-homed host, the local host may have some discretion in the
   choice of a destination locator.  The considerations relating to the
   selection of a destination locator include considerations of local
   routing state (to ensure that the chosen destination locator reflects
   a viable path to the remote endpoint), policy constraints that may
   determine a "best" path to the remote endpoint.  In such situations
   it may also be the case that the source address selection should also
   be considered in relation to the destination locator selection.

   Another common issue is the consideration of the point when a locator
   is not considered to be viable, and the consequences to the transport
   session state.

Huston                   Expires April 24, 2005                [Page 25]

Internet-Draft            Multi6 Architecture               October 2004

   o  Transport Layer Triggers

      A change in state for a currently used path to another path could
      be triggered by indications of packet loss along the current path
      transport-level signaling, or by transport session timeouts,
      assuming an internal signaling mechanism between the transport
      stack element and the locator pool management stack element.
   o  ICMP Triggers

      Path failure within the network may generate an ICMP Destination
      Unreachable ICMP packet being directed back to the sender.  Rather
      than sending this signal to the transport level as an indicator of
      session failure, the IP layer should redirect the notification
      identity module as a trigger for a locator switch.
   o  Routing Triggers

      Alternatively, in the absence of local transport triggers, the
      site exit router could communicate failure of the outbound
      forwarding path in the case where the remote host is multi-homed
      with an associated locator set.  Conventional routing would be
      incapable of detecting a failure in the inbound forwarding path,
      so there are some limitations in the approach of using routing
      triggers to change locator bindings.
   o  Heartbeat Triggers

      An alternative to these approaches is the use of a session
      heartbeat protocol, where failure of the heartbeat would cause the
      session to seek a new locator binding that would re-establish the

   The sensitivity of the locator-switch trigger is a consideration
   here.  A very fine-grained sensitivity of the locator switch trigger
   may generate false triggers arising from short-term transient path
   congestion, while coarse-grained triggers may impose an undue
   performance penalty on the session due to an extended time to detect
   a path failure.

5.3.2  Locator Selection

   The selection of a locator to use for the remote end is obviously
   constrained by the current state of the topology of the network, and
   the primary objective of the selection process is to select a viable
   locator that allows the packet to reach the intended destination
   point.  The selection of a source locator can be considered as an
   indication of preference to the remote end of a preferred locator to
   use for the local end.  However, where there are two or more viable
   locators that could be used, the selection of a particular locator
   may be influenced by a set of additional considerations.

Huston                   Expires April 24, 2005                [Page 26]

Internet-Draft            Multi6 Architecture               October 2004

   The selection of a particular locator from a viable locator set
   implies a selection of one particular network path in preference to
   other viable paths.  An implication of this selection process is that
   the host is that path selection policy and, by inference, traffic
   engineering, is not constrained to a network-based operation of path
   manipulation through adjustment of forwarding state within network
   elements.  There is an interaction between the locator selection
   process and traffic engineering functions.  The use of an address
   selection policy table, as described in RFC 3484 [7] are relevant to
   the selection process.

   The element that performs the locator selection, either as a protocol
   element within the host or a selection undertaken at a site-exit
   router, also determines traffic policy, so the choice of using remote
   packet locator rewriting or host based locator selection shifts the
   policy capability from one element to the other.

   If hosts perform this policy determination, then a more fine grained
   outcome may be achievable, particularly if the anticipated traffic
   characteristics of the application can be signalled to the locator
   selection process.  A further consideration here appears to be that
   hosts may require additional information if they are to make locator
   address selection decisions based on some form of metric of relative
   load currently being imposed on select components of a number of
   end-to-end network paths.  These considerations raise the broader
   issue of traffic engineering being a network function entirely
   independent of host function or an outcome of host interaction with
   the network.  IN the latter case there is also the consideration of
   whether the host is to interact with the network, and, if so, how is
   this interaction is to be signalled to hosts.

