wdiff draft-ietf-mobike-design-02.txt draft-ietf-mobike-design-03.txt

IKEv2 Mobility and Multihoming T. Kivinen
(mobike) Safenet, Inc.
Internet-Draft H. Tschofenig
Expires: August 24, 2005 January 19, 2006 Siemens
February 20,
July 18, 2005

Design of the MOBIKE Protocol
draft-ietf-mobike-design-02.txt
draft-ietf-mobike-design-03.txt

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 becomes
aware will be disclosed, in accordance with
RFC 3668. Section 6 of BCP 79.

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

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

The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt.

The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.

This Internet-Draft will expire on August 24, 2005. January 19, 2006.

Abstract

The MOBIKE (IKEv2 Mobility and Multihoming) working group is
developing protocol extensions to for the Internet Key Exchange Protocol version
2 (IKEv2) to (IKEv2). These extensions should enable its use in the context where there are an efficient management of
IKE and IPsec Security Associations when a host possesses multiple IP
addresses per host or and/or where IP addresses of an IPsec host change over time
(for example, due to mobility).

This document discusses the involved network entities, and the
relationship between IKEv2 signaling and information provided by
other protocols and the rest of the network. protocols. Design decisions for the base MOBIKE protocol,
background information and discussions within the working group are
recorded.

Table of Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4
3. Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.1 Mobility Scenario . . . . . . . . . . . . . . . . . . . . 7
3.2 Multihoming Scenario . . . . . . . . . . . . . . . . . . . 8
3.3 Multihomed Laptop Scenario . . . . . . . . . . . . . . . . 9
4. Framework . . . . . . . . . . . . . . . . . . . . . . . . . . 10
5. Design Considerations . . . . . . . . . . . . . . . . . . . . 13
5.1 Indicating Support for MOBIKE . . . . . . . . . . . . . . 13
5.2 Changing a Preferred Address and Multihoming Multi-homing Support . . . 13
5.2.1 Storing a single or multiple addresses . . . . . . . . 14
5.2.2 Indirect or Direct Indication . . . . . . . . . . . . 15
5.2.3 Connectivity Tests using IKEv2 Dead-Peer Detection . . 16
5.3 Simultaneous Movements . . . . . . . . . . . . . . . . . . 17
5.4 NAT Traversal . . . . . . . . . . . . . . . . . . . . . . 18
5.5 Changing addresses or changing the paths . . . . . . . . . 19 20
5.6 Return Routability Tests . . . . . . . . . . . . . . . . . 20
5.7 Employing MOBIKE results in other protocols . . . . . . . 22 23
5.8 Scope of SA changes . . . . . . . . . . . . . . . . . . . 23 24
5.9 Zero Address Set . . . . . . . . . . . . . . . . . . . . . 24 25
5.10 IPsec Tunnel or Transport Mode . . . . . . . . . . . . . . 25
5.11 Message Representation . . . . . . . . . . . . . . . . . . 25 26
6. Security Considerations . . . . . . . . . . . . . . . . . . . 28
7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 29
8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 30
9. Open Issues . . . . . . . . . . . . . . . . . . . . . . . . . 31
10. References . . . . . . . . . . . . . . . . . . . . . . . . . 32
10.1 Normative references . . . . . . . . . . . . . . . . . . . 32
10.2 Informative References . . . . . . . . . . . . . . . . . . 32
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . 34
Intellectual Property and Copyright Statements . . . . . . . . 35

1. Introduction

The purpose of IKEv2 is used for performing mutual authentication and establishing
and maintaining to mutually authenticate two hosts, establish
one or more IPsec security associations (SAs). Security Associations (SAs) between them, and
subsequently manage these SAs (for example, by rekeying or deleting).
IKEv2 establishes enables the hosts to share information that is relevant to both
the usage of the cryptographic state (such as algorithms that should be employed
(e.g., parameters required by cryptographic algorithms and session keys
keys) and algorithms) as
well to the usage of local security policies, such as non-cryptographic
information (such as IP addresses).

The current about the traffic that should experience protection.

IKEv2 and IPsec documents explicitly say assumes that the IPsec
and an IKE SAs are SA is created implicitly between the IP
address pair that is used during the protocol execution when
establishing the IKEv2 SA. This means that there is that, in each host, only one
IP address pair is stored for the IKEv2 SA,
and this single IP address pair is used SA as an outer endpoint address part of a single IKEv2
protocol session, and, for tunnel mode IPsec SAs. After SAs, the IKE SA is created there is hosts places this
single pair in the outer IP headers. Existing documents make no
way
provision to change them. this pair after an IKE SA is created.

There are scenarios where this one or both of the IP address might change, even
frequently. addresses of this
pair may change during an IPsec session. In some cases principle, the problem could be solved by rekeying IKE SA
and all the corresponding IPsec and IKE SAs could be re-established after the IP
address has changed. However, this can be problematic, as the device
might be too slow to rekey the
SAs that often, and other scenarios the rekeying and required IKEv2
authentication might require for this task. Moreover, manual user interaction (SecurID cards etc).
Due to these reasons, a mechanism to update the IP addresses tied to
(for example when using SecurID cards) might be required as part of
the IPsec and IKEv2 SAs authentication procedure. Therefore, an automatic
mechanism is needed. MOBIKE provides solution to the
problem of the updating neeed that updates the IP addresses stored associated with the
IKE SAs SA and the IPsec SAs. MOBIKE provides such a mechanism.

The charter work of the MOBIKE working group requires IKEv2
[I-D.ietf-ipsec-ikev2], and as IKEv2 assumes therefore this document is
based on the assumption that the RFC2401bis
architecture [I-D.ietf-ipsec-rfc2401bis] mobility and multi-homing extensions
are developed for IKEv2 [I-D.ietf-ipsec-ikev2]. As IKEv2 is used, built on
the architecture described in RFC2401bis [I-D.ietf-ipsec-rfc2401bis],
all protocols developed will use within the MOBIKE working group must be
compatible with both IKEv2 and the architecture described in
RFC2401bis. MOBIKE The document does not aim to neither provide support
IKEv1 [RFC2409] or nor the old RFC2401 architecture described in RFC2401 [RFC2401].

This document is structured as follows. After introducing some
important terms in Section 2 we list a few number of relevant usage scenarios are
discussed in Section 3, to
illustrate possible deployment environments. 3. Section 4 discusses the interoperation of
MOBIKE depends on
information from with other sources (e.g., detect an address change) which
are discussed protocols and processes that may run in Section 4. the local
machine. Finally, Section 5 discusses design considerations effecting
affecting the MOBIKE protocol. The document concludes
with security considerations in Section 6. 6
with security considerations.

2. Terminology

This section introduces some useful terms for a MOBIKE protocol.

Peer:

Within the terminology that is used in this document we refer to IKEv2 endpoints as peers.
document.

Peer:

A peer is an IKEv2 endpoint. In addition, a peer implements the
MOBIKE extensions, as defined in this and therefore IKEv2. related documents.

Available address:

An address is said to be available if the following conditions are
fulfilled:
met:

* The address has been assigned to an interface of the node. interface.

* If the address is an IPv6 address, we additionally require that (a)
that the address is valid as defined in the sense of RFC 2461 [RFC2461], and that
(b) that the address is not tentative as defined in the sense of RFC 2462
[RFC2462]. In other words, we require the address assignment is
complete so that communications can
to be started. complete.

Note that this explicitly allows an address to be optimistic as
defined in [I-D.ietf-ipv6-optimistic-dad].

* If the
sense of [I-D.ietf-ipv6-optimistic-dad] even though
implementations are probably better off using other addresses
as long as there address is an alternative.
* The address IPv6 address, it is a global unicast or
unique site-local address
[I-D.ietf-ipv6-unique-local-addr]. address, as defined in [I-D.ietf-ipv6-unique-
local-addr]. That is, it is not an IPv6
link-local or site-local address. link-local. Where
IPv4 is considered, it is not an RFC 1918 [RFC1918] address.

