DMM WG JC. Zuniga Internet-Draft InterDigital Intended status: Informational CJ. Bernardos Expires: January 17, 2013 UC3M T. Melia Alcatel-Lucent July 16, 2012 DMM Practices and Gap Analysis draft-zuniga-dmm-gap-analysis-01 Abstract This document describes practices for the deployment of existing mobility protocols in a distributed mobility management environment, and identifies the limitations in the current practices with respect to providing the expected functionality. The practices and gap analysis is performed for IP-based mobility protocols, dividing them into two main solution families: client- and network-based. Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. 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." This Internet-Draft will expire on January 17, 2013. Copyright Notice Copyright (c) 2012 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of Zuniga, et al. Expires January 17, 2013 [Page 1] Internet-Draft DMM Gap Analysis July 2012 publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Practices: deployment of existing solutions in a DMM environment . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.1. Client-based mobility . . . . . . . . . . . . . . . . . . 4 2.1.1. Mobile IPv6 / NEMO B.S. . . . . . . . . . . . . . . . 5 2.1.2. Mobile IPv6 Route Optimization . . . . . . . . . . . . 6 2.1.3. Hierarchical Mobile IPv6 . . . . . . . . . . . . . . . 8 2.1.4. Home Agent switch . . . . . . . . . . . . . . . . . . 9 2.1.5. Flow Mobility . . . . . . . . . . . . . . . . . . . . 9 2.1.6. Source Address selection API . . . . . . . . . . . . . 10 2.2. Network-based mobility . . . . . . . . . . . . . . . . . . 10 2.2.1. Proxy Mobile IPv6 . . . . . . . . . . . . . . . . . . 11 2.2.2. Local Routing . . . . . . . . . . . . . . . . . . . . 12 2.2.3. LMA runtime assignment . . . . . . . . . . . . . . . . 12 2.2.4. Source Address Selection . . . . . . . . . . . . . . . 13 2.2.5. Multihoming in PMIPv6 (as per RFC 5213) . . . . . . . 13 3. Gap Analysis: limitations in current practices . . . . . . . . 14 3.1. Client-based mobility . . . . . . . . . . . . . . . . . . 14 3.1.1. REQ1: Distributed deployment . . . . . . . . . . . . . 14 3.1.2. REQ2: Transparency to Upper Layers when needed . . . . 15 3.1.3. REQ3: IPv6 deployment . . . . . . . . . . . . . . . . 15 3.1.4. REQ4: Compatibility . . . . . . . . . . . . . . . . . 16 3.1.5. REQ5: Existing mobility protocols . . . . . . . . . . 16 3.1.6. REQ6: Security considerations . . . . . . . . . . . . 17 3.2. Network-based mobility . . . . . . . . . . . . . . . . . . 17 3.2.1. REQ1: Distributed deployment . . . . . . . . . . . . . 17 3.2.2. REQ2: Transparency to Upper Layers when needed . . . . 18 3.2.3. REQ3: IPv6 deployment . . . . . . . . . . . . . . . . 18 3.2.4. REQ4: Compatibility . . . . . . . . . . . . . . . . . 19 3.2.5. REQ5: Existing mobility protocols . . . . . . . . . . 19 3.2.6. REQ6: Security considerations . . . . . . . . . . . . 19 4. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 19 5. Security Considerations . . . . . . . . . . . . . . . . . . . 19 6. References . . . . . . . . . . . . . . . . . . . . . . . . . . 20 6.1. Normative References . . . . . . . . . . . . . . . . . . . 20 6.2. Informative References . . . . . . . . . . . . . . . . . . 21 Appendix A. Acknowledgments . . . . . . . . . . . . . . . . . . . 21 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 22 Zuniga, et al. Expires January 17, 2013 [Page 2] Internet-Draft DMM Gap Analysis July 2012 1. Introduction The Distributed Mobility Management (DMM) approach aims at setting up IP networks so that traffic is distributed in an optimal way and does not rely on centrally deployed anchors to manage IP mobility sessions. A first step towards the definition of DMM solutions is the definition of the problem of distributed mobility management and the identification of the main requirements for a distributed mobility management solution [I-D.ietf-dmm-requirements], which are summarized below: o REQ1: Distributed deployment. IP mobility, network access and routing solutions provided by DMM must enable a distributed deployment of mobility management of IP sessions so that the traffic can be routed in an optimal manner without traversing centrally deployed mobility anchors. o REQ2: Transparency to Upper Layers when needed. The DMM solutions must provide transparency above the IP layer when needed. Such transparency is needed, when the mobile hosts or entire mobile networks change their point of attachment to the Internet, for the application flows that cannot cope with a change of IP address. Otherwise the support to maintain a stable home IP address or prefix during handover may be declined. o REQ3: IPv6 deployment. The DMM solutions should