5.3.3  Layering Identity

   The consideration of triggering locator switch highlights the
   observation that differing information and context is present in each
   layer of the protocol stack.  This impacts on how identity / locator
   bindings are established, maintained and expired.

   These impacts include questions of what amount of state is kept, by
   which element of the protocol stack, at what level of context
   (dynamic or fixed, and per session or per host).  It also includes
   considerations of state maintenance, such as how stale or superfluous
   state information is detected and removed.  Does only one piece of
   code have to be aware of this identity/locator binding or do multiple
   transport protocols have to be altered to support this functionality?
   If so, are such changes common across all transport protocols, or do
   different protocols require different considerations in their
   treatment of this functionality?

Huston                   Expires April 24, 2005                [Page 27]

Internet-Draft            Multi6 Architecture               October 2004

   It is noted that the set of approaches considered here include
   proposals to place this functionality within the IP layer, with the
   end-to-end transport protocol layer and as a shim between the IP and
   transport protocol layer.

   Placing this identity functionality at the transport protocol layer
   implies that the identity function can be tightly associated with a
   transport session.  In this approach session startup can trigger the
   identity / locator initial binding actions and transport protocol
   timeouts can be used as triggers for locator switch actions.  Session
   termination can trigger expiration of local identity / locator
   binding state.  Where per-session opportunistic identity token values
   are being used, the identity information can be held within the
   overall session state.  In the case of persistent identity token
   values the implementation of the identity can also choose to use
   per-session state, or may choose to pool this information across
   multiple sessions in order to reduce overheads of dynamic discovery
   of identity / locator bindings for remote identities in the case of
   multiple sessions to the same remote endpoint.

   One of the potential drawbacks of placing this functionality within
   the transport protocol layer is that it is possible that each
   transport protocol will require a distinct implementation of identity
   functionality.  This is a considerable constraint in the case of UDP,
   where the UDP transport protocol has no inherent notion of a session

   An alternative approach is to use a distinct protocol element placed
   between the transport and internet layers of the protocol stack.  The
   advantage of this approach is that it would offer a consistent form
   of mapping between identities and locators for all forms of transport
   protocols.  However this protocol element would not be explicitly
   aware of sessions and would either have to discover the appropriate
   identity / locator mapping for all identity-addressed packets passed
   from the transport protocol later, irrespective of whether such a
   mapping exists and whether this is part of a session context, or have
   an additional mechanism of signaling to determine when such a mapping
   is to be discovered and applied.  At this level there is also no
   explicit knowledge of when identity / locator mapping state is no
   longer required, as there is no explicit signaling of when all flows
   to and from a particular destination has stopped and resources
   consumed in supporting state can be released.  Also, such a protocol
   element would not be aware of transport level timeouts, so that
   additional functionality would need to be added to the transport
   protocol to trigger a locator switch at the identity protocol level.
   Support of per-session opportunistic identity structure is more
   challenging in this environment, as the transport protocol layer is
   used to store and manipulate per-session state.  In constructing an

Huston                   Expires April 24, 2005                [Page 28]

Internet-Draft            Multi6 Architecture               October 2004

   identity element at this level of the protocol stack it would appear
   necessary to ensure that an adequate amount of information is being
   passed between the transport, internet protocol and identity protocol
   elements in order to ensure that the identity protocol element is not
   forced into making possibly inaccurate assumptions about the current
   state of active sessions or end-to-end network paths.

   It is also possible to embed this identity function within the
   internet protocol layer of the protocol stack.  As noted in the
   previous section, per session information is not readily available to
   the identity module, so that opportunistic per-session identity
   values would be challenging to support in this approach, as well as
   determining when identity / locator state information should be set
   up and released.  It would also appear necessary to signal transport
   level timeouts to the identity module as a locator switch trigger.
   Some attention needs to be given in this case to synchronizing
   locator switches and IP packet fragmentation, and consideration of
   IPSEC is necessary in this case, in order to avoid making changes to
   the address field in the IP packet header that trigger a condition at
   the remote end where the packet is not recognizable in the correct

5.3.4  Session Startup and Maintenance

   The next issue is that of the difference between the initial session
   startup mode of operation and the maintenance of the session state.