* The address and interface is acceptable for use sending and
receiving traffic according to a local policy.

This definition is reused taken from
[I-D.arkko-multi6dt-failure-detection] [I-D.arkko-multi6dt-failure-
detection]

Locally Operational Address:

An available address is said to be locally operational when if it is available
and its use is locally known to be possible locally. and permitted. This
definition is taken from [I-D.arkko-multi6dt-failure-detection].

Operational address pair:

A pair of operational addresses are said to be an operational
address pair, iff if and only if bidirectional connectivity can be
shown between the two addresses. Note that sometimes it is
necessary to consider a 5-tuple when connectivity on a per-flow level between two
endpoints need needs to be tested. This differentiation might be
necessary to address
load balancers, certain Network Address Translation types or
specific firewalls. This definition is taken from
[I-D.arkko-multi6dt-failure-detection] [I-D.arkko-
multi6dt-failure-detection] and enhanced to fit adapted for the MOBIKE context.
Although it is possible to further differentiate unidirectional
and bidirectional operational address pairs pairs, only bidirectional
connectivity is relevant for to this document and unidirectional
connectivity is out of scope.

Path:

The route taken sequence of routers traversed by the MOBIKE and/or and IPsec packets sent via the IP
address of one peer to a specific destination address of
exchanged between the other
peer. two peers. Note that the this path might may be effected
affected not only by the involved source and destination IP addresses
addresses, but also by the selected transport
protocol and transport protocol identifier. protocol. Since MOBIKE and
IPsec packets have a different appearance on the wire they might
be routed along a different path. path, for example by load balancers.
This definition is taken from [RFC2960] and modified adapted to fit the MOBIKE
context.

Primary Path:

The primary path is sequence of routers traversed by an IP packet that carries the destination and
default source address that will
be put into a packet outbound and destination addresses is said to be the peer by default. Primary
Path. This definition is taken from [RFC2960] and modified adapted to fit the
MOBIKE context.

Preferred Address:

The IP address on which of a peer prefers to receive which MOBIKE messages and IPsec protected data traffic. traffic should
be sent by default. A given peer has only one active preferred
address at a given point in time (ignoring time, except for the small time period
where it needs to switch switches from the an old preferred
address to a new preferred address). address. This
definition is taken from [I-D.ietf-hip-mm] and modified adapted to fit the
MOBIKE context.

Peer Address Set:

We denote the two peers in this Mobike of a MOBIKE session by peer A and peer B.
A peer address set is a the subset of locally operational addresses
of peer A that are is sent to peer B. A policy available at peer A
indicates which addresses to include are included in the peer address set.
Such a policy might be impacted by manual configuration created either manually or
by automatically
through interaction with other protocols mechanisms that indicate new
available addresses.

Terminology for different regarding NAT types, such as types (e.g. Full Cone, Restricted Cone,
Port Restricted Cone and Symmetric, Symmetric), can be found in Section 5 of
[RFC3489]. For mobility related terminology, such as terminology (e.g. Make-before-break
or Break-before-make Break-before-make) see [RFC3753].

3. Scenarios

The

In this section we discuss three typical usage scenarios for the
MOBIKE protocol can be used in different scenarios. Three of
them are discussed below. protocol.

3.1 Mobility Scenario

Figure 1 shows a break-before-make mobility scenario where a mobile
node attaches to, for example a wireless LAN, to obtain connectivity changes its point of network attachment. Prior to some security gateway. This security gateway might connect the change,
the mobile node to had established an IPsec connection with a corporate network, security
gateway which offered, for example, access to a 3G network or to some other corporate network.
The initial IKEv2 exchange takes that facilitated the set up of the IPsec SA(s)
took place along over the path labeled as 'old path' from path'. The involved packets
carried the MN using its old MN's "old" IP address via and were forwarded by the
old "old"
access router (OAR) towards to the security gateway (GW). The IPsec
tunnel mode terminates there but the decapsulated data packet will
typically address another destination. Since only MOBIKE
communication from the MN to the gateway is relevant for this
discussion the end-to-end communication between the MN and some
destination is not shown in Figure 1.

When the MN moves to a new changes its point of network and attachment, it obtains a new
IP address from
a new access router (NAR) it is using statefu address configuration techniques or via the responsibility
stateless address autoconfiguration mechanism. The goal of MOBIKE MOBIKE,
in this scenario, is to avoid
restarting enable the IKEv2 exchange from scratch. As a result, MN and the GW to continue using
the existing SAs and to avoid setting up a new IKE SA. A protocol
exchange, denoted by 'MOBIKE Address Update' , will perform Update', enables the
necessary peers to
update their state update. as necessary.

Note that in a break-before-make mobility scenario the MN obtains a the new IP
address after reachability to it can no longer be reached at the old IP address has been
lost. MOBIKE is also assumed to work in scenarios where the mobile
node is able to establish connectivity with address. In
a make-before-break scenario, the new IP address while
still being MN is, for a given period of time,
reachable at both the old and the new IP address. MOBIKE should work
in both the above scenarios.

(Initial IKEv2 Exchange)
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>v
Old IP +--+ +---+ v
address |MN|------> |OAR| -------------V v
+--+ +---+ Old path V v
. +----+ v>>>>> +--+
.move | R | -------> |GW|
. | | >>>>> | |
v +----+ ^ +--+
+--+ +---+ New path ^ ^
New IP |MN|------> |NAR|--------------^ ^
address +--+ +---+ ^
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>^
(MOBIKE Address Update)

---> = Path taken by data packets
>>>> = Signaling traffic (IKEv2 and MOBIKE)
...> = End host movement

Figure 1: Mobility Scenario

3.2 Multihoming Scenario

Another scenario where MOBIKE might be used usage scenario is between two depicted in Figure 2. In this
scenario, the MOBIKE peers are equipped with multiple interfaces (and
multiple IP addresses).
Figure 2 shows two such multi-homed peers. Peer A has two interface cards with two IP addresses
addresses, IP_A1 and IP_A2. Additionally, Peer IP_A2, and peer B
also has two IP addresses, IP_B1
and IP_B2. Each peer selects one of its IP addresses as the
preferred address which will subsequently be is used for subsequent communication.
Various reasons, such as problems with the
interface card, might (e.g hardware or network link failures), may require
a peer to switch from one IP address interface to
the other one. another.

+------------+ +------------+
| Peer A | *~~~~~~~~~* | Peer B |
| |>>>>>>>>>> * Network *>>>>>>>>>>| |
| IP_A1 +-------->+ +--------->+ IP_B1 |
| | | | | |
| IP_A2 +********>+ +*********>+ IP_B2 |
| | * * | |
+------------+ *~~~~~~~~~* +------------+

---> = Path taken by data packets
>>>> = Signaling traffic (IKEv2 and MOBIKE)
***> = Potential future path through the network
(if Peer A and Peer B change their preferred
address)
Figure 2: Multihoming Scenario

Note that MOBIKE does not aim to support load balancing between
multiple IP addresses. That is, each peer uses only one of the
available IP addresses at a given point in time.

3.3 Multihomed Laptop Scenario

In the multihomed laptop

The third scenario we consider is about a laptop, which has multiple
interface cards and therefore several ways to connect to a the network.
It might may for example have a fixed Ethernet, WLAN, GPRS, Ethernet card, a WLAN interface, a
GPRS adaptor, a Bluetooth interface or USB hardware in order to send IP packets. A number of hardware. Not all
interfaces might are connected to a the network over time depending on at all times for a number of
reasons (e.g., cost, availability of certain link layer technologies,
user convenience). Note The mechanism that a policy for selecting a
network interface based on cost, etc. is out of scope for MOBIKE.
For example, the user can disconnect himself from the fixed Ethernet,
use the office WLAN, and then later leave the office and start using
GPRS during determines which interfaces
are connected to the trip home. At home he might use WLAN. At a certain network at any given point in time multiple interfaces might be connected. As such, the
laptop is a multihomed device. In any case, outside
the IP address scope of the
individual interfaces are subject to change.

