--- 1/draft-ietf-mobike-design-03.txt 2006-02-04 17:26:46.000000000 +0100 +++ 2/draft-ietf-mobike-design-04.txt 2006-02-04 17:26:47.000000000 +0100 @@ -1,19 +1,19 @@ IKEv2 Mobility and Multihoming T. Kivinen (mobike) Safenet, Inc. Internet-Draft H. Tschofenig -Expires: January 19, 2006 Siemens - July 18, 2005 +Expires: April 24, 2006 Siemens + October 21, 2005 Design of the MOBIKE Protocol - draft-ietf-mobike-design-03.txt + draft-ietf-mobike-design-04.txt Status of this Memo 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 becomes aware will be disclosed, in accordance with 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 @@ -24,21 +24,21 @@ 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 January 19, 2006. + This Internet-Draft will expire on April 24, 2006. Copyright Notice Copyright (C) The Internet Society (2005). Abstract The MOBIKE (IKEv2 Mobility and Multihoming) working group is developing extensions for the Internet Key Exchange Protocol version 2 (IKEv2). These extensions should enable an efficient management of @@ -47,51 +47,60 @@ (for example, due to mobility). This document discusses the involved network entities, and the relationship between IKEv2 signaling and information provided by other protocols. Design decisions for the 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 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 . . . . . . . . . 20 - 5.6 Return Routability Tests . . . . . . . . . . . . . . . . . 20 - 5.7 Employing MOBIKE results in other protocols . . . . . . . 23 - 5.8 Scope of SA changes . . . . . . . . . . . . . . . . . . . 24 - 5.9 Zero Address Set . . . . . . . . . . . . . . . . . . . . . 25 - 5.10 IPsec Tunnel or Transport Mode . . . . . . . . . . . . . . 25 - 5.11 Message Representation . . . . . . . . . . . . . . . . . . 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 . . . . . . . . . . . . . . . . . . . . . . . . . 4 + 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 5 + 3. Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . 8 + 3.1. Mobility Scenario . . . . . . . . . . . . . . . . . . . . 8 + 3.2. Multihoming Scenario . . . . . . . . . . . . . . . . . . . 9 + 3.3. Multihomed Laptop Scenario . . . . . . . . . . . . . . . . 10 + 4. Scope of MOBIKE . . . . . . . . . . . . . . . . . . . . . . . 11 + 5. Design Considerations . . . . . . . . . . . . . . . . . . . . 14 + 5.1. Choosing addresses . . . . . . . . . . . . . . . . . . . . 14 + 5.1.1. Inputs and triggers . . . . . . . . . . . . . . . . . 14 + 5.1.2. Connectivity . . . . . . . . . . . . . . . . . . . . . 14 + 5.1.3. Discovering connectivity . . . . . . . . . . . . . . . 15 + 5.1.4. Decision making . . . . . . . . . . . . . . . . . . . 15 + 5.1.5. Suggested approach . . . . . . . . . . . . . . . . . . 15 + 5.2. NAT Traversal . . . . . . . . . . . . . . . . . . . . . . 16 + 5.2.1. Background and constraints . . . . . . . . . . . . . . 16 + 5.2.2. Fundamental restrictions . . . . . . . . . . . . . . . 16 + 5.2.3. Moving to behind NAT and back . . . . . . . . . . . . 16 + 5.2.4. Responder behind NAT . . . . . . . . . . . . . . . . . 17 + 5.2.5. NAT Prevention . . . . . . . . . . . . . . . . . . . . 17 + 5.2.6. Suggested approach . . . . . . . . . . . . . . . . . . 17 + 5.3. Scope of SA changes . . . . . . . . . . . . . . . . . . . 18 + 5.4. Zero address set functionality . . . . . . . . . . . . . . 19 + 5.5. Return routability test . . . . . . . . . . . . . . . . . 19 + 5.5.1. Employing MOBIKE results in other protocols . . . . . 22 + 5.5.2. Suggested approach . . . . . . . . . . . . . . . . . . 23 + 5.6. IPsec Tunnel or Transport Mode . . . . . . . . . . . . . . 23 + 6. Protocol detail issues . . . . . . . . . . . . . . . . . . . . 24 + 6.1. Indicating support for mobike . . . . . . . . . . . . . . 24 + 6.2. Path Testing and Window size . . . . . . . . . . . . . . . 25 + 6.3. Message presentation . . . . . . . . . . . . . . . . . . . 26 + 6.4. Updating address list . . . . . . . . . . . . . . . . . . 27 + 7. Security Considerations . . . . . . . . . . . . . . . . . . . 28 + 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 29 + 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 30 + 10. References . . . . . . . . . . . . . . . . . . . . . . . . . . 31 + 10.1. Normative references . . . . . . . . . . . . . . . . . . . 31 + 10.2. Informative References . . . . . . . . . . . . . . . . . . 31 + Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 34 + Intellectual Property and Copyright Statements . . . . . . . . . . 35 1. Introduction The purpose of IKEv2 is to mutually authenticate two hosts, establish one or more IPsec Security Associations (SAs) between them, and subsequently manage these SAs (for example, by rekeying or deleting). IKEv2 enables the hosts to share information that is relevant to both the usage of the cryptographic algorithms that should be employed (e.g., parameters required by cryptographic algorithms and session keys) and to the usage of local security policies, such as @@ -105,39 +114,38 @@ single pair in the outer IP headers. Existing documents make no provision to change this pair after an IKE SA is created. There are scenarios where one or both of the IP addresses of this pair may change during an IPsec session. In principle, the IKE SA and all corresponding IPsec SAs could be re-established after the IP address has changed. However, this can be problematic, as the device might be too slow for this task. Moreover, manual user interaction (for example when using SecurID cards) might be required as part of the IKEv2 authentication procedure. Therefore, an automatic - mechanism is neeed that updates the IP addresses associated with the + mechanism is need that updates the IP addresses associated with the IKE SA and the IPsec SAs. MOBIKE provides such a mechanism. The work of the MOBIKE working group and therefore this document is based on the assumption that the mobility and multi-homing extensions are developed for IKEv2 [I-D.ietf-ipsec-ikev2]. As IKEv2 is built on the architecture described in RFC2401bis [I-D.ietf-ipsec-rfc2401bis], all protocols developed within the MOBIKE working group must be compatible with both IKEv2 and the architecture described in RFC2401bis. The document does not aim to neither provide support IKEv1 [RFC2409] nor the architecture described in RFC2401 [RFC2401]. This document is structured as follows. After introducing some important terms in Section 2 a number of relevant usage scenarios are - discussed in Section 3. Section 4 discusses the interoperation of - MOBIKE with other protocols and processes that may run in the local - machine. Finally, Section 5 discusses design considerations - affecting the MOBIKE protocol. The document concludes in Section 6 - with security considerations. + discussed in Section 3. The next section Section 4 will describe the + scope of the MOBIKE protocol. Finally, Section 5 discusses design + considerations affecting the MOBIKE protocol. The document concludes + in Section 7 with security considerations. 