target IPv6 as primary deployment and should not be tailored specifically to support IPv4, in particular in situations where private IPv4 addresses and/or NATs are used. o REQ4: Compatibility. The DMM solution should be able to work between trusted administrative domains when allowed by the security measures deployed between these domains. Furthermore, the DMM solution must be able to co-exist with existing network deployment and end hosts so that the existing deployment can continue to be supported. For example, depending on the environment in which DMM is deployed, the DMM solutions may need to be compatible with other existing mobility protocols that are deployed in that environment or may need to be interoperable with the network or the mobile hosts/routers that do not support the DMM enabling protocol. o REQ5: Existing mobility protocols. A DMM solution should first consider reusing and extending the existing mobility protocols before specifying new protocols. Zuniga, et al. Expires January 17, 2013 [Page 3] Internet-Draft DMM Gap Analysis July 2012 o REQ6: Security considerations. The protocol solutions for DMM must consider security, for example authentication and authorization mechanisms that allow a legitimate mobile host/ router to access to the DMM service, protection of signaling messages of the protocol solutions in terms of authentication, data integrity, and data confidentiality, opti-in or opt-out data confidentiality to signaling messages depending on network environments or user requirements. We next first analyze existing practices of deployment of IP mobility solutions in a DMM environment [I-D.perkins-dmm-matrix], [I-D.patil-dmm-issues-and-approaches2dmm]. After that, a gap analysis is conducted, identifying what can be achieved with existing solutions and what is missing in order to meet the DMM requirements identified in [I-D.ietf-dmm-requirements]. 2. Practices: deployment of existing solutions in a DMM environment This section documents practices for the deployment of existing mobility protocols in a distributed mobility management (DMM) environment. This analysis is limited in scope to existing IPv6- based mobility protocols, such as Mobile IPv6 [RFC6275], NEMO Basic Support Protocol [RFC3963], Proxy Mobile IPv6 [RFC5213], and their extensions, such as Hierarchical Mobile IPv6 [RFC5380], Mobile IPv6 Fast Handovers [RFC5568] or Localized Routing for Proxy Mobile IPv6 [I-D.ietf-netext-pmip-lr], among others. The section is divided in two parts: client-based and network-based mobility. 2.1. Client-based mobility Mobile IPv6 (MIPv6) [RFC6275] and its extension to support mobile networks, the NEMO Basic Support protocol (NEMO B.S.) [RFC3963] are the main client-based IP mobility protocols. They heavily rely on the figure of the Home Agent (HA), a centralized anchor, to provide mobile nodes (hosts and routers) with mobility support. We next describe how Mobile IPv6/NEMO and several additional protocol extensions can be deployed to meet some of the DMM requirements [I-D.ietf-dmm-requirements]. Zuniga, et al. Expires January 17, 2013 [Page 4] Internet-Draft DMM Gap Analysis July 2012 2.1.1. Mobile IPv6 / NEMO B.S. +-----+ +-----+ | CN1 | | CN2 | +-----+ +-----+ | | +------------------------------------------------+ ( ) ( ------- ------- ) ( | HA1 | MN1 operator's | HA2 | ) ( ------- core ------- ) ( ) +------------------------------------------------+ / | | \ / | | \ / | | \ / | | \ -+----- ---+--- - -+--- -----+- | AR1 | | AR2 | | AR3 | | AR4 | ---+--- ---+--- ---+--- ---+--- | | | | (o) (o) (o) (o) x x x x x x (o) (o) | | +--+--+ +--+--+ ( anchored ) | MN1 | ( anchored ) | MN2 | ( at HA1 ) +-----+ ( at HA2 ) +-----+ Figure 1: Distributed operation of Mobile IPv6 / NEMO B.S. Due to the heavy dependance on the home agent role, Mobile IPv6 and NEMO B.S. plain vanilla protocols (i.e., without additional extensions) cannot be easily deployed in a distributed fashion. One approach would be to deploy several HAs (like in Figure 1, and assign to each MN the one closest to its topological location [RFC4640], [RFC5026], [RFC6611]. In the example shown in Figure 1, MN1 is assigned HA1 (and a home address anchored by HA1), while MN2 is assigned HA2. Note that current Mobile IPv6 / NEMO B.S. specifications do not allow the use of multiple home agents by a mobile node simultaneously, and therefore the benefits of this deployment model shown here are limited. For example, if MN1 moves and attaches to AR4, the path followed by data packets would be suboptimal, as they have to traverse HA1, which is no longer close to the topological attachment point of MN1. Zuniga, et al. Expires January 17, 2013 [Page 5] Internet-Draft DMM Gap Analysis July 2012 2.1.2. Mobile IPv6 Route