The user desires to keep the established IKE-SA and IPsec SAs alive
all the time without MOBIKE protocol and, as such, this document.
However, as the need laptop changes its point of attachment to re-run the initial IKEv2 exchange
which could require user interaction as part
network, the set of IP addresses under which the authentication
process. Furthermore, even laptop is reachable,
changes too.

Even if IP addresses change due to interface switching or mobility,
the IP source and destination address obtained via the configuration payloads within IKEv2 and used inside
remain unaffected. The IP address obtained via the IKEv2
configuration payloads allow the configuration of the inner IP
address of the IPsec
tunnel remains unaffected, i.e., tunnel. As such, applications might not detect
any change at all.

4. Framework

The working group will develop a MOBIKE protocol which needs to
perform the following functionality: operations:

o Ability to inform the other peer about the peer address set

o Ability to inform the other peer about the preferred address

o Ability to test connectivity along a path and thereby to detect an outage
situation

o Ability to change the preferred address

o Ability to change the peer address set

o Ability to deal with Network Address Translation devices

The technical details of these functions are discussed below.
Although the interaction with other protocols is important for MOBIKE will have to operate correctly interact with other mechanisms, the
working group is chartered to leave this aspect outside the scope.

When a MOBIKE peer initiates a protocol exchange with its MOBIKE peer it needs to define a
peer address set based on the IP addresses available
addresses. It might to it. The peer
may want to make this peer address set available to the other peer. The IKEv2
Initiator does not need to explicitly indicate
its preferred which of the addresses in the
peer address since it set is already using its preferred address. The outgoing IKEv2 and MOBIKE messages use this This is because the
Initiator has to place its preferred address as into the source IP
address and field of the MOBIKE peer IP header with the initial message exchange.
Additionally, the Initiator expects incoming signaling messages to be sent to
arrive at this address. Interaction with
other protocols at the MOBIKE host is required to build the The peer address set and the preferred address.
address are defined based on interaction with other components within
a host. In some cases cases, the peer address set is may be available before
the initial protocol exchange and does not change during the lifetime
of the IKE-SA. The preferred address might change due to policy
reasons. Section 3 describes three scenarios in which the peer
address set is modified (by adding or deleting addresses). In these
scenarios the preferred address
needs to be changed may change as well.

Modifying

A modification of the peer address set or changing a change of the preferred
address typically is
effected by the host's result of the MOBIKE peer's local policy and
by the interaction with other protocols (such as DHCP or IPv6
Neighbor Discovery).

Figure 3 shows an example protocol interaction at between a pair of
MOBIKE peer. peers. MOBIKE interacts with the IPsec engine using the
PF_KEY interface
[RFC2367] to API [RFC2367]. Using this API, the MOBIKE daemon can create
entries in the Security Association (SAD) and Security Policy Databases.
Databases (SPD). The IPsec engine might may also interact with IKEv2 and MOBIKE. Established state at
MOBIKE daemon using this API. The content of the Security Policy and
Security Association Databases determines what traffic is protected
with IPsec databases has an impact in which fashion. MOBIKE, on the incoming and outgoing data traffic. MOBIKE other hand, receives
information from other protocols running in a number of sources that may run both in kernel-mode
and
user-mode, as previously mentioned. in user-mode. Information relevant for MOBIKE
is might be stored in
a database. The contents of such a database, referred as Trigger database, that guides along with the
occurrence of events of which the MOBIKE in its decisions process is notified, form
the basis on which MOBIKE decides regarding the set of available
addresses, the peer address set, and the preferred address. Policies might
may also affect the selection process.

Building and maintaining

The a peer address set and selecting or changing
a the preferred address based on locally available information is,
however, insufficient. This information needs to be
available to the other peer and in peer. In order to address various certain failure cases it is
necessary to test
cases, MOBIKE should perform connectivity along tests between the peers
(potentially over a number of different paths). Although a path. A number of
address pairs might may be available for connectivity tests but such tests, the most important
are the source and is
the pair (source address, destination IP address address) of the primary path
since these addresses are path.
This is because this pair is selected for sending and receiving
MOBIKE signaling messages and for sending and receiving IPsec protected data traffic. If a problem along this primary
path is detected (e.g., due to a router failure) it is necessary to
switch to the a new primary path. Optionally, periodic testing In order to be able to do so quickly,
it may be helpful to perform connectivity tests of other paths might be useful to
determine when
periodically. Such a technique would also help in identifying
previously disconnected path becomes paths that become operational.

+-------------+ +---------+
|User-space | | MOBIKE |
|Protocols | +-->>| Daemon |
|relevant for | | | |
|MOBIKE | | +---------+
+-------------+ | ^
User Space ^ | ^
++++++++++++++++++++++++++++ API ++++++ API ++++ PF_KEY ++++++++
Kernel Space v | v
_______ | v
+-------------+ / \ | +--------------+
|Routing | / Trigger \ | | IPsec |
|Protocols |<<-->>| Database |<<-+ +>+ Engine |
| | \ / | | (+Databases) |
+-----+---+---+ \_______/ | +------+-------+
^ ^ ^ | ^
| +---------------+-------------+--------+-----+
| v | | |
| +-------------+ | | |
I | |Kernel-space | | | | I
n | +-------->+Protocols +<----+-----+ | | n
t v v |relevant for | | v v v t
e +----+---+-+ |MOBIKE | | +-+--+-----+-+ e
r | Input | +-------------+ | | Outgoing | r
f | Packet +<--------------------------+ | Interface | f
==a>|Processing|===============================| Processing |=a>
c | | | | c
e +----------+ +------------+ e
s s
===> = IP packets arriving/leaving a MOBIKE node
<-> = control and configuration operations

Figure 3: Framework

Please note that Figure 3 is only one way illustrates an example of implementing MOBIKE. how a MOBIKE
implementation could work. Hence, it serves illustrative purposes
only.

Extensions of the PF_KEY interface required by MOBIKE are also within
the scope of the working group. Finally, certain optimizations in for
wireless
environment will environments are also be covered.

5. Design Considerations

This section discusses aspects affecting the design of the MOBIKE
protocol.

5.1 Indicating Support for MOBIKE

In order for MOBIKE to function correctly, function, both peers must implement
this extension. We propose extensions to IKEv2 described below for the MOBIKE support.
extension of IKEv2. If only one peer supports MOBIKE, then or none of the peers supports MOBIKE,
then, whenever an IP address changes, IKEv2 will just run IKEv2. Specifically, have to be re-run in
order to create a node new IKE SA and the respective IPsec SAs. In
MOBIKE, a peer needs to be able confident that its address change messages
are understood by the other peer. If these messages are not
understood, it is possible that connectivity between the peers is
lost.

One way to
guarantee ensure that a peer receives feedback on whether or not its address change
messages are understood by its
corresponding peer. Otherwise an address change may render the
connection useless, and it other peer, is important that both sides realize this
as early by using IKEv2
messaging for MOBIKE and to mark some messages as possible.

Ensuring that "critical".
According to the IKEv2 specification, such messages are understood can in be arranged either have to be
understood by marking some IKEv2 payloads critical so that they are either
processed the receiver, or an error message has to be returned to
the sender.

A second way to ensure receipt of the above-mentioned feedback is returned, by
using Vendor ID payloads or via a Notify.