2. Terminology This section introduces the terminology that is used in this document. Peer: A peer is an IKEv2 endpoint. In addition, a peer implements the MOBIKE extensions, as defined in this and related documents. @@ -160,23 +168,21 @@ * If the address is an IPv6 address, it is a global unicast or unique site-local address, as defined in [I-D.ietf-ipv6-unique- local-addr]. That is, it is not an IPv6 link-local. Where IPv4 is considered, it is not an RFC 1918 [RFC1918] address. * The address and interface is acceptable for sending and receiving traffic according to a local policy. This definition is taken from [I-D.arkko-multi6dt-failure- - detection] - - . + detection]. Locally Operational Address: An address is said to be locally operational if it is available and its use is locally known to be possible 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 @@ -232,33 +238,33 @@ Terminology regarding NAT types (e.g. Full Cone, Restricted Cone, Port Restricted Cone and Symmetric), can be found in Section 5 of [RFC3489]. For mobility related terminology (e.g. Make-before-break or Break-before-make) see [RFC3753]. 3. Scenarios In this section we discuss three typical usage scenarios for the MOBIKE protocol. -3.1 Mobility Scenario +3.1. Mobility Scenario Figure 1 shows a break-before-make mobility scenario where a mobile node changes its point of network attachment. Prior to the change, the mobile node had established an IPsec connection with a security gateway which offered, for example, access to a corporate network. The IKEv2 exchange that facilitated the set up of the IPsec SA(s) took place over the path labeled as 'old path'. The involved packets carried the MN's "old" IP address and were forwarded by the "old" access router (OAR) to the security gateway (GW). When the MN changes its point of network attachment, it obtains a new - IP address using statefu address configuration techniques or via the + IP address using stateful address configuration techniques or via the stateless address autoconfiguration mechanism. The goal of MOBIKE, in this scenario, is to enable the 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', enables the peers to update their state as necessary. Note that in a break-before-make scenario the MN obtains the new IP address after it can no longer be reached at the old IP address. In a make-before-break scenario, the MN is, for a given period of time, reachable at both the old and the new IP address. MOBIKE should work @@ -278,21 +284,21 @@ 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 +3.2. Multihoming Scenario Another MOBIKE usage scenario is depicted in Figure 2. In this scenario, the MOBIKE peers are equipped with multiple interfaces (and multiple IP addresses). Peer A has two interface cards with two IP addresses, IP_A1 and IP_A2, and peer B has two IP addresses, IP_B1 and IP_B2. Each peer selects one of its IP addresses as the preferred address which is used for subsequent communication. Various reasons, (e.g hardware or network link failures), may require a peer to switch from one interface to another. @@ -303,27 +309,27 @@ | | | | | | | 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 + 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 +3.3. Multihomed Laptop Scenario The third scenario we consider is about a laptop, which has multiple interface cards and therefore several ways to connect to the network. It may for example have a fixed Ethernet card, a WLAN interface, a GPRS adaptor, a Bluetooth interface or USB hardware. Not all interfaces are connected to the network at all times for a number of reasons (e.g., cost, availability of certain link layer technologies, user convenience). The mechanism that determines which interfaces are connected to the network at any given point in time is outside the scope of the MOBIKE protocol and, as such, this document. @@ -331,64 +337,55 @@ network, the set of IP addresses under which the laptop is reachable, changes too. Even if IP addresses change due to interface switching or mobility, the IP address obtained via the configuration payloads within IKEv2 remain unaffected. The IP address obtained via the IKEv2 configuration payloads allow the configuration of the inner IP address of the IPsec tunnel. As such, applications might not detect any change at all. -4. Framework +4. Scope of MOBIKE - The working group will develop a MOBIKE protocol which needs to - perform the following operations: + Getting mobility and multihoming actually working requires lots of + different components working together, including coordinating things + and making consistent decisions in several link layers, the IP + layers, different mobility mechanisms in those layers, and IPsec/IKE. + Most of those aspects are beyond the scope of MOBIKE: The MOBIKE + focuses on what two peers need to agree in IKEv2 level (like new + message formats and some aspects of their processing) for + interoperability. + + The MOBIKE is not trying to be full mobility protocol; there is no + support for simultaneous movement or rendezvous mechanism, and there + is no support for route optimization etc. This current design + document focuses mainly on the tunnel mode, everything going inside + the tunnel is unaffected by the changes in the tunnel header IP + address, and this is the mobility feature provided by the MOBIKE. + I.e. the applications running through the MOBIKE IPsec tunnel cannot + even detect the movement, their IP address etc stay constant. + + A MOBIKE protocol should be able to perform the following operations: o inform the other peer about the peer address set o inform the other peer about the preferred address o test connectivity along a path and thereby to detect an outage situation o change the preferred address o change the peer address set o Ability to deal with Network Address Translation devices - The technical details of these functions are discussed below. - Although MOBIKE will have to interact with other mechanisms, the - working group is chartered to leave this aspect outside the scope. - - When a MOBIKE peer initiates a protocol exchange it needs to define a - peer address set based on the IP addresses available to it. The peer - may want to make this set available to the other peer. The IKEv2 - Initiator does not need to indicate which of the addresses in the - peer address set is its preferred address. This is because the - Initiator has to place its preferred address into the source IP - address field of the IP header with the initial message exchange. - Additionally, the Initiator expects incoming signaling messages to - arrive at this address. The peer address set and the preferred - address are defined based on interaction with other components within - a host. In some cases, the peer address set 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 may change as well. - - A modification of the peer address set or a change of the preferred - address typically is the 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 between a pair of MOBIKE peers. MOBIKE interacts with the IPsec engine using the PF_KEY API [RFC2367]. Using this API, the MOBIKE daemon can create