Optimization +-----+ +-----+ | CN1 | | CN2 | +-----+ +-----+ | | +------------------------------------------------+ ( ) ( ------- ) ( | HA1 | MN1 operator's ) ( ------- core ) ( ) +------------------------------------------------+ / | / | / | / | -+----- ---+--- | AR1 | | AR2 | ---+--- ---+--- | | (o) (o) x x MN1 AR2 HA1 CN1 CN2 x | | | | | (o) |<-------+---------------->| | RO mode | | | | | | +--+--+ |<=======+=======>|<--------------->| BT mode | MN1 | | | | | | +-----+ Figure 2: Mobile IPv6 Route Optimization One of the main goals of DMM is to avoid the suboptimal routing caused by centralized anchoring. By default, Mobile IPv6 (and NEMO B.S.) uses the so-called Bidirectional Tunnel (BT) mode, in which data traffic is always encapsulated between the MN and its HA. Mobile IPv6 also specifies the Route Optimization (RO) mode, which allows the MN to update its current location on the CNs, and then use the direct path between them An example is shown in Figure 2, in which MN1 is using BT mode with CN2 and RO mode with CN1. Note that this RO mode has several drawbacks: o The RO mode is only supported by Mobile IPv6. There is no route optimization support standardized for the NEMO B. S. protocol, although there are many different solution proposed, mainly as academic exercises. Zuniga, et al. Expires January 17, 2013 [Page 6] Internet-Draft DMM Gap Analysis July 2012 o The RO mode requires additional signaling, which adds some protocol overhead. o The signaling required to enable RO involves the home agent, and it is repeated periodically because of security reasons [RFC4225]. This basically means that the HA remains as single point of failure, because the Mobile IPv6 RO mode does not mean HA-less operation. o The RO mode requires additional support on the correspondent node (CN). Zuniga, et al. Expires January 17, 2013 [Page 7] Internet-Draft DMM Gap Analysis July 2012 2.1.3. Hierarchical Mobile IPv6 +-----+ +-----+ | CN1 | | CN2 | +-----+ +-----+ | | +------------------------------------------------+ ( ) ( ------- ) ( | HA1 | MN1 operator's ) ( ------- core ) ( ) +------------------------------------------------+ / | \ / | \ / | \ / | \ --+----- ---+---- +--+---- | MAP1 | | MAP2 | | MAP3 | ---+---- ---+---- ---+---- / \ / \ / \ / \ / \ / \ / \ / \ / \ --+-- --+-- --+-- --+-- --+-- --+-- |AR1| |AR2| |AR3| |AR4| |AR5| |AR6| ----- ----- ----- ----- ----- ----- | (o) x x MN1 AR2 MAP1 HA1 CN1 CN2 x | | | | | | (o) |<=====+======+======>+----->| | | | | | | | | +--+--+ |<=====+=====>+<------------------->| | MN1 | | | | | | | +-----+ Figure 3: Mobile IPv6 Route Optimization Hierarchical Mobile IPv6 (HMIPv6) [RFC5380] allows reducing the amount of mobility signaling as well as the overall handover performance of Mobile IPv6, by introducing a new hierarchy level to handle local mobility. The Mobility Anchor Point (MAP) entity is introduced as a local mobility handling node deployed closer to the mobile node. When HMIPv6 is used, the MN two different temporal addresses: the Regional Care-of Address (RCoA) and the Local Care-of Address (LCoA). Zuniga, et al. Expires January 17, 2013 [Page 8] Internet-Draft DMM Gap Analysis July 2012 The RCoA is anchored at one MAP, that plays the role of local home agent, while the LCoA is anchored at the access router level. The mobile node used the RCoA as the CoA signaled to its home agent. Therefore, while roaming within a local domain handled by the same MAP, the mobile node does not need to update its home agent (i.e., the mobile node does not change RCoA). The use of HMIPv6 allows some certain route optimization, as a mobile node may decide to directly use the RCoA as source address for a communication with a given correspondent node, if the MN does not expect to move outside the local domain during the lifetime of the communication. This can be seen as a potential DMM mode of operation. In the example shown in Figure 3, MN1 is using its global HoA to communicate with CN1, while it is using its RCoA to communicate with CN2. Additionally, a local domain might have several MAPs deployed, enabling different kind of HMIPv6 deployments (e.g., flat and distributed). HMIPv6 specification supports a flexible selection of the MAP (e.g., based on the distance between the MN and the MAP, taking into consideration the expected mobility pattern of the MN, etc.). 2.1.4. Home Agent switch The Home Agent switch specification [RFC5142] defines a new mobility header for signaling a mobile node that it should acquire a new home agent. Although the purposes of this specification do not include the case of changing the mobile node's home address, as that might imply loss of connectivity for ongoing connections, it could be used to force the change of home agent in those situations where there are no active sessions running that cannot cope themselves with a change of home address. 