The first solution approach is to use Vendor ID payloads that are exchanged during the initial IKEv2 exchange using a specific string denoting MOBIKE to
signal the support of the MOBIKE protocol. This ensures that in all
cases a MOBIKE capable node knows
exchange. These payloads would then indicate whether its or not a given
peer supports the MOBIKE or
not.

The second solution protocol.

A third approach uses would use the Notify payload which is also used for
NAT detection (via NAT_DETECTION_SOURCE_IP and
NAT_DETECTION_DESTINATION_IP).
NAT_DETECTION_DESTINATION_IP payloads).

Both a Vendor ID and a Notify payload might may be used to indicate the
support of certain extensions.

Note that the node a MOBIKE peer could also attempt to execute MOBIKE
opportunistically with the critical bit set when an address change
has occurred. The drawback of this approach is, however, that the an
unnecessary MOBIKE message exchange is introduced.

Although Vendor ID payloads and Notifications are technically
equivalent
equivalent, Notifications are already used in IKEv2 as a capability
negotiation mechanism mechanism. Hence, Notifications and is therefore Vendor ID payloads
are the preferred mechanism. mechanisms.

5.2 Changing a Preferred Address and Multihoming Multi-homing Support

From MOBIKE's point of view multihoming view, support for multi-homing is integrated inherently
provided by
supporting the fact that it manages a peer address set of peer addresses, rather
than a single address and
protocol address. Further, MOBIKE provides mechanisms to change to use
a new peer's preferred IP address.

From a protocol point of view each Each peer needs to learn the
preferred address and the peer address set either implicitly or explicitly. set.

5.2.1 Storing a single or multiple addresses

One design decision is whether an IKE-SA should store be associated with a
single IP address or multiple IP addresses as part of the peer address set. addresses. One option is that the other end a
peer can provide a list of addresses to its counterpart which can
then be used as destination addresses.

Note that MOBIKE does not support load balancing, i.e., only one IP
address is set to a preferred address at a time and changing the
preferred address typically requires some MOBIKE signaling.

Another option is to only communicate one address to the other peer
and both peers only use that address when communicating. If this
preferred
address cannot be used anymore then an address update is sent to the
other peer changing that changes the preferred address.

Alternatively, if peer A provides A, for example,provides a peer address set
with multiple IP addresses then peer B can recover from a failure of
the preferred address on its
own, meaning that when without further communication with peer A. That
is, if it detects that the primary path does not work anymore it can
either switch to a new preferred address locally (i.e., causing changing the
source IP address of outgoing MOBIKE messages to
have a non-preferred address) messages) or to try an IP
address from A's peer address set. set (i.e., changing the destination
address). If peer B only received a single IP address as the A's
peer IP address set from peer A
for A then peer B can only change its own preferred address. The other end has Peer B
would have to wait for an address update from peer A with a new IP
address in order to fix the problem.

The main advantage about using of storing only a single IP address for the remote host
peer is that it makes retransmission easy, and handling easier. Moreover, it also
simplifies the recovery procedure. The peer, peer whose IP address changed, changed
must start the recovery process and send the new IP address to the
other peer.
Failures However, connectivity failures along the path are not
well covered addressed with advertising a single IP address.

The single IP address approach will does not work if both peers happen to
loose change
their IP address addresses at the same time (due to, say, the failure of
one of the links that time, for example if both nodes hosts move
simultaneously, even though multiple addresses are connected to). It may also
require available to the
two peers. The IKEv2 implementation might also require window size
to be larger than 1, especially if only
direct indications are used. This is 1 because the host MOBIKE peer needs to be able to send
the IP address change notifications before it can switch switches to another address, and depending
address. Depending on the occurrence of return routability checks,
retransmissions policies etc, it and similar message exchanges a window size
larger than 1 might be hard required to make the protocol
such that it works deal with window size of 1 too. more than one pending
response at the same time. Furthermore, the single IP address
approach does not really benefit much from indirect indications as
the peer receiving these indications might not be able to fix the
situation by itself (e.g., even if a peer receives an ICMP host
unreachable message for the old IP address, it cannot try other another IP addresses,
address, since they are unknown). it does not know any).

The problems with IP address lists are lie mostly in its their complexity.
Notification and recovery processes are more complex. complicated. Both ends
can recover from the IP address changes. However, both peers should
not attempt to recover at the same time or nearly the same time as it
this could cause them to lose connectivity. The MOBIKE protocol is
required to prevent this.

The previous discussion is independent of the question of how many
addresses to send in a single MOBIKE message. A MOBIKE message might
be able to carry more than one IP address (with the IP address list
approach) or a single address only. The latter approach has the
advantage of dealing with NAT traversal in a proper fashion. A NAT
cannot does not change addresses
carried inside the MOBIKE message but it can change IP address (and
transport layer addresses) in the IP header and Transport Protocol
header (if NAT traversal is enabled). enabled as part of configuration or
dynamically through the NAT discovery mechanism. Furthermore, a
MOBIKE message carrying the peer address set could be idempotent
(i.e., always resending the full address list) or the protocol may
allow add/delete operations to be specified.
[I-D.dupont-ikev2-addrmgmt], [I-D.dupont-ikev2-
addrmgmt], for example, offers an approach which defines add/delete
operations. The same is true for the dynamic address reconfiguration
extension for SCTP [I-D.ietf-tsvwg-addip-sctp].

5.2.2 Indirect or Direct Indication

An indication to change the preferred IP address might be either
direct or indirect.

Direct indication:

A direct indication means that the other peer will send an message
with the address change. Such a message This can, for example, be accomplished
by having MOBIKE sending an authenticated address update
notification with a different preferred address.

Indirect indication:

An indirect indication to change the preferred address can be
obtained by observing the environment. An indirect indication
might, for example, be be the receipt of an ICMP message or
information of about a link failure. This information should be seen
as a hint and might should not directly cause any changes to a change of the remote peer's
preferred address. Depending on the local policy policy, MOBIKE might may
decide to perform a dead-peer detection exchange for the preferred
address pair (or another address pair from the peer address set).
When a peer detects that the other end started to use a different
source IP address than before, it might want to authorize the new
preferred address (if not already authorized). A peer might also
start Authorization aims
to ensure that a connectivity test of this particular address. If a
bidirectionally operational address pair peer is selected then MOBIKE
needs to communicate this new preferred address allowed to its remote
peer.

MOBIKE will utilize both mechanisms, direct and indirect indications, use the indicated
address. Claiming to make be at an arbitrary address without
performing a more intelligent decision which return-routability test or without verifying that the
IP address pair to select as is listed within a certificate opens the preferred door for
various denial of service attacks. Hence a peer may also start a
connectivity test of this particular address. The

If more information will be is available to the MOBIKE daemon then a more
intelligent decision regarding the faster selection of a new primary path
can be selected among the
available candiates. made.

5.2.3 Connectivity Tests using IKEv2 Dead-Peer Detection

This section discusses the suitability of the IKEv2 dead-peer
detection (DPD) mechanism for connectivity tests between address
pairs. The basic IKEv2 DPD mechanism is not modified by MOBIKE but
it needs to be investigated whether it can be used for MOBIKE
purposes as well.

The IKEv2 DPD mechanism is done by involves sending an empty informational
exchange packet to a given address of the other peer, in which case remote peer. On receipt,
the other end will
respond remote peer responds with an acknowledgement. If no
acknowledgement is received after a certain timeout period (and after
couple of retransmissions), the
other remote peer is considered to be not reachable. Note that
reachable at the address in question. On the other hand, receipt of
IPsec protected data traffic acknowledgement is also a guarantee that the other peer is alive.
reachable at the address in question.