   In a split endpoint identifier / locator environment there needs to
   be at least one initial locator associated with an endpoint
   identifier in order to establish an initial connection between the
   two hosts.  This locator could be loaded into the DNS in a
   conventional fashion, or, if the endpoint identifier is a
   distinguished address value, the initial communication could be
   established using the endpoint identifier in the role of a locator
   (i.e.  using this as a conventional address).

   The initial actions in establishing a session would be similar.  If
   the session is based on specification of a FQDN, the FQDN is first
   mapped to an endpoint identity value, and this endpoint identity
   value could then be mapped to a locator set.  The locators in this
   set are then candidate locators for use in establishing an initial
   synchronized state between the two hosts.  Once the state is
   established it is then possible to update the initial locator set
   with the current set of useable locators.  This update could be part
   of the initial synchronization actions, or deferred until required.

   This leads to the concept of the use of a 'distinguished' locator
   that acts as the endpoint identifier, and a pool of alternative

Huston                   Expires April 24, 2005                [Page 29]

Internet-Draft            Multi6 Architecture               October 2004

   locators that are associated with this 'home' locator.  This
   association may be statically defined, using referential pointers in
   a third party referral structure (such as the DNS), or dynamically
   added to the session through the actions of the endpoint identity
   protocol stack element, or both.

   If opportunistic identities are used where the identity is not a
   fixed discoverable value but one that is generated in the context of
   a session then additional actions must be performed at session
   startup.  In this case there is still the need for defined locators
   that are used to establish a session, but then an additional step is
   required to generate session keys and exchange these values in order
   to support the identity equivalence of multiple locators within the
   ensuing session.  This may take the form of a capability exchange and
   an additional handshake and associated token value exchange within
   the transport protocol if an in-band approach is being used, or it
   may take the form of a distinct protocol exchange at the level of the
   identity protocol element, performed out-of-band from the transport

   Some approaches are capable of a further distinction, namely that of
   initial session establishment and that of establishment of additional
   shared state within the session to allow multiple locators to be
   treated as being bound to a common endpoint identity.  It is not
   strictly necessary that such additional actions be performed at
   session startup, but it appears that such actions need to be
   performed prior to any loss of end-to-end connectivity on the
   selected initial locator, so that any delay in this additional state
   exchange does increase the risk of session disruption due to
   connectivity changes.

   This raises a further question of whether the identity / locator
   split is a capability negotiation performed per session or per remote
   end, or whether the use of a distinguished identity value by the
   upper level application to identify the remote end triggers the
   identity / locator mapping functionality further down in the protocol
   stack at the transport level, and that this is performed without any
   further capability negotiation within the session.

   Within the steps related to session startup there is also the
   consideration that the passive end of the connection follows a
   process where it may need to verify the proposed new address
   contained in the source address of incoming packets before using it
   as a destination address for outgoing packets.  It is not necessarily
   the case that the sender's choice of source address reflects a valid
   path from the receiver back to the source.  While using this offered
   address appears to offer a low overhead response to connection
   attempts, if this response fails the receiver may need to discover

Huston                   Expires April 24, 2005                [Page 30]

Internet-Draft            Multi6 Architecture               October 2004

   the full locator set of the remote end through some locator discovery
   mechanism in order to establish whether there is a viable locator
   that can use a forwarding path that reaches the remote end.

   Alternatively, the passive end would use the initially offered
   locator and if this is successful leave it to the identity modules in
   each stack to exchange information to establish the current complete
   locator set for each end.  This approach implies that the active end
   of a communication needs to cycle through all of its associated
   locators as source addresses until it receives a response or exhausts
   its locator set.  While this may extend the time to confirm that no
   path exists to the remote end, it has the potential to improve the
   characteristics of the initial exchange against denial of service
   attacks that could force the remote end to engage in a high volume of
   spurious locator lookups.