entries in the Security Association (SAD) and Security Policy Databases (SPD). The IPsec engine may also interact with IKEv2 and MOBIKE daemon using this API. The content of the Security Policy and Security Association Databases determines what traffic is protected with IPsec in which fashion. MOBIKE, on the other hand, receives information from a number of sources that may run both in kernel-mode and in user-mode. Information relevant for MOBIKE might be stored in @@ -452,371 +449,308 @@ Extensions of the PF_KEY interface required by MOBIKE are also within the scope of the working group. Finally, certain optimizations for wireless environments are also 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, both peers must implement the MOBIKE - extension of IKEv2. If one or none of the peers supports MOBIKE, - then, whenever an IP address changes, IKEv2 will have to be re-run in - order to create a new IKE SA and the respective IPsec SAs. In - MOBIKE, a peer needs to be 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 ensure that a peer receives feedback on whether or not its - messages are understood by the other peer, is by using IKEv2 - messaging for MOBIKE and to mark some messages as "critical". - According to the IKEv2 specification, such messages either have to be - understood by 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 by - using Vendor ID payloads that are exchanged during the initial IKEv2 - exchange. These payloads would then indicate whether or not a given - peer supports the MOBIKE protocol. +5.1. Choosing addresses - A third approach would use the Notify payload which is also used for - NAT detection (via NAT_DETECTION_SOURCE_IP and - NAT_DETECTION_DESTINATION_IP payloads). + One of the core aspects of the MOBIKE protocol is the selection of + the address for the IPsec packets we send. Choosing addresses for + the IKEv2 request is somewhat separate problem: in many cases, they + will be the same (and in some design choice they will always be the + same). - Both a Vendor ID and a Notify payload may be used to indicate the - support of certain extensions. +5.1.1. Inputs and triggers - Note that 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 an - unnecessary MOBIKE message exchange is introduced. + How the address changes are triggered are largerly beyond the scope + of MOBIKE. The triggers can include e.g. changes in the set of + addresses, various link-layer indications, failing dead peer + detection, and changes in preferences and policies. Furthermore, + there may be less reliable sources of information (such as lack of + IPsec packets and ICMP packets) that do not trigger any changes + directly, but rather cause DPD to be performed sooner than it + otherwise would have been (and if that DPD fails, that may trigger + changing of addresses). - Although Vendor ID payloads and Notifications are technically - equivalent, Notifications are already used in IKEv2 as a capability - negotiation mechanism. Hence, Notifications and Vendor ID payloads - are the preferred mechanisms. + These triggers are largerly the same as for, e.g. Mobile IP, and are + beyond the scope of MOBIKE. -5.2 Changing a Preferred Address and Multi-homing Support +5.1.2. Connectivity - From MOBIKE's point of view, support for multi-homing is inherently - provided by the fact that it manages a set of peer addresses, rather - than a single address. Further, MOBIKE provides mechanisms to change - a peer's preferred IP address. Each peer needs to learn the - preferred address and the peer address set. + There can be two kind of "failures" in the connectivity; local or + middle. Local failure is a property of an address (interface), while + the failures in the middle is property of address pair. MOBIKE does + not assume full connectivity, but it does not try to use + unidirectional address pairs (multi6 has discussed about how to deal + with unidirectional paths). Unidirectional address pairs is + complicated issue, and supporting it would require abandoning the + principle that you always send the IKEv2 reply to the address the + request came from. Because of that MOBIKE decided to deal only with + bidirectional address pairs. It does consider unidirectional address + pairs as broken, and not use them, but the connection will not break + even if unidirectional address pairs are present, provided there is + at least one bidirectional address pair it can use. -5.2.1 Storing a single or multiple addresses + Note, that MOBIKE is not really concerned about the actual path used, + it cannot even detect if some path is unidirectional, if the same + address pair has some other unidirectional path back. Ingress + filters might still cause such path pairs to be unusable, and in that + case MOBIKE will detect that there is no connection between address + pair. - One design decision is whether an IKE-SA should be associated with a - single IP address or multiple IP addresses. One option is that a - peer can provide a list of addresses to its counterpart which can - then be used as destination addresses. + In a sense having both IPv4 and IPv6 address is basically a case of + partial connectivity (putting both IPv4 and IPv6 address in the same + IP header does not work). The main difference is that it is known + beforehand, and there is no need to discover that IPv4/IPv6 + combination does not work. - 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. +5.1.3. Discovering connectivity - Another option is to only communicate one address to the other peer - and both peers only use that address when communicating. If this - address cannot be used anymore then an address update is sent to the - other peer that changes the preferred address. + To detect connectivity, i.e failures in the middle, MOBIKE needs to + have some kind of probe which it can send to the other end and get a + reply back to that. If it will see the reply it knows the connection + works, if it does not see the reply after multiple retransmissions it + may assume that the address pair tested is broken. - Alternatively, if peer A, for example,provides a peer address set - with multiple IP addresses then peer B can recover from a failure of - the preferred address 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., changing the - source IP address of outgoing MOBIKE messages) or to try an IP - address from A's peer address set (i.e., changing the destination - address). If peer B only received a single IP address from peer A - for A then peer B can only change its own preferred address. 