When reusing dead-peer detection in MOBIKE for connectivity tests it
seems to be reasonable to try other IP addresses (if they are
available) in case of an unsuccessful connectivity test for a given
address pair. Dead-peer detection messages using a different address
pair should use the same message-id as the original dead-peer
detection message (i.e. they are simply retransmissions of the
dead-peer dead-
peer detection packet using different destination IP address). If
different message-id is used then it violates constraints placed by
the IKEv2 specification on the DPD message with regard to the
mandatory ACK for each message-id, causing the IKEv2 SA to be
deleted.

If MOBIKE strictly follows the guidelines of the dead-peer detection
mechanism in IKEv2 then an IKE-SA should be marked as dead and
deleted if the connectivity test for all available address pairs
fails. Note that this is not in-line with the approach used, for
example, in SCTP where a failed connectivity test for an address does
not lead to (a) the IP address or IP address pair to be marked as
dead and (b) delete state. Connectivity tests will be continued for
the address pairs in hope that the problem will recover soon. This
comparison with SCTP aims to point at another IETF protocol that aims
to address the multi-homing problem (although with a different scope
and a different layer).

Note that as IKEv2 implementations might may have a window size of 1, which
prevents it from initiating and
therefore be unable to initiate a dead-peer detection exchange while
doing
another exchange. exchange is pending. As a result, all other exchanges experience
the identical retransmission policy as used for the dead-peer
detection.

The dead-peer detection for the other IP address pairs can also be
executed simultaneously (with a window size larger than 1), meaning
that after the initial timeout of the preferred address expires, we
send packets simultaneously to all other address pairs. It is
necessary to differentiate individual acknowledgement messages in
order to determine which address pair is operational. Therefore the
source IP address of the acknowledgement should match the destination
IP towards the message that was previously sent.

Also the other peer is most likely going to reply only to the first
packet it receives, and that first packet might not be the best (by
whatever criteria) address pair. The reason the other peer is only
responding are
subject to the first packet it receives is that implementations
should not send retransmissions if they have just sent out an identical response message. This is retransmission policy as used for the dead-
peer detection. To use a different policy for different message
types seems to protect be reasonable.

The dead-peer detection mechanism for the packet
multiplication problem, which other IP address pairs can happen
also be executed simultaneously if some node in the network
queues up packets and then sends them to window size larger than 1,
meaning that after the destination. If initial timeout period of the
destination were to reply preferred
address expires, DPD packets are sent simultaneously to all of them then the other end will
again see multiple packets, and will reply
address pairs. It is necessary to all differentiate acknowledgement
messages in order to determine which address pair is operational.
The source IP address of them, etc. the acknowledgement can be used for this
purpose.

The protocol should also be nice to the network, meaning, that when
some core router link goes down, and a large number of MOBIKE clients notice it,
this, they should not start sending a large number of messages while
trying to recover from the problem. This might may be especially
bad particularly
unfortunate because packets are may be dropped because of congestion in
the congested network. first place. If MOBIKE clients simultaneously try to test all
possible address pairs by executing connectivity tests then the
congestion problem only gets worse.

Also note that the IKEv2 dead-peer detection is not sufficient for
the return routability check. See Section 5.6 for more information.

5.3 Simultaneous Movements

MOBIKE does not aim to provide a full mobility solution which that allows
simultaneous movements. Instead, the MOBIKE working group focuses on
selected scenarios as described in Section 3. Some of the scenarios
assume that one peer has a fixed set of addresses (from which some
subset might be in use). Thus it cannot move to the an address that is
unknown to the other peer. Situations in which both peers move and the
movement notifications cannot reach the other peer
simultaneously are outside the scope of the MOBIKE WG. MOBIKE has
not being been chartered to deal with the rendezvous problem, or with the
introduction of any new entities in the network.

Note that if only a single address is stored in the peer address set
(instead of an address list) we might end up in the case where it
seems that both peers changed their addresses at the same time. time (e.g.,
if both nodes change their addresses at the same time). This is
something that the MOBIKE protocol must deal with.

Three cases can be differentiated:

o Two mobile nodes obtain a new address at the same time, and then
being unable to tell each other where they are (in a
break-before-make break-before-
make scenario). This problem is called the rendezvous problem,
and is traditionally solved by introducing another third entity,
for example, the home agents (in Mobile IPv4/IPv6) or forwarding
agents (in the Host Identity Protocol). Essentially, solving this
problem requires the existence of a
stable additional infrastructure node somewhere. nodes.

o Simultaneous changes to addresses such that at least one of the
new addresses is known to the other peer before the change
occurred.

o No simultaneous changes at all.

5.4 NAT Traversal

IKEv2 has support of supports legacy NAT traversal built-in. traversal. This feature is known as NAT-T
which allows IKEv2 to work even if a NAT along the path between the
Initiator and the Responder modified modifies the source and possibly the
destination IP address. With NAPT even the transport protocol
identifiers are modified (which then requires UDP encapsulation for subsequent
exchanged IPsec protected data traffic). Therefore, the MOBIKE
daemon needs to obtain to required IP address information is taken from informationfrom the IP
header (if a NAT was detected who rewrote that modified the IP header information) header) rather
than from the protected payload. This problem is not new and
was discovered during the design is an
issues of every mobility protocol where the only
valuable most important
information exchanged is the IP address information. .

One of the goals in the MOBIKE protocol is to securely exchange one
or more addresses of the peer address set and to securely set the
primary address. If no other protocol is used to securely retrieve
the IP address and port information allocated by the NAT then it is
not possible to tackle all attacks against MOBIKE. Section 6
discusses this aspect in more detail. Various solution approaches to solve
the problem have been presented:

o Securely retrieving IP address and port information allocated by
the NAT using a protocol different from MOBIKE. This approach is
outside the scope of the MOBIKE working group since other working
groups, such as MIDCOM and NSIS, already deal with this problem.
The MOBIKE protocol can benefit from the interaction with these
protocols.
protocols but the interaction with these protocols it outside the
scope of this document.

o Design a protocol in such a way that NAT boxes are able to inspect
(or even participate) in the protocol exchange. This approach was
taken with HIP the Host Identity Protocol but is not an option for
IKEv2 and MOBIKE. MOBIKE since most IKEv2 messages are encrypted with the
established IKE SA. This prevents the NAT from learning required
information from the protocol exchange in a similar fashion as in
HIP.

o Disable NAT-T support by indicating the desire never to use information
from the (unauthenticated) header. While this approach prevents
some security problems it effectively disallows the protocol to
work in certain environments. environments with NATs.

There is no way to distinguish the cases where whether there is a NAT device
along the path that modifies the header information in packets or whether an
adversary mounts mounting an attack. If a NAT is detected in during the
creation of an IKE SA
creation, SA, that should automatically disable the MOBIKE
extensions and use NAT-T.

A design question is whether NAT detection capabilities should be
enabled only during the initial IKEv2 exchange or even at also during
subsequent message exchanges. If MOBIKE is executed with no NAT
along the path when the IKE SA was created, then a NAT which appears
after moving to a new network cannot be dealt with if NAT detection
is enabled only during the initial exchange. Hence, it might be is desirable
to also support a scenario where a MOBIKE peer moves from a network which
does subnet
that is not deploy behind a NAT to a network which uses a private address
space. that is.

A NAT prevention mechanism can be used to make sure that no adversary
can interact with the protocol if no NAT is expected between the
Initiator and the Responder.

The (reference? Explanation?)

Whether or not MOBIKE should support for NAT traversal is certainly one of the most
important design decisions with an impact on other protocol aspects. decisions.

5.5 Changing addresses or changing the paths

A design decision is whether it is enough for the MOBIKE protocol to
detect dead address, addresses, or does it need also needs to detect also dead paths. Dead
address detection means that we only detect that we cannot get
packets through refers to that remote address by using the local IP address ability to establish whether or not a
given by the local IP-stack (i.e., local IP address selected normally by
the routing information). pair is operational. Dead path detection means that we need refers to
try all possible local interfaces/IP addresses for each remote
addresses, i.e., find all possible paths between
the hosts and try
them all ability to see which of them work establish whether or not all possible (local/remote)
address pairs are operational (or at least find one working
path). such pair).

