Distributed Mobility Management (DMM) M. Liebsch Internet-Draft NEC Intended status: Informational 8 July 2024 Expires: 9 January 2025 Mobile Traffic Steering draft-liebsch-dmm-mts-00.txt Abstract The evolution of cellular mobile communication systems is aligned with an increasing demand for customized deployments, energy efficiency, dynamic re-configurability and the integration and use of other network technologies, such as non-cellular radio access technologies and non-terrestrial networks. In order to achieve and maintain the expected service quality and continuity, such systems should be designed and controllable end-to-end, taking all involved network domains and segments into account. This document discusses an end-to-end system from an advanced use cases perspective and substantiates the demand for solutions to share information and enable control interfaces between all connected network domains, including the mobile communication system and the transport network that stretches up to the data networks that host service instances. Two architectural principles are described and discussed in the view of existing or new IETF technology that enables end-to-end mobile traffic treatment and steering in such complex and dynamically changing networks. 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 https://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 9 January 2025. Liebsch Expires 9 January 2025 [Page 1] Internet-Draft Mobile Traffic Steering July 2024 Copyright Notice Copyright (c) 2024 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 (https://trustee.ietf.org/ license-info) in effect on the date of 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 Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License. Table of Contents 1. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 2 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 3. Reference Architecture in the view of advanced end-to-end operations . . . . . . . . . . . . . . . . . . . . . . . 5 4. System Evolution and Use Cases . . . . . . . . . . . . . . . 6 4.1. General directions and impact . . . . . . . . . . . . . . 6 4.2. MCS proactive UPA relocation . . . . . . . . . . . . . . 8 4.3. MCS reactive UPA relocation . . . . . . . . . . . . . . . 9 4.4. DPN ephemerality . . . . . . . . . . . . . . . . . . . . 9 4.5. Communication between two mobile users . . . . . . . . . 9 5. Framework and Deployment Options . . . . . . . . . . . . . . 10 5.1. Mobile User Plane and Data Plane aspects . . . . . . . . 10 5.2. Dedicated Control Plane . . . . . . . . . . . . . . . . . 11 5.3. Decentralized Control Plane . . . . . . . . . . . . . . . 12 6. Design Recommendations and Information Models . . . . . . . . 12 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 8. Security Considerations . . . . . . . . . . . . . . . . . . . 13 9. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 13 10. References . . . . . . . . . . . . . . . . . . . . . . . . . 13 10.1. Normative References . . . . . . . . . . . . . . . . . . 13 10.2. Informative References . . . . . . . . . . . . . . . . . 13 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 14 1. Terminology The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here. Liebsch Expires 9 January 2025 [Page 2] Internet-Draft Mobile Traffic Steering July 2024 2. Introduction The evolution of cellular mobile communication systems resulted in a clear separation of control plane and user plane functions. The control plane comprises functions for security, mobile subscriber management, handover, and mobility management. The user plane functions represent anchor points for a user's traffic and enforces policies to user plane traffic, such as for traffic engineering or chargeable event monitoring and reporting. Compared to early standards, today's mobile communication systems support the demand of mobile subscribers as well as mobile network operators better in terms of deployment options and route optimization. This includes the decentralized deployment and selection of user plane anchors, which can, e.g., be topologically located on the path between a mobile user and a user's currently connected and used application service. Such flexible deployment of user plane anchors is aligned with a raising interest in distributing compute resources. Edge Computing enables the provisioning of compute resources, which are topologically close to mobile users. This helps on the one hand side to reduce end-to-end latency between a mobile user and a service or a compute resource, that performs a certain computation task for the user. On the other hand, it keeps data for certain use cases local, which can be leveraged for certain analytics tasks, where data is only of local interest, or for meeting some privacy and security objectives. Figure 1 depicts the end-to-end view of a mobile user, which connects to an application service function (ASF) through a mobile communication system (MCS). The ASF serves a user's request at the data plane layer, while the service may have a dedicated application control function (ACF). The user, as an example, may connect to the ACF to configure the service and gets served by the associated ASF. The service components are deployed in a data network, which may be represented by a central cloud, or a distributed data center with cloud