2.1.5. Flow Mobility There exist different protocols meant to support flow mobility with Mobile IPv6, namely the multiple care-of address registration [RFC5648], the flow bindings in Mobile IPv6 and NEMO B.S. [RFC6089] and the traffic selectors for flow bindings [RFC6088]. The use of these extensions allows a mobile node to associate different flows with different care-of addresses that the mobile owns at a given time. This could also be used, combined with the route optimization support, to improve the paths followed by data packets. Zuniga, et al. Expires January 17, 2013 [Page 9] Internet-Draft DMM Gap Analysis July 2012 2.1.6. Source Address selection API The IPv6 socket API for source address selection [RFC5014], [RFC3484] can be used by an application running on a mobile node to express its preference of using a home address or a care-of address in a given connection. This allows, for example, that an application which can survive an IP address change to always prefer the use of a care-of address. Similarly, and as mentioned in [RFC6275], a mobile node can also prefer the use of a care-of address for sessions that are going to finish before the mobile node hands off to a different attachment point (e.g., short-lived connections like DNS dialogs). 2.2. Network-based mobility Proxy Mobile IPv6 (PMIPv6) [RFC5213] and GPRS Tunneling Protocol (GTP) [3GPP.29.060] are the main network-based IP mobility protocols. PMIPv6 relies on the figure of the Local Mobility Anchor (LMA) to provide mobile nodes with mobility support, without requiring the involvement of the mobile nodes, and supplying the required functionality by the Mobile Access Gateway (MAG). We next describe how PMIPv6 and several additional protocol extensions can be deployed to meet some of the DMM requirements [I-D.ietf-dmm-requirements]. Zuniga, et al. Expires January 17, 2013 [Page 10] Internet-Draft DMM Gap Analysis July 2012 2.2.1. Proxy Mobile IPv6 +-----+ +-----+ | CN1 | | CN2 | +-----+ +-----+ | | +------------------------------------------------+ ( ) ( -------- -------- ) ( | LMA1 | MN1 operator's | LMA2 | ) ( -------- core -------- ) ( ) +------------------------------------------------+ / | | \ / | | \ / | | \ / | | \ --+----- ----+--- ---+---- -----+-- | MAG1 | | MAG2 | | MAG3 | | MAG4 | ---+---- ----+--- ---+---- ---+---- | | | | (o) (o) (o) (o) x x x x x x (o) (o) | | +--+--+ +--+--+ ( anchored ) | MN1 | ( anchored ) | MN2 | ( at LMA1 ) +-----+ ( at LMA2 ) +-----+ Figure 4: Distributed operation of Proxy Mobile IPv6 As with Mobile IPv6, plain Proxy Mobile IPv6 operation cannot be easily decentralized. One simple, but still suboptimal, approach would be to deploy several local mobility anchors and use a topological position based assignment to attaching mobile nodes (an example is shown in Figure 4. This assignment can be static or dynamic (as described in Section 2.2.3. The main advantage of this simple approach is that the IP address anchor (i.e., the LMA) is placed close to the mobile node, and therefore resulting paths are close-to-optimal. On the other hand, as soon as the mobile node moves, the resulting path starts to deviate from the optimal one, unless an inter-LMA mobility protocol is in place (which is missing today). Zuniga, et al. Expires January 17, 2013 [Page 11] Internet-Draft DMM Gap Analysis July 2012 2.2.2. Local Routing [I-D.ietf-netext-pmip-lr] enables optimal routing in Proxy Mobile IPv6 in three cases: two MNs attached to the same MAG and LMA, two MNs attached to different MAGs but same LMA and two MNs attached to the same MAG with different LMAs. In these three cases, data traffic between two mobile nodes does not traverse the LMA(s), thus providing some form of distribution, since the traffic is locally router at the edge. The main disadvantage of this approach is that it only tackles the MN-to-MN communication scenario, and only under certain circumstances. In the context of 3GPP the closest analogy is the use of the X2 interface between two eNBs to directly exchange data traffic during handover procedures. 3GPP does not foresee the use of local routing at any other point of the network given the structure of the EPS bearer model. 