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 connectivity tests do require to take in to account the + congestion problems, because the connection failure might be because + of congestion, and the MOBIKE should not make it worse by sending + lots of probe packets. - The main advantage of storing only a single IP address for the remote - peer is that it makes retransmission handling easier. Moreover, it - simplifies the recovery procedure. The peer whose IP address changed - must start the recovery process and send the new IP address to the - other peer. However, connectivity failures along the path are not - well addressed with advertising a single IP address. +5.1.4. Decision making - The single IP address approach does not work if both peers change - their IP addresses at the same time, for example if both hosts move - simultaneously, even though multiple addresses are available to the - two peers. The IKEv2 implementation might also require window size - to be larger than 1 because the MOBIKE peer needs to be able to send - the IP address change notifications before it switches to another - address. Depending on the occurrence of return routability checks, - retransmissions policies and similar message exchanges a window size - larger than 1 might be required to deal with 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 another IP - address, since it does not know any). + One of the core decisions to the MOBIKE protocol is to who makes the + decisions to fix situations, i.e. symmetry in decision making vs. + asymmetry in decisions. Symmetric decision making may cause the + different peers to make different decisions, thus causing asymmetric + upstream/downstream traffic. In mobility case it is desirable that + the mobile peer can move both upstream and downstream traffic to some + particular interface, and this requires asymmetric decision making. - The problems with IP address lists lie mostly in their complexity. - Notification and recovery processes are more 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 - this could cause them to lose connectivity. The MOBIKE protocol is - required to prevent this. + Working with stateful packet filters and NATs is easier if same + address pair is used in both upstream and downstream directions. + Also in common cases only the peer behind NAT can actually do actions + to recover from the connectivity problems, as it might be that the + other peer is not able to initiate any connections to the peer behind + NAT. - 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. A NAT 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 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], 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]. +5.1.5. Suggested approach -5.2.2 Indirect or Direct Indication + Because of those issues listed above, the MOBIKE protocol decided to + select method where the initiator will decide which addresses are + used. This has the benefits that it makes one peer to decide, thus + we cannot end up in the asymmetric decisions, and it also works best + with NATs, as the initiator is the most common peer to be behind NAT, + and thus is the only peer which can recover in those cases. - An indication to change the preferred IP address might be either - direct or indirect. +5.2. NAT Traversal - Direct indication: +5.2.1. Background and constraints - A direct indication means that the other peer will send an message - with the address change. This can, for example, be accomplished - by having MOBIKE sending an authenticated address update - notification with a different preferred address. + Another core aspect of the MOBIKE was the co-operation and working + with NATs. In IKEv2 the tunnel header IP addresses are not sent + inside the IKEv2 payloads, and thus there is no need to do unilateral + self-address fixing (UNSAF). The tunnel header IP addresses are + taken from the outer IP header of the IKE packets, thus they are + already processed by the NAT. - Indirect indication: + The NAT detection payloads are used to detect if the addresses in the + IP header were modified by a NAT between the peers, and that enables + UDP encapsulation of ESP packets if needed. MOBIKE is not to change + how IKEv2 NAT-T works, in particular, any kind of UNSAF or explicit + interaction with NATs (e.g. midcom or nsis natfw) are beyond the + scope. The MOBIKE will need to define how MOBIKE and NAT-T are used + together. - An indirect indication to change the preferred address can be - obtained by observing the environment. An indirect indication - might, for example, be the receipt of an ICMP message or - information about a link failure. This information should be seen - as a hint and should not cause a change of the remote peer's - preferred address. Depending on the local policy, MOBIKE 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). Authorization aims - to ensure that a particular peer is allowed to use the indicated - address. Claiming to be at an arbitrary address without - performing a return-routability test or without verifying that the - IP address is listed within a certificate opens the door for - various denial of service attacks. Hence a peer may also start a - connectivity test of this particular address. + The NAT-T support should also be optional, i.e. if the IKEv2 + implementation does not implement NAT-T, since it is not required in + some particular environment, implementing MOBIKE should not require + adding support for NAT-T as well. - If more information is available to the MOBIKE daemon then a more - intelligent decision regarding the selection of a new primary path - can be made. + The property of being behind NAT is actually property of the address + pair, thus one peer can have multiple IP-addresses and some of those + might be behind NAT and some might not be behind NAT. -5.2.3 Connectivity Tests using IKEv2 Dead-Peer Detection +5.2.2. Fundamental restrictions - 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. + There are some cases which cannot be made work with the restrictions + provided by the MOBIKE charter. One of those cases is the case where + the party "outside" a symmetric NAT changes its address to something + not known by the the other peer (and old address has stopped + working). It cannot send a packet containing the new addresses to + the peer, because the NAT does not contain the necessary state. + Furthermore, since the party behind the NAT does not know the new IP + address, it cannot cause the NAT state to be created. - The IKEv2 DPD mechanism involves sending an empty informational - exchange packet to a given address of the remote peer. On receipt, - the remote peer responds with an acknowledgement. If no - acknowledgement is received after a certain timeout period (and after - couple of retransmissions), the remote peer is considered to be not - reachable at the address in question. On the other hand, receipt of - IPsec protected acknowledgement is a guarantee that the other peer is - reachable at the address in question. + This case could be solved using some rendezvous mechanism outside + IKEv2, but that is beyond the scope of MOBIKE. - 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 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. +5.2.3. Moving to behind NAT and back - 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). + MOBIKE provides mechanism where peer not initially behind the NAT, + can move to behind NAT, when new address pair is selected. MOBIKE + does not need to detect when someone attach NAT to the currently + working address pair, i.e. the NAT detection is only done