computing resources. In alignment with mobile communication system standards, the MCS is separated into a radio access network (RAN) part and a mobile core network, which comprises the MCN control plane and the user plane anchors. The RAN offers cellular and non- cellular radio access technologies, such as WiFi, to a mobile user to connect to the MCS for mobile access to a target service. While mobile communication standards focus on the components of the MCS and their operation, details on the network in between the MCS and a data network are often treated out-of-scope of the mobile communication standard and deployment specific. Such network may indeed be deployment specific and differs in the type of network Liebsch Expires 9 January 2025 [Page 3] Internet-Draft Mobile Traffic Steering July 2024 nodes and protocols that route or switch traffic in between the MCS and the data network. Figure 1 depicts this network as a set of data plane nodes (DPN), which can, for example, be routers with redundant paths and implementing MPLS or SRv6 in their transport plane. This document addresses relevant system components in an end-to-end reference architecture to discuss advanced use cases and associated deployment options. In particular use cases are addressed, where components or network functions on the end-to-end path change, move or discontinue operation while the mobile user is still connected to the service. The objective of this document is to analyze these use cases and elaborate how existing or new IETF technology can serve as enabler to accomplish service continuity for a mobile user in such agile and dynamic system by means of controlled steering and treatment of a mobile user's end-to-end traffic. Each case is analyzed and discussed in the view of technical challenges and operational aspects in each of the two described solution options. Advanced use cases are in-line with the view in industry, research community as well as pre-standards effort. The following deployment- and operational aspects of session and service continuity are included: 0 Mid-session relocation of a mobile user's UPA, e.g. due to user mobility, UPA failover or load balancing. o Deployments with moving or ephemeral system-relevant nodes on the end-to-end path. These nodes include system components, such as radio access network components and a MCS's UPA that are on- board of a moving resource, such as a low earth orbit (LEO) satellite, or an energy constrained node whose schedule enforces a power save- or inactive mode. o Mid-session relocation of a mobile user's serving data network, e.g. due to the service resource's mobility, service failover, energy/costs or quality of service reasons. This document includes the description of two solution options, that complement a MCS without interference by means of well inter- connected and collaborative control- and data planes. Operational aspects of the two solution options are described and semantics as well as information models, that apply to relevant reference points, are specified. Liebsch Expires 9 January 2025 [Page 4] Internet-Draft Mobile Traffic Steering July 2024 3. Reference Architecture in the view of advanced end-to-end operations Figure 1 depicts a reference architecture for end-to-end operations, which includes communication between a mobile user and an Application Service Function (ASF) as well as user-to-user communications. The ACF can be a service control instance hosted in a central or a distributed cloud resources, or a workload placed by the mobile user to an assigned compute resource in either a central or a decentralized cloud. While the MCS ensures mobile connectivity and data services between a mobile user and its UPA, a Transport Network comprises a network of data plane nodes (DPN) and ensures routing of a mobile user's traffic between it's UPA and one or multiple data networks. The Transport Network may implement redundant paths and select the most suitable route based on the topological location and associated IP address of the ASF and mobile user, respectively. The mobile user's IP address may be topologically correct and fit the UPA's network, or it does not match the network where the mobile user's assigned UPA is located. The latter case applies, for example, to mobile user subscriptions, which have static IP addresses assigned and represents a view that the IETF's Distributed Mobility Management (DMM) group considers for a deployment of distributed UPAs. Routing of topologically incorrect mobile device IP addresses can be tackled by overlay transport mechanisms, such as encapsulation or label switching, or host routes that apply to the relevant DPNs in the transport network in between the mobile user's UPA and its connected data networks. With reference to Figure 1, the DPNs that are depicted close to the data network and to the MCS can represent the associated domains' PE router, or be independent routers in the domain of the transport network provider that connect to the PE routers of the MCS domain and the data network domain respectively. The reference architecture's domains, the MCS, the transport network and the data network, may share the same AS or be located in different ASs. The relevance and scope of solutions for mobile traffic steering in advanced and anticipated use cases, that this document describes in Section 4, depend on the involved ASs. This aspect will be addressed in Section 6 Liebsch Expires 9 January 2025 [Page 5] Internet-Draft Mobile Traffic Steering July 2024 +----------MCS----------+ Transport Network +--Data Network--+ | | | | | +-------------------+ | +---+ | +---+ | | | MCS Control Plane | | +---+DPN+---+ | |ACF| | | +--+-------------+--+ | | +---+ | | +-+-+ | | | | | | | | | | Mobile | +--+--+ +--+--+ | +-+-+ +-+-+ | +-+-+ | User-----+ RAN +===/===+ UPA +----+DPN+-------+DPN+--+-----+ASF| | | +-----+ +-----+ | +---+ +---+ | +---+ | | | | | +-----------------------+ +----------------+ Figure 1: Reference End-to-End Architecture. 