2.2.3. LMA runtime assignment [RFC6463] specifies a runtime local mobility anchor assignment functionality and corresponding mobility options for Proxy Mobile IPv6. This runtime local mobility anchor assignment takes place during a Proxy Binding Update and a Proxy Binding Acknowledgment message exchange between a mobile access gateway and a local mobility anchor. While this mechanism mainly aims for load-balancing purposes, it can also be used to select an optimal LMA from a point of view of routing. If properly complemented by an inter-LMA mobility protocol, it could also be used as part of a global DMM solution. Even without that solution, a runtime LMA assignment can be used to change the assigned LMA of an MN, for example when no session is alive (or when those running can survive an IP address change). In the context of 3GPP network similar considerations have been discussed with respect to SIPTO above the RAN or at the macro. When a MN is allowed to get SIPTO service a geographically close P-GW is selected during connection time. That is, upon establishment of the PDN connection, the MN is configured with an IP address belonging to an optimal P-GW from a routing point of view. The drawback is that if the MN moves out of the region covered by that P-GW it will need to disconnect from the network and reconnect again since no seamless session continuity is provided. In this sense applications that do not survive and IP address change will result in an unpredictable behavior. Zuniga, et al. Expires January 17, 2013 [Page 12] Internet-Draft DMM Gap Analysis July 2012 2.2.4. Source Address Selection Also in the context of network based mobility source address selection API can be considered as a mean to achieve better routing (or using different anchors) and similar considerations of section 2.1.6 apply here. For instance a MN connected to a PMIPv6 domain could attach two different wireless network interfaces to two different MAGs, hence configuring a different set of HNPs on both interfaces (potentially combining both IPv4 and IPv6). Based on application requirements or operator's policies the connection manager logic could instruct the IP stack on the MN to route selected traffic on a specific wireless interface. It should be noted that source address selection mostly provides a better routing but not session continuity in case a session should be anchored and a different LMA. In the context of 3GPP networks two ongoing study items are currently addressing the issue of selecting a wireless interface or an IP address for a specific application. The study item DIDA (Data IDentification in ANDSF) is addressing the need to map an application ID to a specific wireless interface, while the study item OPIIS (Operator Policies for IP Interface Selection) is addressing the need of selecting the right APN for a given application. Taking into account that there is a one to one link between APN and PDN connection (IP address) the second study item clearly addresses from a 3GPP perspective the same problem space as RFC 3484. 2.2.5. Multihoming in PMIPv6 (as per RFC 5213) PMIPv6 provides some multihoming support. RFC 5213 defines that the LMA can maintain one mobility session per attached interface and that upon handover the full set of HNPs can be moved to another interface in case of inter-technology handover (MAGs providing different wireless access technology) or maintained on the same interface in case of intra-technology handover (MAGs providing the same wireless access technology). A MN can also attach (as described in section 2.2.4) two different interfaces to the PMIPv6 domain, hence resulting in a multihomed device being able to send/receive traffic sequentially or simultaneously from both network interfaces. [REF to flow mob draft] extends the base RFC5213 capabilities in the sense that a mobility session can now be shared across two different access networks. It derives that a selected flow could be routed through different paths hence achieving some sort of better routing. Yet all the traffic is anchored to centralized anchor points. In the context of 3GPP networks the MAPCON feature addresses the use of multiple PDN connections, hence the use of multiple wireless interfaces either sequentially or simultaneously. Zuniga, et al. Expires January 17, 2013 [Page 13] Internet-Draft DMM Gap Analysis July 2012 3. Gap Analysis: limitations in current practices This section identifies the limitations in the current practices (documented in Section 2) with respect to the requirements listed in [I-D.ietf-dmm-requirements]. The section is also divided in two parts: client-based and network- based mobility. Each section analyzes how well the requirements listed in [I-D.ietf-dmm-requirements] are covered/met by the current practices, highlighting existing limitations and gaps. The remaining of this section will be provided in a future version of this document. 