when MOBIKE + changes the address pair used. - Note that IKEv2 implementations may have a window size of 1, and - therefore be unable to initiate a dead-peer detection exchange while - another exchange is pending. As a result, all other exchanges are - subject to an identical retransmission policy as used for the dead- - peer detection. To use a different policy for different message - types seems to be reasonable. + Similarly the MOBIKE provides mechanism to move from the address pair + having NAT to the address pair not having NAT. - The dead-peer detection mechanism for the other IP address pairs can - also be executed simultaneously if the window size larger than 1, - meaning that after the initial timeout period of the preferred - address expires, DPD packets are sent simultaneously to all other - address pairs. It is necessary to differentiate acknowledgement - messages in order to determine which address pair is operational. - The source IP address of the acknowledgement can be used for this - purpose. + As we only use one address pair at time, effectively MOBIKE peer is + either behind NAT or not behind NAT, but each address change can + change the situation. Because of this and because initiator always + chooses the addresses it is enough to send keepalive packets only to + that one address pair. - The protocol should also be nice to the network, meaning, that when - some core link goes down, and a large number of MOBIKE clients notice - this, they should not start sending a large number of messages while - trying to recover from the problem. This may be particularly - unfortunate because packets may be dropped because of congestion in - the first place. If MOBIKE clients simultaneously try to test all - possible address pairs by executing connectivity tests then the - congestion problem only gets worse. +5.2.4. Responder behind NAT - Also note that the IKEv2 dead-peer detection is not sufficient for - the return routability check. See Section 5.6 for more information. + MOBIKE can work in cases where the responder is behind static NAT, + but in that case the initiator needs to know all possible addresses + where the responder can move to, i.e. responder cannot move to the + address which is not known by the initiator. -5.3 Simultaneous Movements + If the responder is behind NAPT then it might need to communicate + with the NAT to create mapping so initiator can connect to it. Those + external hole punching mechanisms are beyond the scope of MOBIKE. - MOBIKE does not aim to provide a mobility solution 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 an address that is - unknown to the other peer. Situations in which both peers move - simultaneously are outside the scope of the MOBIKE WG. MOBIKE has - not been chartered to deal with the rendezvous problem, or with the - introduction of new entities in the network. + In case the responder is behind NAPT then also finding the port + numbers used by the responder, is outside the scope of MOBIKE. - 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 (e.g., - if both nodes change their addresses at the same time). This is - something that the MOBIKE protocol must deal with. +5.2.5. NAT Prevention - Three cases can be differentiated: + One new feature created by the MOBIKE, is the NAT prevention, i.e. if + we detect NAT between the peers, we do not allow that address pair to + be used. This can be used to protect IP-addresses in cases where it + is known by the configuration that there is no NAT between the nodes + (for example IPv6, or fixed site-to-site VPN). This gives extra + protection against 3rd party bombing attacks (attacker cannot divert + the traffic to some 3rd party). This feature means that we + authenticate the IP-address and detect if they were changed. As this + is done on purpose to break the connectivity if NAT is detect, and + decided by the configuration, there is no need to do UNSAF + processing. - 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 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 additional infrastructure nodes. +5.2.6. Suggested approach - o Simultaneous changes to addresses such that at least one of the - new addresses is known to the other peer before the change - occurred. + The working group decided that MOBIKE uses NAT-T mechanisms from the + IKEv2 protocol as much as possible, but decided to change the dynamic + address update for IKEv2 packets to MUST NOT (it would break path + testing using IKEv2 packets). Working group also decided to only + send keepalives to the current address pair. - o No simultaneous changes at all. +5.3. Scope of SA changes -5.4 NAT Traversal + Most sections of this document discuss design considerations for + updating and maintaining addresses in the database entries that + relate to an IKE-SA. However, changing the preferred address also + affects the entries of the IPsec SA database. The outer tunnel + header addresses (source 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. - IKEv2 supports legacy NAT 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 modifies the source and possibly the - destination IP address. With NAPT even the transport protocol - identifiers are modified (which then requires UDP encapsulation for - exchanged IPsec protected data traffic). Therefore, the MOBIKE - daemon needs to obtain to required IP address informationfrom the IP - header (if a NAT was detected that modified the IP header) rather - than from the protected payload. This problem is not new and is an - issues of every mobility protocol where the most important - information exchanged is the IP address . + 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). 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. - 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 approaches to solve - the problem have been presented: + If IPsec SAs should be updated separately then a more efficient + format than the notification payload is needed to preserve bandwidth. + A notification payload can only store one SPI per payload. A + separate payload could have list of IPsec SA SPIs and new preferred + address. If there 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 from IKE SA, i.e. no need to store IP + addresses per IPsec SA. Note that IKEv2 [I-D.ietf-ipsec-ikev2] + already requires implementations to keep track which IPsec SAs are + created using which IKE SA. - 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 but the interaction with these protocols it outside the - scope of this document. + If we do allow each IPsec 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. - 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 the Host Identity Protocol but is not an option for - IKEv2 and 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. + 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. - o Disable NAT-T 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 environments with NATs. + MOBIKE protocol decided to move all IPsec SAs implicitly when the IKE + SA address pair changes. If more granular handling of the IPsec SA + is