4. System Evolution and Use Cases 4.1. General directions and impact While recent standards for MCSs are still being evolved and extended, the research community and standards organizations started to elaborate on a new MCS generation and advanced use cases. Key objectives include tighter integration of a variety of radio access technologies, including non-terrestrial networks, e.g. satellites, support of energy-efficiency schemes, and runtime changes of service instances and network functions. Customized deployments place a set of UPAs topologically close to or even inside distributed data networks, which may host compute resources for service instances or temporary workload that has been offloaded from a mobile user. Liebsch Expires 9 January 2025 [Page 6] Internet-Draft Mobile Traffic Steering July 2024 In case of user mobility, the MCS may assign a new UPA to the mobile device, which can lead to sub-optimal routes between a data network and the mobile user. While the relocation of the UPA is handled by the MCS, awareness of such change is required in the transport network and the data network to update traffic treatment. For efficiency reasons, even the use of temporarily available and ephemeral resources is considered, which includes, for example, energy constrained or solar powered resources, mobile resources, such as vehicles, drones or satellites. Industry develops satellites with radio units to enable inter-satellite links and to connect to ground stations as well as mobile users on earth. Recent directions consider regenerative payload on satellites, which goes beyond the use of a satellite as a communication relay but offers on-board compute, storage and networking resources. This enables the deployment of functions from mobile communication system, such as a UPA, and Internet technology, such as a routing function, on satellites. Even the placement of suitable compute resources on- board of satellites is considered, which makes the pre-destined to host application service instances and other type of workload. These directions enable highly efficient deployments in terms of customization, resources usage, energy saving, communication latency, connectivity and Internet coverage, as well as workload placement and distribution. On the other hand, the dynamics of system components' availability as well as their geographic and topological location requires sharing of information between the involved system components across domains, i.e. MCS, transport network and data network. Exposing such information and offering control interfaces between these domains allows timely configuration of alternative paths to continue a service and steer mobile data plane traffic between a mobile user and its service or between two mobile users. In the discussed end-to-end system, a solution should also enable traffic engineering in the traversed network segments and data networks. UPA relocation and other dynamic changes in the network may result in reordering of the data plane packets. Solutions for mobile traffic treatment and steering should provide suitable enablers to mitigate the impact of such changes to packet re-ordering. [I-D.zzhang-dmm-5gdn-end-marker] describes the use of End Markers to address this problem. The following sub-sections describe selected use cases, their impact to the end-to-end system and the need for mobile traffic steering solutions. Section 5 introduces two architectural principles and options to tackle mobile traffic steering, which are subsequently discussed in the view of enabling technology, functional operation and required semantics on relevant control plane interfaces. Liebsch Expires 9 January 2025 [Page 7] Internet-Draft Mobile Traffic Steering July 2024 4.2. MCS proactive UPA relocation Figure 2 depicts a use case, where the MCS relocates the mobile user's current UPA, UPA1, to a new UPA, UPA2. The resulting route between service instance ASF1 in the data network and UPA2 may differ from the previous route to UPA1. In case the IP address of the mobile user's device needs to survive in order to not break the current session with the used service at ASF1, the transport network needs to enforce rules for traffic steering, e.g. by applying host routes on tha path's DPNs, by applying source routing or an overlay route, i.e. IP tunnel. In case the user's IP address needs to continue only as long as the data session with ASF1 takes, the MCS may assign a new, topologically correct IP address from the network of UPA2 to the mobile user device. After the mobile user terminates its data session with ASF1, new data sessions will use the new IP address and transient host routes in the transport network