While performing just an one address change is simpler, the existence of
locally operational addresses are is not, however, a guarantee that
communications can be established with the peer. A failure in the
routing infrastructure can prevent the sent packets from reaching
their destination. Or a failure of an interface on one side can be
related to the failure of an interface on the other side, making it
necessary to change more than one address at a time.

5.6 Return Routability Tests

Setting a new

Changing the preferred address which is and subsequently used using it for
communication is associated with an authorization decision: Is a peer
allowed to use this address? Does this peer own this address? Two
mechanisms have been proposed in the past to allow a peer to
determine the answer to this question:

o The addresses a peer is using are part of a certificate.
[RFC3554] which is introduced by this approach. If the other peer
is, for example, a security gateway with a limited set of fixed IP
addresses, then the security gateway may have a certificate with
all the IP addresses appear in the certificate.

o A return routability check is performed by the remote peer before
the address is used
to ensure updated in that peer's Security Association
Database. This is done in order provide a certain degree of
confidence to the remote peer that local peer is reachable at the
indicated address.

Without performing taking an authorization decision a malicious peer can
redirect traffic towards a third party or a blackhole.

An IP address

A MOBIKE peer should not be used use an IP addressed provided by another
MOBIKE peer as a primary address without computing the authorization
decision. If the addresses are part of
the certificate then it is not necessary to execute the weaker return
routability test. the certificate then it is
not necessary to execute the weaker return-routability test. The
return-routability test is a form of authorization check, although it
provides weaker guarantees then the inclusion of the IP address as
part of a certificate. If multiple addresses are communicated to another
peer as part of the
remote peer address set then some of these addresses
might may be already verified even
if the primary address is still operational.

Another option is to use the [I-D.dupont-mipv6-3bombing] approach
which suggests to do perform a return routability test only if you have to
send when an
address update needs to be sent from some other address other than the
indicated preferred address.

Finally it would be possible not to execute return routability checks
at all. In case of indirect change notifications then we only move to the
new preferred address after successful dead-peer detection (i.e., a
response to a DPD test) on the new address, which is already a return
routability check. With a direct notifications notification the authenticated peer
may have provided an authenticated IP address, thus we could address. Thus it is would be
possible to simply trust the other end. MOBIKE peer to provide a proper IP
address. There is no way external attacker an adversary can cause any attacks, but we are successfully launch an
attack by injecting faked addresses since it does not protected from know the IKE SA
and the corresponding keying material.A protection against an
internal attacker, i.e. the authenticated peer forwarding its traffic
to the new address. address, is not provided. This might be an issue when
extensions are added to IKEv2 that do not require authentication of
end points (e.g., opportunistic security using anonymous Diffie-
Hellman). On the other hand we
do know the identity of the peer in that
case.

As such, it The identity of the IKEv2 Initiator and the IKEv2 Responder
can take various forms: IP address, FQDN, DN, email address alike
identifiers, etc.

It seems that there it is also a policy issue when to schedule a
return routability test.

The basic format of the return routability check could be similar to
dead-peer detection, but the problem is that if that fails then the
IKEv2 specification requires the IKE SA to be deleted. Because of
this we might need to do some kind a different type of other exchange. exchange is required and thereby avoiding
modifications to the IKEv2 specification.

There are potential attacks if a return routability check does not
include some kind of nonce. In this attack the The valid end point sends could send an
address update notification for the a third party, trying to get all the
traffic to be sent there, causing a denial of service attack. If the
return routability checks does not contain any cookies or other
random information not known by to the other end, then that valid node
could reply to the return routability checks even when it cannot see
the request. This might cause the other end a peer to turn move the traffic to
there, even when a
location where the true original recipient cannot be reached at
this address. reached.

The IKEv2 NAT-T mechanism does not perform any return routability checks.

It simply uses the last seen source IP address used by the other peer
as the destination address to send response packets. An adversary
can change those IP addresses, and can cause the response packets to
be sent to wrong IP address. The situation is self-fixing when the
adversary is no longer able to modify packets and the first packet
with true an unmodified IP address reaches the other peer. In a certain
sense mobility handling makes Mobility
environments make this attack more difficult for an adversary since
it needs requires the adversary to be located somewhere along the path where on the individual
paths ({CoA1, ..., CoAn} towards the destination IP address) have a
shared path or if the adversary is located near the MOBIKE client
then it needs to follow the users user mobility patterns. With IKEv2
NAT-T, the genuine client can cause third party bombing by
redirecting all the traffic pointed to him to third party. As the
MOBIKE protocol tries to provide equal or better security than IKEv2
NAT-T the MOBIKE protocol mechanism it should protect against these attacks.

There might may be also return routability information available from the other
parts of the system too (as shown in Figure 3), but the checks done might
may have a different quality. There are multiple levels for return
routability checks:

o None, no tests

o A party willing to answer the return routability check is on located
along the path to the claimed address. address (). This is the basic form
of return routability test.

o There is an answer from the tested address, and that answer was
authenticated (including the address) to be from our peer.
authenticated, integrity and replay protected.

o There was an authenticated authenticated, integrity and replay protected answer
from the peer, but it is not guaranteed to be from originate at the tested
address or path to it (because the peer can construct a response
without seeing the request).

The basic return routability checks do not protect against 3rd party
bombing if the attacker is along the path, as the attacker can
forward the return routability checks to the real peer (even if those
packets are cryptographically authenticated).

If the address to be tested is carried inside the MOBIKE packet too, payload,
then the adversary cannot forward packets, thus it prevents forward packets. Thus 3rd party
bombings. bombings
are prevented.

If the reply packet can be constructed without seeing the request
packet (for example, if there is no nonce, challenge or similar
mechanism to show liveness), then the genuine peer can cause 3rd
party bombing, by replying to those requests without seeing them at
all.

Other levels might only provide information a guarantee that there is someone a node at
the IP address which replied to the request. There might not be an is no indication that
as to whether or not the reply was freshly generated or repeated, is fresh, and whether or not the
request might may have been transmitted from a different source address.

Requirements for a MOBIKE protocol for the return routability test
might be able to verify that there is the same (cryptographically)
authenticated node at the other peer and it can now receive packets
from the IP address it claims to have.

5.7 Employing MOBIKE results in other protocols

If MOBIKE has learned about new locations or verified the validity of
an
a remote address through a return routability test, check, can this
information be useful for other protocols?

When considering the basic MOBIKE VPN scenario, the answer is no.
Transport and application layer protocols running inside the VPN
tunnel have no consideration about are unaware of the outer addresses or their status.

Similarly, IP layer tunnel termination at a gateway rather than a
host endpoint limits the benefits for "other protocols" that could be
informed -- all application protocols at the other side are running
in a node that is unaware
of IPsec, IKE, or MOBIKE.

However, it is conceivable that future uses or extensions of the
MOBIKE protocol make such information distribution useful. For
instance, if transport mode MOBIKE and SCTP were made to work
together, it would likely potentially be useful for SCTP to learn about the
new addresses at the same time as MOBIKE learns them. MOBIKE. Similarly, various IP
layer mechanisms might may make use of the fact that a return routability
test of a specific type has already been performed. However, in all of these cases careful consideration care should be
applied. This consideration should answer to questions such as
whether other alternative sources exist for the information, whether
dependencies are created between parts that prior to this had no
dependencies, and what the impacts
exercised in terms of number of messages or
latency are. all these situations .

[I-D.crocker-celp] discussed discusses the use of common locator information
pools in a IPv6 multihoming multi-homing context; it assumed that both transport
and IP layer solutions would are be used for providing multihoming, in order to support multi-homing,
and that it would be beneficial for different protocols to coordinate
their results in some manner, way, for instance by sharing experienced throughput
information for of address pairs. This may apply to MOBIKE as well,
assuming it co-exists with non-IPsec protocols that are faced with
the same multiaddressing or similar multi-homing choices.