required, then multiple IKE SAs can be created one for each set of + IPsec SAs needed. - There is no way to distinguish the whether there is a NAT device - along the path that modifies the header information in packets or an - adversary mounting an attack. If a NAT is detected during the - creation of an IKE SA, that should automatically disable the MOBIKE - extensions and use NAT-T. +5.4. Zero address set functionality - A design question is whether NAT detection capabilities should be - enabled only during the initial IKEv2 exchange or 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 is desirable - to also support a scenario where a MOBIKE peer moves from a subnet - that is not behind a NAT to a network that is. + One of the features which is potentially useful is for the peer to + announce that it will now disconnect for some time, i.e. it will not + be reachable at all. For instance, a laptop might go to suspend + mode. In this case the 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 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. allow TCP/IP sessions + to notice the link is broken. - 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. (reference? Explanation?) + The local policy could then indicate how long the peer should allow + remote peers to remain disconnected. - Whether or not MOBIKE should support NAT traversal is one of the most - important design decisions. + From a technical point of view this feature addresses two aspects: -5.5 Changing addresses or changing the paths + o There is no need to transmit IPsec data traffic. IPsec protected + data can be dropped which saves bandwidth. This does not provide + a functional benefit, i.e., nothing breaks if this feature is not + provided. - A design decision is whether it is enough for the MOBIKE protocol to - detect dead addresses, or it also needs to detect dead paths. Dead - address detection refers to the ability to establish whether or not a - given IP address pair is operational. Dead path detection refers to - the ability to establish whether or not all possible (local/remote) - address pairs are operational (or at least find one such pair). + o MOBIKE signaling messages are also ignored. The IKE-SA must not + be deleted and the suspend functionality (realized with the zero + address set) may require the IKE-SA to be tagged with a lifetime + value since the IKE-SA should not be kept in alive for an + undefined period of time. Note that IKEv2 does not require that + the IKE-SA has a lifetime associated with it. In order to prevent + the IKE-SA from being deleted the dead-peer detection mechanism + needs to be suspended as well. - While performing just one address change is simpler, the existence of - locally operational addresses 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. + Due to the fact that this extension could be complicated and there + was no clear need for it, a first version of the MOBIKE protocol will + not provide this feature. -5.6 Return Routability Tests +5.5. Return routability test Changing the preferred address and subsequently 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 + [RFC3554] introduced 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 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 taking an authorization decision a malicious peer can @@ -845,45 +779,38 @@ may have provided an authenticated IP address. Thus it is would be possible to simply trust the MOBIKE peer to provide a proper IP address. There is no way an adversary can successfully launch an attack by injecting faked addresses since it does not 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, 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 know the identity of the peer in that - case. The identity of the IKEv2 Initiator and the IKEv2 Responder - can take various forms: IP address, FQDN, DN, email address alike - identifiers, etc. + case. - It seems that there it is also a policy issue when to schedule a - return routability test. + There is also a policy issue when to schedule a return routability + test. Before moving traffic? After moving traffic? 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 a different type of 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. The valid end point could send an - address update notification for 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 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 a peer to move the traffic to a - location where the original recipient cannot be reached. + dead-peer detection. There are potential attacks if a return + routability check does not include some kind of nonce. The valid end + point could send an address update notification for 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 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 a peer to move + the traffic to a location where the original recipient cannot be + reached. The IKEv2 NAT-T mechanism does not perform 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 an unmodified IP address reaches the other peer. Mobility environments make this attack more difficult for an adversary since it requires the adversary to be located somewhere 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 @@ -894,22 +821,22 @@ NAT-T mechanism it should protect against these attacks. There may be return routability information available from the other parts of the system too (as shown in Figure 3), but the checks done 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 located - along the path to the claimed address (). This is the basic form - of return routability test. + along the path to the claimed address. This is the basic form of + return routability test. o There is an answer from the tested address, and that answer was authenticated, integrity and replay protected. o There was an authenticated, integrity and replay protected answer from the peer, but it is not guaranteed to originate at the tested address or path to it (because the peer can construct a response without seeing the request). The return routability checks do not protect against 3rd party @@ -925,21 +852,21 @@ 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 a guarantee that there is a node at the IP address which replied to the request. There is no indication as to whether or not the reply is fresh, and whether or not the request may have been transmitted from a different source address. -5.7 Employing MOBIKE results in other protocols +5.5.1. Employing MOBIKE results in other protocols If MOBIKE has learned about new locations or verified the validity of a remote address through a return routability 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 are unaware of the outer addresses or their status. Similarly, IP layer tunnel termination at a gateway rather than a @@ -959,150 +886,178 @@ [I-D.crocker-celp] discusses the use of common locator information pools in a IPv6 multi-homing context; it assumed that both transport and IP layer solutions are be used in order to support multi-homing, and that it would be beneficial for different protocols to coordinate