3.1. Client-based mobility 3.1.1. REQ1: Distributed deployment MIPv6 / NEMO B. S. A careful home agent deployment and policy configuration of the Mobile IPv6 / NEMO B.S. protocols can achieve some distribution. However, as soon as the mobile node moves and changes its initial attachment point, the anchors are no longer placed optimally, incurring in sub-optimal routes. If the mobile node is not expected to move within a limited area, this configuration might be considered sufficient. Otherwise, additional mechanisms to support dynamic anchoring would be needed. Mobile IPv6 RO The use of route optimization support enables a close-to anchor-less operation, which effectively can be considered as a fully distributed configuration. However, as explained before in this document, the home agent is still used for the signaling and therefore remains as a critical centralized component. Additionally, there is no RO support for network mobility standardized. HMIPv6 The use of hierarchical mobile IPv6 can be seen as a step forward compared to a careful deployment of multiple home agents and its proper configuration, as it allows a mobile node to roam within a local domain, reducing the handover latency as well as the signaling overhead. If used together with mobile IPv6, traffic still has to traverse the centralized home agent, and therefore no distributed operation is achieved. HA switch The home agent switch specification can be used to enable obtaining more benefits from a multiple-HA deployment, as the mobile node could be instructed to switch to a closer home agent. This switch can only performed without packet connectivity loss, Zuniga, et al. Expires January 17, 2013 [Page 14] Internet-Draft DMM Gap Analysis July 2012 if done at periods of time in which the mobile node does not have any active connection running, and therefore the change of home address would require sessions reestablishment. Flow mobility Previous considerations could also apply here, being the scenario extended by the use of multiple attached interfaces. SA selection API The use of proper source address selection decisions, enabled by smart connection managers, or mobility aware applications using a selection API, would allow to benefit the most from deployments exhibiting multiple anchors. 3.1.2. REQ2: Transparency to Upper Layers when needed MIPv6 / NEMO B. S. As a mobility protocol is used, the solution is transparent to the upper layers. However, as described before, this transparency comes with the cost of suboptimal routes if the mobile nodes moves away from its initial attachment point. Mobile IPv6 RO The use of the route optimization support is transparent to the upper layers. HMIPv6 The use of HMIPv6 is transparent to the upper layers. HA switch The use of the home agent switch functionality is not transparent to the upper layers, as a change of home agent normally implies a change of home address. Therefore, it is only recommended to switch home agent when there is no active session running on the mobile node. Flow mobility The use of flow mobility mechanisms is transparent to the upper layers. SA selection API The use of an intelligent source address mechanisms is transparent to the upper layers if performed by the connection manager. However if the selection is performed by the applications themselves, via the use of the API, then applications have to be mobility-aware. 3.1.3. REQ3: IPv6 deployment MIPv6 / NEMO B. S. Mobile IPv6 / NEMO B.S. protocols primarily support IPv6, although there are some extensions defined to also offer some IPv4 support. Zuniga, et al. Expires January 17, 2013 [Page 15] Internet-Draft DMM Gap Analysis July 2012 Mobile IPv6 RO Route optimization only supports IPv6. HMIPv6 HMIPv6 is only defined for IPv6. HA switch The home agent switch specification supports only IPv6, although the use of the defined mechanisms to support dual stack IPv4/IPv6 mobile nodes would also enable some IPv4 support in this case. Flow mobility Flow mobility is only defined for IPv6. SA selection API The use of source address selection mechanisms support both IPv6 and IPv4. 3.1.4. REQ4: Compatibility MIPv6 / NEMO B. S. This approach would be compatible with other protocols and work between trusted administrative domains, although as described before its operation would not provide the benefits of a fully distributed mechanism. Mobile IPv6 RO This approach would be compatible with other protocols and work between trusted administrative domains, as long as mobile IPv6 is allowed. However, as highlighted before, mobile IPv6 route optimization requires specific support on the correspondent nodes. HMIPv6 HMIPv6 is compatible with other protocols. HA switch This approach would be compatible with other protocols and work between trusted administrative domains. Flow mobility This approach would be compatible with other protocols and work between trusted administrative domains. SA selection API This approach has no impact in terms of compatibility or use between trusted administrative domains. 