5.3.5  Dynamic Capability Negotiation

   The common aspect of these approaches is that they all involve
   changes to the end-to-end interaction, as both endS of the
   communication need to be aware of this separation.  The implication
   is that this form of support for multi-homing is relatively sweeping
   in its scope, as the necessary changes to support multi-homing extend
   beyond changes to the hosts and/or routers within the multi-homed
   site and encompass changes to the IPv6 protocol itself.  It would be
   prudent when considering these changes to evaluate associated
   mechanisms that allow the communicating endpoints to discover each
   other's capabilities and only enable this form of split identity /
   locator functionality when it is established that both ends can
   support it.

   It is a corollary of this form of negotiated capability that it is
   not strictly necessary that only one form of functionality can be
   negotiated in this way.  If the adoption of a particular endpoint
   identity / locator mapping scheme is the outcome of a negotiation
   between the endpoints then it would be possible to negotiate to use
   one of a number of possible approaches.  There is some interaction
   between the approach used and the form of endpoint identity, and some
   care needs to be taken that any form of acceptable outcome of the
   endpoint identity capability negotiation is one that allows the upper
   level application to continue to operate.

5.3.6  Identity Uniqueness and Stability

   When considering the properties of long-lived identities, it is
   reasonable to assume that the identity assignation is not necessarily
   one that is permanent and unchangeable.  In the case of structured
   identity spaces the identity value reflects a distribution hierarchy.

Huston                   Expires April 24, 2005                [Page 31]

Internet-Draft            Multi6 Architecture               October 2004

   There are a number of circumstances where a change of identity value
   is appropriate.  For example, if an endpoint device is moved across
   administrative realms of this distribution hierarchy it is likely
   that the endpoint's identity value will be re-assigned to reflect the
   new realm.  It is also reasonable to assume that an endpoint may have
   more than one identity at any point in time.  RFC 3014 [8] provides a
   rationale for such a use of multiple identities.

   If an endpoint's identity can change over time, and if an endpoint
   can be identified by more than one identity at any single point in
   time, then some further characteristics of endpoint identifiers
   should be defined.  These relate to the constancy of an endpoint
   identity within an application, and the question of whether a
   transport session relies on a single endpoint identity value, and, if
   so, whether an endpoint identity can be changed within a transport
   session, and under what conditions the old identity can continue to
   be used following any such change.  If the endpoint identity is a
   long-lived reference to a remote endpoint, and if multiple identities
   can exist for a single unique endpoint, then the question arises as
   to whether applications can compare identities for equivalence, and
   whether it is necessary for applications to recognize the condition
   where different identities refer to the same endpoint.  These
   identities may be used within applications within a single host, or
   may be identifies being used on applications on different hosts.

6.  Functional Decomposition of Multi-Homing Approaches

   The following sections provide a framework for the characterization
   of multi-homing approaches through a decomposition of the functions
   associated with session establishment, maintenance and completion in
   the context of a multi-homed environment.

6.1  Establishing Session State

      What form of token is passed to the transport layer from the upper
      level protocol element as an identification of the local protocol
      What form of token is passed to the transport layer from the upper
      level protocol element as an identification of the remote session
      What form of token is used by the upper level protocol element as
      a self-identification mechanism for use within the application
      Does the identity protocol element need to create a mapping from
      the upper level protocol's local and remote identity tokens into
      an identity token that identifies the session? If so, then is this
      translation performed before or after the initial session packet
      exchange handshake?