their results in some way, for instance by sharing throughput information of address pairs. This may apply to MOBIKE as well, assuming it co-exists with non-IPsec protocols that are faced with the same 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?) + protocol design and may be addressed in future work. -5.8 Scope of SA changes +5.5.2. Suggested approach - Most sections of this document discuss design considerations for - updating and maintaining addresses in the database entries that - relate to an IKE-SA. However, changing the preferred address also - affects the entries of the IPsec SA database. The outer tunnel - header addresses (source 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. + MOBIKE protocol selected to use IKEv2 INFORMATIONAL exchanges as a + return routability tests, but added random cookie there to prevent + redirections done by authenticated attacker. Return routability + tests are done by default before moving the traffic. However these + tests are optional. Nodes MAY also perform these tests upon their + own initiative at other times. - 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. + It is worth noting that the return routability test in MOBIKE is not + he same as return routability test in MIP6: The MIP6 WG decided that + it is not necessary to do return routability tests between the mobile + node and the home agent at all. - If IPsec SAs should be updated separately then a more efficient - format than the notification payload is 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 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. +5.6. IPsec Tunnel or Transport Mode - If we do allow each IPsec 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. + Current MOBIKE design is focused only on the VPN type usage and + tunnel mode. Transport mode behavior would also be useful, but will + be discussed in future documents. - 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. +6. Protocol detail issues -5.9 Zero Address Set +6.1. Indicating support for mobike - One of the features which is potentially useful is for the peer to - announce that it will now disconnect for some time, i.e. it will not - be reachable at all. For instance, a laptop might go to suspend - mode. In this case the 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 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. allow TCP/IP sessions - to notice the link is broken. + In order for MOBIKE to function, both peers must implement the MOBIKE + extension of IKEv2. If one or none of the peers supports MOBIKE, + then, whenever an IP address changes, IKEv2 will have to be re-run in + order to create a new IKE SA and the respective IPsec SAs. In + MOBIKE, a peer needs to be 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. - The local policy could then indicate how long the peer should allow - remote peers to remain disconnected. + One way to ensure that a peer receives feedback on whether or not its + messages are understood by the other peer, is by using IKEv2 + messaging for MOBIKE and to mark some messages as "critical". + According to the IKEv2 specification, such messages either have to be + understood by the receiver, or an error message has to be returned to + the sender. - From a technical point of view this feature addresses two aspects: + A second way to ensure receipt of the above-mentioned feedback is by + using Vendor ID payloads that are exchanged during the initial IKEv2 + exchange. These payloads would then indicate whether or not a given + peer supports the MOBIKE protocol. - 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., nothing breaks if this feature - is not provided. + A third approach would use the Notify payload which is also used for + NAT detection (via NAT_DETECTION_SOURCE_IP and + NAT_DETECTION_DESTINATION_IP payloads). - o MOBIKE signaling messages are also ignored. The IKE-SA must not - be deleted and the suspend functionality (realized with the zero - address set) may require the IKE-SA to be tagged with a lifetime - value since the IKE-SA should not be kept in alive for an - undefined period of time. Note that IKEv2 does not require that - the IKE-SA has a lifetime associated with it. In order to prevent - the IKE-SA from being deleted the dead-peer detection mechanism - needs to be suspended as well. + Both a Vendor ID and a Notify payload may be used to indicate the + support of certain extensions. - Due to the fact that this extension would be complicated, a first - version of the MOBIKE protocol will not provide this feature. + Note that 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 an + unnecessary message exchange is introduced. -5.10 IPsec Tunnel or Transport Mode + Although Vendor ID payloads and Notifications are technically + equivalent, Notifications are already used in IKEv2 as a capability + negotiation mechanism. Hence, notification payloads are used in the + MOBIKE to indicate support of it. - 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. + Also as the information of the support of the MOBIKE is not needed + during the IKE_SA_INIT exchange, the indication of the support is + done inside the IKE_AUTH exchange. The reason for this is to need to + keep the IKE_SA_INIT messages as small as possible, so they do not + get fragmented. The idea is that responder can do stateless + processing of the first IKE_SA_INIT packet, and request cookie from + the other end if it is under attack. To mandate responder to be able + to reassemble initial IKE_SA_INIT packets would not allow fully + stateless processing of the initial IKE_SA_INIT packets. -5.11 Message Representation +6.2. Path Testing and Window size + + As the IKEv2 has the window of outgoing messages, and the sender is + not allowed to violate that window (meaning, that if the window is + full, then he cannot send packets), it do cause some complications to + the path testing. The another complication created by IKEv2 is that + once the message is first time sent to the other end, it cannot be + modified in its future retransmissions. This makes it impossible to + know what packet actually reached first to the other end. We cannot + use IP headers to find out which packet reached the other end first, + as if responder gets retransmissions of the packet it has already + replied (and those replies might have been lost due unidirectional + address pair), it will retransmit the previous reply using the new + address pair of the request. Because of this it might be possible + that the responder has already used the IP-address information from + the header of the packet, and the reply packet ending up to the + initiator has different address pair. + + Another complication comes from the NAT-T. The current IKEv2 + document says that if NAT-T is enabled the node not behind NAT SHOULD + detect if the IP-address changes in the incoming authenticated + packets, and update the remote peers addresses accordingly. This + works fine with the NAT-T, but it causes some complications in the + MOBIKE, as it needs an ability to probe the another address