3.1.5. REQ5: Existing mobility protocols MIPv6 / NEMO B. S. This approach is based on existing protocols [RFC6275] and [RFC3963]. Mobile IPv6 RO This approach is based on existing protocol [RFC6275]. Zuniga, et al. Expires January 17, 2013 [Page 16] Internet-Draft DMM Gap Analysis July 2012 HMIPv6 This approach is based on existing protocol [RFC5380]. HA switch This approach is based on existing protocol [RFC5142]. Flow mobility This approach is based on existing protocols [RFC5648], [RFC6089] and [RFC6088]. SA selection API This approach is based on existing protocols [RFC3484] and [RFC5014]. 3.1.6. REQ6: Security considerations MIPv6 / NEMO B. S. This approach include security considerations. Mobile IPv6 RO This approach include security considerations. HMIPv6 This approach include security considerations. HA switch This approach include security considerations. Flow mobility This approach include security considerations. SA selection API This approach does not have security issues. 3.2. Network-based mobility 3.2.1. REQ1: Distributed deployment Local Routing As mentioned it enables optimal routing in three cases: the LMA manages the traffic of two mobile nodes connected to teh same MAG, the LMA manages the traffic of two mobile nodes connected to different MAGs, the MAG manages the traffic of two mobile nodes connected to different LMAs. LR does not consider the case where the traffic should be optimized considering different MAGs and different LMAs. Inter LMA communication is not in scope. LR only enables better routing and does not consider the distribution of mobility anchors as such. LMA Runtime Assignment The LMA runtime assignment is used to allocate an optimal LMA mostly for load balancing purposes (for instance in scenarios where LMAs run in datacenter alike infrastructure). It can be used to allocate a different LMA based on other policies such as routing although is not clear how the technology can be used to achieve distributed mobility management, especially considering scalability issues. Zuniga, et al. Expires January 17, 2013 [Page 17] Internet-Draft DMM Gap Analysis July 2012 Source Address Selection It can help in selecting a given IP source address although the specs have many limitations (for instance prefer IPv6 over IPv4, prefer HoA instead of CoA) and the socket extensions [RFC5014] require changes in the node. This solution alone is not sufficient to achieve anchors distribution in case of session continuity requirements. Multihoming in PMIPv6 As summarized in the previous section a single mobility session belongs to a single LMA (at the most the same mobility session is shared across two access networks). As of today there is no possibility to distribute anchors and to move the session between different LMAs. 3.2.2. REQ2: Transparency to Upper Layers when needed Local Routing During HO the standard mechanisms are used. In this sense if there is a MAG change while LR is enabled signaling is exchanged to inform the target MAG that upon handover LR should be re-established. The inter LMA case is not supported. For this solution the mobility context is always up, all the traffic receive seamless service. LMA Runtime Assignment Seamless support is provided as per RFC 5213. For this solution the mobility context is always up, all the traffic receive seamless service. Source Address Selection No seamless support is provided since it requires solutions such as IP flow mobility for PMIPv6. Multihoming in PMIPv6 Seamless support falls back to standard PMIPv6 operations extended for IP flow mobility support. For this solution the mobility context is always up, all the traffic receive seamless service. 3.2.3. REQ3: IPv6 deployment Local Routing It supports both IPv4 and IPv6. LMA Runtime Assignment It supports both IPv4 and IPv6. Source Address Selection It supports both IPv4 and IPv6. Multihoming in PMIPv6 It supports both IPv4 and IPv6. Zuniga, et al. Expires January 17, 2013 [Page 18] Internet-Draft DMM Gap Analysis July 2012 3.2.4. REQ4: Compatibility Local Routing Since it extends RFC 5213 compatibility with previous technology is provided. LMA Runtime Assignment Since it extends RFC 5213 compatibility with previous technology is provided. Source Address Selection To enable the full set of use cases mentioned above extensions are required thus impacting the landscape of mobile devices. The extensions should not impact the network. Multihoming in PMIPv6 Since it extends RFC 5213 compatibility is provided. 3.2.5. REQ5: Existing mobility protocols Local Routing It reuses RFC 5213. LMA Runtime Assignment It reuses RFC 5213. Source Address Selection This is terminal only. Multihoming in PMIPv6 It reuses RFC 5213. 