Huston                   Expires April 24, 2005                [Page 32]

Internet-Draft            Multi6 Architecture               October 2004

      How does the session initiator establish that the remote end of
      the session can support the multi-homing capabilities in its
      protocol stack? If not, does the multi-homing capable protocol
      element report a session establishment failure to the upper level
      protocol, or silently fall back to a non-multi-homed protocol
      How do the endpoints discover the locator set available for each
      other endpoint (locator discovery)?
      What mechanisms are used to perform locator selection at each end
      for the local selection of source and destination locators?
      What form of mechanism is used to ensure that the selected site
      exit path matches the selected packet source locator?

6.2  Rehoming Triggers

      What triggers are used to identify that a switch of locators is
      Are the triggers based on the end-to-end transport session and/or
      on notification of state changes within the network path from the
      What triggers can be used to indicate the direction of the failed
      path in order to trigger the appropriate locator repair function?

6.3  Rehoming Locator Pair Selection

      What parameters are used to determine the selection of a locator
      to use to reference the local endpoint?
      If the remote endpoint is multi-homed, what parameters are used to
      determine the selection of a locator to use to reference the
      remote endpoint?
      Must a change of an egress site exit router be accompanied by a
      change in source and / or destination locators?
      How can new locators be added to the locator pool of an existing

6.4  Locator Change

      What are the preconditions that are necessary for a locator
      How can the locator change be confirmed by both ends?
      What interactions are necessary for synchronization of locator
      change and transport session behavior?

6.5  Removal of Session State

      How is identity / locator binding state removal synchronized with
      session closure?

Huston                   Expires April 24, 2005                [Page 33]

Internet-Draft            Multi6 Architecture               October 2004

      What binding information is cached for possible future use?

7.  IANA Considerations

   This document has no associated IANA actions.

8.  Security Considerations

   There are a significant number of security considerations that result
   from the action of distinguishing within the protocol suite endpoint
   identity and locator identity.

   It is not proposed to enumerate these considerations in detail within
   this document, but to reference a distinct document that describes
   the security considerations of this domain [6].

9.  Acknowledgements

   The author acknowledges the assistance from the following reviewers:
   Brian Carpenter, Kurtis Lundqvist, Erik Nordmark, Iljitsch van
   Beijnum, Marcelo Bagnulo and Joe Touch.

10  Informative References

   [1]  Abley, J., Black, B. and V. Gill, "Goals for IPv6
        Site-Multihoming Architectures", RFC 3582, August 2003.

   [2]  Lear, E., "Things MULTI6 Developers should think about", Work in
        progress: Internet Drafts
        draft-ietf-multi6-things-to-think-about-00.txt, June 2004,

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

   [4]  Nikander, P., Arkko, J., Aura, T., Montenegro, G. and E.
        Nordmark, "Mobile IP version 6 Route Optimization Security
        Design Background", Work in progress: Internet Drafts
        draft-ietf-multi6-things-to-think-abou, July 2004,

   [5]  Saltzer, J., Reed, D. and D. Clark, "End-to-End Arguments in
        System Design", ACM TOCS Vol 2, Number 4, pp 277-288, November

Huston                   Expires April 24, 2005                [Page 34]

Internet-Draft            Multi6 Architecture               October 2004

   [6]  Nordmark, E. and T. Li, "Threats relating to IPv6 multi-homing
        solutions", Work in progress: Internet Drafts
        draft-ietf-multi6-multihoming-threats-01.txt, July 2004, <http:/

   [7]  Draves, R., "Default Address Selection for Internet Protocol
        version 6 (IPv6)", RFC 3484, February 2003.

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

Author's Address

   Geoff Huston

Huston                   Expires April 24, 2005                [Page 35]

Internet-Draft            Multi6 Architecture               October 2004

Intellectual Property Statement

   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at

   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

Disclaimer of Validity

   This document and the information contained herein are provided on an

Copyright Statement

   Copyright (C) The Internet Society (2004).  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.


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

Huston                   Expires April 24, 2005                [Page 36]

Html markup produced by rfcmarkup 1.129d, available from https://tools.ietf.org/tools/rfcmarkup/