can be removed. Figure 2 depicts also the change of the serving ASF, which may happen also as a result of user mobility. As example, in an automotive use case, vehicles represent mobile users and connect to a most suitable, nearby ASF that is hosted by one of many available distributed cloud networks. Due to mobility, the MCS may relocate the mobile user's UPA from UPA1 to UPA2. In order to keep routing paths short and latency low, the ASF may be relocated from ASF1 to ASF2, as ASF2 is deployed in a more suitable data network. +----------MCS----------+ Transport Network +--Data Network--+ | | | | | +-------------------+ | +---+ | +---+ | | | MCS Control Plane | | +---+DPN+---+ | |ACF| | | +--+-------------+--+ | | +---+ | | +-+-+ | | | | | | | | | | Mobile | +--+--+ +---+--+ | +-+-+ +-+-+ | +--+-+ | User-----+ RAN +===/==+ UPA1 +----+DPN+-------+DPN+--+----+ASF1| | | +--+--+ +------+ | +---+ +-+-+ | +----+ | | | | | +----------------+ | | | | +----------------+ | | +------+ | +---+ +-+-+ | +----+ | | +---------+ UPA2 +----+DPN+-------|DPN|-------+ASF2| | | +------+ | +---+ +---+ | +----+ | +-----------------------+ +----------------+ Figure 2: Mid-session UPA relocation, e.g. due to user mobility. Liebsch Expires 9 January 2025 [Page 8] Internet-Draft Mobile Traffic Steering July 2024 4.3. MCS reactive UPA relocation In this case the source and rationale behind a UPF relocation in the MCS is different and results from a decision to change the serving ASF from ASF1 to ASF2 (Figure 2). This may be because of an anticipation that the targeted geographic region of the mobile user gets better service from a data network that hosts ASF2. As further example, the transport network applies changes in its routing strategy and as a result, a route between the MCS and a new ASF instance will result in better service quality and lower latency. The new situation in the service data network and the transport network can be exposed to the MCS, which may relocate the UPF from UPF1 to UPF2 to optimize the end-to-end path. 4.4. DPN ephemerality Components from the transport network or the MCS may be of ephemeral nature and discontinue service at a predictable or unpredictable point in time. Predictable discontinuity may be due to a scheduled power saving plan or mobility of the component. The latter case applies, as example, to LEO satellites in space, which are non- stationary. Advanced use cases consider the components of a radio base station and a UPA instance from the MCS as well as a DPN instance on-board of each satellite. In such case, the mobile user on earth or a High Altitude Platform Station (HAPS) is associated with the satellite's radio base station, the UPF and a DPN for further routing of the mobile traffic. At some point in time, the mobile user is not covered by the satellite's radio base station anymore and a different satellite, including its on-board radio base station, the UPF and DPN, takes over service the mobile user. The end-to-end system needs to adapt to the changed triple of a radio base station, UPA and DPN in terms of traffic steering. Taking an example from a 3GPP-based MCS, a joint deployment of a radio base station, UPA and virtual routing function can be realized per the description in [I-D.zzhang-dmm-mup-evolution]. 4.5. Communication between two mobile users A MCS provide may deploy distributed UPAs in order to shorten the route and resulting latency for the communication between two mobile users. Figure 3 depicts such case where mobile user1 has UPA1 assigned, while mobile user2 has UPA3 assigned. Mobile traffic between the two mobile users traverses UPA1 and UPA3 as well as the transport network. In case the MCS relocates the mobile user1's UPA1 to UPA2, a different route may apply. In case of demand for IP address continuity or enforcement of particular traffic engineering rules in the DPNs, the change in the UPFs must be notified to the Liebsch Expires 9 January 2025 [Page 9] Internet-Draft Mobile Traffic Steering July 2024 transport network to apply adjusted policies in the relevant DPNs. +----------MCS----------+ Transport Network | | | +-------------------+ | | | MCS Control Plane | | | +--+-------------+--+ | | | | | Mobile | +--+--+ +---+--+ | +-+-+ User1-----+ RAN +===/==+ UPA1 +----+DPN+-----+ | +--+--+ +------+ | +---+ | | |I | | | || | | | || +------+ | +---+ +-+-+ | +=========+ UPA2 +----+DPN+---|DPN| | +------+ | +---+ +-+-+ Mobile | +-----+ +------+ | +---+ | User 2----+ RAN +======+ UPA3 +----+DPN+-----+ | +-----+ +------+ | +---+ +-----------------------+ Figure 3: Mid-session UPA relocation, e.g. due to one user's mobility. 