3.2.6. REQ6: Security considerations Local Routing It reuses RFC 5213 as such same security considerations apply. LMA Runtime Assignment It reuses RFC 5213 as such same security considerations apply. Source Address Selection There is not signaling involved here. Multihoming in PMIPv6 It reuses RFC 5213 as such same security considerations apply. 4. IANA Considerations No IANA considerations. 5. Security Considerations TBD. Zuniga, et al. Expires January 17, 2013 [Page 19] Internet-Draft DMM Gap Analysis July 2012 6. References 6.1. Normative References [RFC3484] Draves, R., "Default Address Selection for Internet Protocol version 6 (IPv6)", RFC 3484, February 2003. [RFC3963] Devarapalli, V., Wakikawa, R., Petrescu, A., and P. Thubert, "Network Mobility (NEMO) Basic Support Protocol", RFC 3963, January 2005. [RFC5026] Giaretta, G., Kempf, J., and V. Devarapalli, "Mobile IPv6 Bootstrapping in Split Scenario", RFC 5026, October 2007. [RFC5142] Haley, B., Devarapalli, V., Deng, H., and J. Kempf, "Mobility Header Home Agent Switch Message", RFC 5142, January 2008. [RFC5213] Gundavelli, S., Leung, K., Devarapalli, V., Chowdhury, K., and B. Patil, "Proxy Mobile IPv6", RFC 5213, August 2008. [RFC5380] Soliman, H., Castelluccia, C., ElMalki, K., and L. Bellier, "Hierarchical Mobile IPv6 (HMIPv6) Mobility Management", RFC 5380, October 2008. [RFC5568] Koodli, R., "Mobile IPv6 Fast Handovers", RFC 5568, July 2009. [RFC5648] Wakikawa, R., Devarapalli, V., Tsirtsis, G., Ernst, T., and K. Nagami, "Multiple Care-of Addresses Registration", RFC 5648, October 2009. [RFC6088] Tsirtsis, G., Giarreta, G., Soliman, H., and N. Montavont, "Traffic Selectors for Flow Bindings", RFC 6088, January 2011. [RFC6089] Tsirtsis, G., Soliman, H., Montavont, N., Giaretta, G., and K. Kuladinithi, "Flow Bindings in Mobile IPv6 and Network Mobility (NEMO) Basic Support", RFC 6089, January 2011. [RFC6275] Perkins, C., Johnson, D., and J. Arkko, "Mobility Support in IPv6", RFC 6275, July 2011. [RFC6463] Korhonen, J., Gundavelli, S., Yokota, H., and X. Cui, "Runtime Local Mobility Anchor (LMA) Assignment Support for Proxy Mobile IPv6", RFC 6463, February 2012. Zuniga, et al. Expires January 17, 2013 [Page 20] Internet-Draft DMM Gap Analysis July 2012 [RFC6611] Chowdhury, K. and A. Yegin, "Mobile IPv6 (MIPv6) Bootstrapping for the Integrated Scenario", RFC 6611, May 2012. 6.2. Informative References [3GPP.29.060] 3GPP, "General Packet Radio Service (GPRS); GPRS Tunnelling Protocol (GTP) across the Gn and Gp interface", 3GPP TS 29.060 3.19.0, March 2004. [I-D.ietf-dmm-requirements] Chan, A., "Requirements of distributed mobility management", draft-ietf-dmm-requirements-01 (work in progress), July 2012. [I-D.ietf-netext-pmip-lr] Krishnan, S., Koodli, R., Loureiro, P., Wu, W., and A. Dutta, "Localized Routing for Proxy Mobile IPv6", draft-ietf-netext-pmip-lr-10 (work in progress), May 2012. [I-D.patil-dmm-issues-and-approaches2dmm] Patil, B., Williams, C., and J. Korhonen, "Approaches to Distributed mobility management using Mobile IPv6 and its extensions", draft-patil-dmm-issues-and-approaches2dmm-00 (work in progress), March 2012. [I-D.perkins-dmm-matrix] Perkins, C., Liu, D., and W. Luo, "DMM Comparison Matrix", draft-perkins-dmm-matrix-03 (work in progress), March 2012. [RFC4225] Nikander, P., Arkko, J., Aura, T., Montenegro, G., and E. Nordmark, "Mobile IP Version 6 Route Optimization Security Design Background", RFC 4225, December 2005. [RFC4640] Patel, A. and G. Giaretta, "Problem Statement for bootstrapping Mobile IPv6 (MIPv6)", RFC 4640, September 2006. [RFC5014] Nordmark, E., Chakrabarti, S., and J. Laganier, "IPv6 Socket API for Source Address Selection", RFC 5014, September 2007. Appendix A. Acknowledgments The work of Carlos J. Bernardos and Telemaco Melia has been partially Zuniga, et al. Expires January 17, 2013 [Page 21] Internet-Draft DMM Gap Analysis July 2012 supported by the European Community's Seventh Framework Programme (FP7-ICT-2009-5) under grant agreement n. 258053 (MEDIEVAL project). The work of Carlos J. Bernardos has also been partially supported by the Ministry of Science and Innovation of Spain under the QUARTET project (TIN2009-13992-C02-01). Authors' Addresses Juan Carlos Zuniga InterDigital Communications, LLC 1000 Sherbrooke Street West, 10th floor Montreal, Quebec H3A 3G4 Canada Email: JuanCarlos.Zuniga@InterDigital.com URI: http://www.InterDigital.com/ Carlos J. Bernardos Universidad Carlos III de Madrid Av. Universidad, 30 Leganes, Madrid 28911 Spain Phone: +34 91624 6236 Email: cjbc@it.uc3m.es URI: http://www.it.uc3m.es/cjbc/ Telemaco Melia Alcatel-Lucent Bell Labs Route de Villejust Nozay, Ile de France 91620 France Email: telemaco.melia@alcatel-lucent.com Zuniga, et al. Expires January 17, 2013 [Page 22]