Nevertheless, all of this is outside the scope of current MOBIKE base
protocol design and may be addressed in future work. (so why do you
elaborate so much on this stuff?)

5.8 Scope of SA changes

Most sections of this document discuss design considerations for
updating and maintaining addresses for in the database entries that
relate to an IKE-SA. However, changing the preferred address also has an impact for
affects the entries of the IPsec SAs. SA database. The outer tunnel
header addresses (source IP and destination IP addresses) need to be
modified according to the primary path to allow the IPsec protected
data traffic to travel along the same path as the MOBIKE packets (if
we only consider the IP header information). If the MOBIKE messages
and the IPsec protected data traffic travel along a different path
then NAT handling is severely complicated.

The basic question is then how the IPsec SAs are changed to use the
new address pair (the same address pair as the MOBIKE signaling
traffic -- the primary path). One option is that when the IKE SA
address is changed then automatically all IPsec SAs associated with
it are moved along with it to new address pair. Another option is to
have a separate exchange to move the IPsec SAs separately.

If IPsec SAs should be updated separately then a more efficient
format than the notification payload is needed. needed to preserve bandwidth.
A notification payload can only store one SPI per payload. A
separate payload which would have list of IPsec SA SPIs and new
preferred address. If there are is a large number of IPsec SAs, those
payloads can be quite large unless ranges of SPI values are
supported. If we automatically move all IPsec SAs when the IKE SA
moves, then we only need to keep track which IKE SA was used to
create the IPsec SA, and fetch the IP addresses. Note that IKEv2
[I-D.ietf-ipsec-ikev2] already requires implementations to keep track
which IPsec SAs are created using which IKE SA.

If we do allow each IPsec SAs SA address set to be updated separately,
then we can support scenarios, where the machine has fast and/or
cheap connections and slow and/or expensive connections, and it wants
to allow moving some of the SAs to the slower and/or more expensive
connection, and prevent the move, for example, of the news video
stream from the WLAN to the GPRS link.

On the other hand, even if we tie the IKE SA update to the IPsec SA
update, then we can create separate IKE SAs for this scenario, e.g.,
we create one IKE SA which have both links as endpoints, and it is
used for important traffic, and then we create another IKE SA which
have only the fast and/or cheap connection, which is then used for
that kind of bulk traffic.

5.9 Zero Address Set

One of the features which might be is potentially useful would be is for the peer to
announce the other end that it will now disconnect for some time,
i.e., i.e. it will not
be reachable at all. For instance, a laptop might go to suspend
mode. In this case the peer it could send address notification with zero
new addresses, which means that it will not have any valid addresses
anymore. The responder of that kind of notification would then
acknowledge that, and could then temporarily disable all SAs. SAs and
therefore stop sending traffic. If any of the SAs gets any packets
they are simply dropped. This could also include some kind of ACK
spoofing to keep the TCP/IP sessions alive (or simply set the TCP/IP
keepalives and timeouts large enough not to cause problems), or it
could simply be left to the applications, e.g., e.g. allow TCP/IP sessions
to notice the link is broken.

The local policy could then decide how long the peer would allow
other peers to be disconnected, e.g., whether this is only allowed
for few minutes, or do they allow users to disconnect Friday evening
and reconnect Monday morning (consuming resources during weekend too,
but on the other hand not more than is normally used during week
days, but we do not need lots of extra resources on then indicate how long the Monday
morning to support all those people connecting back peer should allow
remote peers to network). remain disconnected.

From a technical point of view this feature addresses two aspects:

o There is no need to transmit IPsec data traffic. IPsec protected
data needs to be dropped which saves bandwidth. This does not
provide a functional benefit i.e, benefit, i.e., nothing breaks if this feature
is not provided.

o MOBIKE signaling messages are also ignored and need to be
suspended. ignored. The IKE-SA must not
be deleted and the suspend functionality (realized with the zero
address set) might may require the IKE-SA to be tagged with a lifetime
value since the IKE-SA
will should not be kept in memory alive for an arbitrary amount
undefined period of time. Note that IKEv2 does not require that
the IKE-SA has no a lifetime associated with it. In order to prevent
the IKE-SA to be from being deleted the dead-peer detection mechanism
needs to be suspended as well.

Due to the enhanced complexity of fact that this extension would be complicated, a first
version of the MOBIKE protocol will not provide this feature.

5.10 IPsec Tunnel or Transport Mode

Current MOBIKE design is focused only on the VPN type usage and
tunnel mode. Transport mode behaviour would also be useful, but will
be discussed in future documents.

5.11 Message Representation

The basic IP address change notifications can be sent either via an
informational exchange already specified in the IKEv2, or via a
MOBIKE specific message exchange. Using informational exchange has
the main advantage that it is already specified in the IKEv2 and
implementations incorporated incorporate the functionality already.

Another question is the basic format of the address update notifications.
The address update notifications can include multiple addresses, of
which some can may be IPv4 and some IPv6 addresses. The number of
addresses is most likely going to be quite small (less limited in typical environments
(with less than
10). 10 addresses). The format might may need to indicate a
preference value for each address. Furthermore, one of the addresses in the peer address set
must be labeled as the preferred address. The format could either contain the a
preference number, giving out number that determines the relative order of the
addresses, or it could simply be ordered ordered, according to preference,
list of IP addresses in the
order of the most preferred first. addresses. While two addresses can have the same
preference value an ordered list avoids this problem. situation.

Even if load balancing is currently outside the scope of MOBIKE,
there might be
future work to include this feature. might include. The format selected format needs to be flexible
enough to include additional information (e.g., (e.g. to enable load
balancing). This might may be
something like one realized with an reserved field, which can
later be used to store additional information. As there is may arise
other potential information which might may have to be tied to the an address in the
future, a reserved field seems like a prudent design in any case.

There are two basic formats for putting that place IP address list in to the
exchange, we can include lists into a message.
One includes each IP address as separate payload (where the payload
order indicates the preference value, or the payload itself might
include the preference value), or we can put the IP address list as
one payload to the exchange, and that one payload will then have
internal format which includes the list of IP addresses.

Having multiple payloads with each one having carrying one IP address
makes the protocol probably easier to parse, as we can already use
the normal IKEv2 payload parsing procedures to get the list out. procedures.. It also offers an easy
way for the extensions, as the payload probably contains only the
type of the IP address (or the type is encoded to the payload type),
and the IP address itself, and as each payload already has length
associated to it, we can detect if there is any extra data after the
IP address. Some implementations might have problems parsing too long of a list of IKEv2 payloads,
payloads that are longer than a certain threshold, but if the sender
sends them in the most preferred first, the other end receiver can simply only take use the
first addresses and use them. addresses.

Having all IP addresses in one big MOBIKE specified internal format
provides more compact encoding, and keeps the MOBIKE implementation
more concentrated to one module.

The next It also avoids problems of packets
arriving in an order different from what they were sent.

Another choice is which type of payloads to use. IKEv2 already
specifies a notify payload. It includes some extra fields (SPI size,
SPI, protocol etc), which gives 4 bytes of the extra overhead, and
there is the notify data field, which could include the MOBIKE
specific data.

Another option would be to have a custom payload type, which then
includes the information needed for the MOBIKE protocol.

MOBIKE might send the full peer address list once one of the IP
addresses changes (either addresses are added, removed, the order
changes or the preferred address is updated) or an incremental
update. Sending incremental updates provides more compact packets
(meaning we can support more IP addresses), but on the other hand
have more problems in the synchronization and packet reordering
cases, i.e., the incremental updates must be processed in order, but
for full updates we can simply use the most recent one, and ignore
old ones, even if they arrive after the most recent one (IKEv2
packets have message id which is incremented for each packet, thus we
know the sending order easily).