pairs, + without breaking the old one. + + One approach to fix those would be to add completely new protocol + that is outside the IKE SA message id limitations (window code), + outside identical retransmission requirements, and outside the + dynamic address updating of the NAT-T. + + Another approach is to make the protocol so that it does not violate + window restrictions and does not require changing the packet is sent, + and change the dynamic address updating of NAT-T to MUST NOT in case + MOBIKE is used. To not to violate window restrictions, it means that + the addresses of the currently ongoing exchange needs to be changed, + to test different paths. To not to require changing the packet after + it is first sent, requires that the protocol needs to restart from + the beginning in case packet was retransmitted to different addresses + (so sender does not know which packet was the one that responder got + first, i.e. which IP-addresses it used). + + MOBIKE protocol decided to use normal IKEv2 exchanges for the path + testing, and decided to change the dynamic address updating of NAT-T + to MUST NOT. + +6.3. Message presentation The 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 incorporate the functionality already. Another question is the format of the address update notifications. The address update notifications can include multiple addresses, of which some may be IPv4 and some IPv6 addresses. The number of addresses is most likely going to be limited in typical environments (with less than 10 addresses). The format may need to indicate a preference value for each address. The format could either contain a preference number that determines the relative order of the addresses, or it could simply be ordered, according to preference, list of IP addresses. While two addresses can have the same preference value an ordered list avoids this situation. Even if load balancing is currently outside the scope of MOBIKE, - future work might include. The selected format needs to be flexible - enough to include additional information (e.g. to enable load - balancing). This may be realized with an reserved field, which can - later be used to store additional information. As there may arise - other information which may have to be tied to an address in the - future, a reserved field seems like a prudent design in any case. + future work might include support for it. The selected format needs + to be flexible enough to include additional information (e.g. to + enable load balancing). This may be realized with an reserved field, + which can later be used to store additional information. As there + may arise other information which may have to be tied to an address + in the future, a reserved field seems like a prudent design in any + case. There are two formats that place IP address 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.. It also offers an easy + the normal IKEv2 payload parsing 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 IKEv2 payloads that are longer than a certain threshold, but if the sender sends them in the most preferred first, the receiver can only use the first addresses. Having all IP addresses in one big MOBIKE specified internal format @@ -1112,108 +1067,99 @@ 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 + MOBIKE decided to use IKEv2 NOTIFY payloads, and put only one data + item per notify, i.e. there will be one NOTIFY payload for each item + to be sent. + +6.4. Updating address list + + Because of the initiator decides, the initiator needs to know all the + addresses used by the responder. The responder needs also that list + in case it happens move to the address unknown by the initiator, and + needs to send address update notify to the initiator, and it might + need to try different addresses for the initiator. + + MOBIKE could send the full peer address list every time any 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. + MOBIKE decided to use protocol format, where both ends can send full + list of their addresses to the other end, and that list overwrites + the previous list. To support NAT-T the IP-addresses of the received + packet is added to the list (and they are not present in the list). -6. Security Considerations +7. Security Considerations As all the messages are already authenticated by the IKEv2 there is no problem that any attackers would modify the 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, and that do not cause any direct actions. 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. One type of attack that needs to be taken care of in the MOBIKE - protocol is the "flooding attack" type. See [I-D.ietf-mip6-ro-sec] + protocol is the 'flooding attack' type. See [I-D.ietf-mip6-ro-sec] and [Aur02] for more information about flooding attacks. -7. IANA Considerations +8. IANA Considerations This document does not introduce any IANA considerations. -8. Acknowledgments +9. 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, 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 corresponding countermeasures. - - o Some text is 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 - http://www.vpnc.org/ietf-mobike/issues.html. - 10. References -10.1 Normative references +10.1. Normative references [I-D.ietf-ipsec-ikev2] Kaufman, C., "Internet Key Exchange (IKEv2) Protocol", 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", draft-ietf-ipsec-rfc2401bis-06 (work in progress), April 2005. -10.2 Informative References +10.2. Informative References [I-D.arkko-multi6dt-failure-detection] Arkko, J., "Failure Detection and Locator Selection in Multi6", 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 @@ -1223,23 +1169,23 @@ Dupont, F., "A note about 3rd party bombing in Mobile IPv6", 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", 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", draft-ietf-hip-mm-01 (work in - progress), February 2005. + Nikander, P., "End-Host Mobility and Multihoming with the + Host Identity Protocol", draft-ietf-hip-mm-02 (work in + progress), July 2005. [I-D.crocker-celp] Crocker, D., "Framework for Common Endpoint Locator Pools", draft-crocker-celp-00 (work in progress), February 2004. [RFC3489] Rosenberg, J., Weinberger, J., Huitema, C., and R. Mahy, "STUN - Simple Traversal of User Datagram Protocol (UDP) Through Network Address Translators (NATs)", RFC 3489, March 2003. @@ -1262,22 +1208,22 @@ Dupont, F., "Address Management for IKE version 2", draft-dupont-ikev2-addrmgmt-07 (work in progress), May 2005. [RFC3554] Bellovin, S., Ioannidis, J., Keromytis, 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", draft-ietf-ipv6-optimistic-dad-05 (work in - progress), February 2005. + IPv6", draft-ietf-ipv6-optimistic-dad-06 (work in + progress), September 2005. [I-D.ietf-ipv6-unique-local-addr] Hinden, R. and B. Haberman, "Unique Local IPv6 Unicast Addresses", draft-ietf-ipv6-unique-local-addr-09 (work in progress), January 2005. [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and E. Lear, "Address Allocation for Private Internets", BCP 5, RFC 1918, February 1996.