Note that each peer needs to communicate its peer address set to the
other peer.

6. Security Considerations

As all the messages are already authenticated by the IKEv2 there is
no problem that any attackers would modify the actual contents of the
packets. The IP addresses in the IP header of the packets are not
authenticated, thus the protocol defined must take care that they are
only used as an indication that something might be different, they
should and
that do not cause any direct actions.

Attacker

An attacker can also spoof ICMP error messages in an effort to
confuse the peers about which addresses are not working. At worst
this causes denial of service and/or the use of non-preferred addresses,
so it is not that serious.
addresses.

One type of attack that needs to be taken care of in the MOBIKE
protocol is also various flooding attacks. the "flooding attack" type. See [I-D.ietf-mip6-ro-sec]
and [Aur02] for more information about flooding attacks.

[EDITOR's NOTE: This section needs more work.]

7. IANA Considerations

This document does not introduce any IANA considerations.

8. Acknowledgments

This document is the result of discussions in the MOBIKE working
group. The authors would like to thank Jari Arkko, Pasi Eronen,
Francis Dupont, Mohan Parthasarathy, Paul Hoffman, Bill Sommerfeld,
James Kempf, Vijay Devarapalli, Atul Sharma, Bora Akyol, Joe Touch,
Udo Schilcher, Tom Henderson Henderson, Andreas Pashalidis and Maureen Stillman
for their input.

We would like to particularly thank Pasi Eronen for tracking open
issues on the MOBIKE mailing list. He helped us to make good
progress on the document.

9. Open Issues

This document is not yet complete, the following open issues need to
be addressed in a future version:

o Section 4 needs an example to better illustrate the functionality
of MOBIKE

o Section 6 requires a more detailed discussion about the potential
security vulnerabilities and their solution. corresponding countermeasures.

o Some text is need needed to address the implications of unidirectional
connectivity aspect for MOBIKE (see also issue #19).

o A discussion about the scalability aspects of connectivity tests
would be benefical.

o More details are necessary to describe the implications of NAT
traversal for MOBIKE.

A complete list of issues is available at TBD.
http://www.vpnc.org/ietf-mobike/issues.html.

10. References

10.1 Normative references

[I-D.ietf-ipsec-ikev2]
Kaufman, C., "Internet Key Exchange (IKEv2) Protocol",
Internet-Draft draft-ietf-ipsec-ikev2-17,
draft-ietf-ipsec-ikev2-17 (work in progress),
October 2004.

[I-D.ietf-ipsec-rfc2401bis]
Kent, S. and K. Seo, "Security Architecture for the
Internet Protocol",
Internet-Draft draft-ietf-ipsec-rfc2401bis-05, December
2004. draft-ietf-ipsec-rfc2401bis-06 (work
in progress), April 2005.

10.2 Informative References

[I-D.arkko-multi6dt-failure-detection]
Arkko, J., "Failure Detection and Locator Selection in
Multi6",
Internet-Draft draft-arkko-multi6dt-failure-detection-00, draft-arkko-multi6dt-failure-detection-00 (work
in progress), October 2004.

[RFC2409] Harkins, D. and D. Carrel, "The Internet Key Exchange
(IKE)", RFC 2409, November 1998.

[RFC2401] Kent, S. and R. Atkinson, "Security Architecture for the
Internet Protocol", RFC 2401, November 1998.

[I-D.dupont-mipv6-3bombing]
Dupont, F., "A note about 3rd party bombing in Mobile
IPv6", Internet-Draft draft-dupont-mipv6-3bombing-01,
January draft-dupont-mipv6-3bombing-02 (work in progress),
June 2005.

[I-D.ietf-mip6-ro-sec]
Nikander, P., "Mobile IP version 6 Route Optimization
Security Design Background",
Internet-Draft draft-ietf-mip6-ro-sec-02, October 2004. draft-ietf-mip6-ro-sec-03
(work in progress), May 2005.

[I-D.ietf-hip-mm]
Nikander, P., "End-Host Mobility and Multi-Homing with
Host Identity Protocol",
Internet-Draft draft-ietf-hip-mm-00, October 2004. draft-ietf-hip-mm-01 (work in
progress), February 2005.

[I-D.crocker-celp]
Crocker, D., "Framework for Common Endpoint Locator
Pools", Internet-Draft draft-crocker-celp-00, draft-crocker-celp-00 (work in progress),
February 2004.

[RFC3489] Rosenberg, J., Weinberger, J., Huitema, C. C., and R. Mahy,
"STUN - Simple Traversal of User Datagram Protocol (UDP)
Through Network Address Translators (NATs)", RFC 3489,
March 2003.

[RFC2960] Stewart, R., Xie, Q., Morneault, K., Sharp, C.,
Schwarzbauer, H., Taylor, T., Rytina, I., Kalla, M.,
Zhang, L. L., and V. Paxson, "Stream Control Transmission
Protocol", RFC 2960, October 2000.

[RFC3753] Manner, J. and M. Kojo, "Mobility Related Terminology",
RFC 3753, June 2004.

[I-D.ietf-tsvwg-addip-sctp]
Stewart, R., "Stream Control Transmission Protocol (SCTP)
Dynamic Address Reconfiguration",
Internet-Draft draft-ietf-tsvwg-addip-sctp-10, January
draft-ietf-tsvwg-addip-sctp-12 (work in progress),
June 2005.

[I-D.dupont-ikev2-addrmgmt]
Dupont, F., "Address Management for IKE version 2",
Internet-Draft draft-dupont-ikev2-addrmgmt-06, October
2004.
draft-dupont-ikev2-addrmgmt-07 (work in progress),
May 2005.

[RFC3554] Bellovin, S., Ioannidis, J., Keromytis, A. A., and R.
Stewart, "On the Use of Stream Control Transmission
Protocol (SCTP) with IPsec", RFC 3554, July 2003.

[I-D.ietf-ipv6-optimistic-dad]
Moore, N., "Optimistic Duplicate Address Detection for
IPv6", Internet-Draft draft-ietf-ipv6-optimistic-dad-05, draft-ietf-ipv6-optimistic-dad-05 (work in
progress), February 2005.

[I-D.ietf-ipv6-unique-local-addr]
Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast
Addresses",
Internet-Draft draft-ietf-ipv6-unique-local-addr-09, draft-ietf-ipv6-unique-local-addr-09 (work in
progress), January 2005.

[RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G. G., and
E. Lear, "Address Allocation for Private Internets",
BCP 5, RFC 1918, February 1996.

[RFC2367] McDonald, D., Metz, C. C., and B. Phan, "PF_KEY Key
Management API, Version 2", RFC 2367, July 1998.

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

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

[Aur02] Aura, T., Roe, M. M., and J. Arkko, "Security of Internet
Location Management", In Proc. 18th Annual Computer
Security Applications Conference, pages 78-87, Las Vegas,
NV USA, December 2002.

Authors' Addresses

Tero Kivinen
Safenet, Inc.
Fredrikinkatu 47
HELSINKI FIN-00100
FI

Email: kivinen@safenet-inc.com

Hannes Tschofenig
Siemens
Otto-Hahn-Ring 6
Munich, Bavaria 81739
Germany

Email: Hannes.Tschofenig@siemens.com
URI: http://www.tschofenig.com

Intellectual Property Statement

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

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

The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights that may cover technology that may be required to implement
this standard. Please address the information to the IETF at
ietf-ipr@ietf.org.

Disclaimer of Validity

This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Copyright (C) The Internet Society (2005). This document is subject
to the rights, licenses and restrictions contained in BCP 78, and
except as set forth therein, the authors retain all their rights.

Acknowledgment

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