5. Framework and Deployment Options This section discusses two architecture principles and deployment options that enable end-to-end awareness of changes in the network and its configuration that have impact on a mobile user's access to a service and the expected service quality. One option leverages a transport network controller (TN Controller) that shares relevant states and information with the MCS through a control plane interface. The same interface is used to receive notification from the MCS and to apply control commands in the MCS. The second option is based on decentralized control and requires tighter coupling of the transport network's routing functions with the MCS. 5.1. Mobile User Plane and Data Plane aspects A mobile user plane applies to the scope of the MCS and is managed by the MCS control plane to enable bi-directional data traffic between a mobile user's device and its assigned UPA. The network in between the UPA and a data network includes DPNs of the transport network and PE routers. This document focuses on IETF technology that applies to the control and data plane in the transport network and data network. The transport network's data plane transits into the MCS domain at a Liebsch Expires 9 January 2025 [Page 10] Internet-Draft Mobile Traffic Steering July 2024 PE router or the UPA. 5.2. Dedicated Control Plane Figure 4 illustrates a deployment with a dedicated TN controller. The TN controller leverages an interface to the MCS control plane to receive notifications, such as during the change of a mobile user's UPA, and to apply changes in the configuration of a mobile user's states in the MCS, such as enforcement of a UPA change in alignment with a change in the transport network or data network. The data network may also use an interface either to the TN controller or to the MCS control plane, which is used to enforce re-configurations in the transport network and/or the MCS in alignment with changes in the data network or a mobile user's service, e.g. ASF relocation or QoS settings. The transport network's DPNs may make use of a dedicated PE router to reach the MCS's UPAs. In this deployment option, the control of the UPAs is clearly separated from the DPN control, though the two control planes, the MCS control plane and the TN controller, are connected through a control plane interface. The MCS control plane may enforce rules in a UPA for uplink traffic treatment, e.g. to forward a mobile user's traffic to a particular DPN. +----------MCS----------+ Transport Network +--Data Network--+ | | | | | +-------------------+ | +-------------+ | +---+ | | | MCS Control Plane +-----+TN Controller+---------+ACF| | | +--+-------------+--+ | +-------------+ | +---+ | | | | | | | | | | Mobile | +--+--+ +--+---+ | +-+-+ +-+-+ | +--+-+ | User-----+ RAN +===/==+ UPA1 +----+DPN+-------+DPN+--+----+ASF1| | | +-----+ +------+ | +---+ +-+-+ | +----+ | | || | | +----------------+ | || | : | +----------------+ | || +------+ | +-+-+ +-+-+ | +----+ | | +=========+ UPA2 +----+DPN+-------|DPN|-------+ASF2| | | +------+ | +---+ +---+ | +----+ | +-----------------------+ +----------------+ Figure 4: End-to-end architecture with a dedicated transport network controller and loose UPA-DPN coupling Liebsch Expires 9 January 2025 [Page 11] Internet-Draft Mobile Traffic Steering July 2024 5.3. Decentralized Control Plane Figure 5 illustrates a decentralized deployment without a dedicated TN controller. DPNs expose routes through a routing protocol. To cover cases where the MCS relocates a mobile user's UPA, the DPN must either tightly couple with the UPAs and apply local APIs to learn about the mobile user and it's changes configuration in the MCS, or the UPF applies a protocol to share relevant states and information with the DPN. This document does not consider a dedicated control interface between the MCS control plane and the DPN. Relevant semantic to enable the DPN to propagate updated routes towards the transport network and the data network must be available at the UPA. This may require an extension to the MCS control plane to configure the UPA with information that is relevant for mobile traffic treatment and steering in the transport network and the data network. +----------MCS----------+ Transport Network +--Data Network--+ | | | | | +-------------------+ | | +---+ | | | MCS Control Plane | | | +ACF| | | +--+-------------+--+ | | +---+ | | | | | | | | Mobile | +--+--+ +--+---+---+ +-+-+ | +--+-+ | User-----+ RAN +===/==+ UPA1 +DPN+-----------+DPN+---+----+ASF1| | | +-----+ +------+---+ +-+-+ | +----+ | | || | | +----------------+ | || | | +----------------+ | || +------+---+ +-+-+ | +----+ | | +=========+ UPA2 +DPN+-----------|DPN|--------+ASF2| | | +------+---+ +---+ | +----+ | +-----------------------+ +----------------+ Figure 5: End-to-end architecture with a routing protocol-based propagation of traffic steering policies and tight UPA-DPN coupling 6. Design Recommendations and Information Models This section will be compiled in an updated version of this draft. Details in this section depend on further analysis of operational aspects and required control plane interface semantics for enabling the different use cases in the end-to-end architecture with and without a dedicated TN controller. Liebsch Expires 9 January 2025 [Page 12] Internet-Draft Mobile Traffic Steering July 2024 Dependency on MCS standards organizations in extending control plane semantics will be addressed in this section. 7. IANA Considerations This document has no request to IANA. 8. Security Considerations TBD. 9. Acknowledgments This document includes results from discussion with various IETF and non-IETF delegates. Many thanks to Jeffrey Zhang, Jari Mutikainen, Tianji Jiang and David Lake for their immediate interest in this topic and feedback at various opportunities and through different channels. Many thanks also to Joel Halpern for his clear and apt feedback during the public side meeting@IETF118 and to Sri Gundavelli and Satoru Matsushima for their support and advice on this matter. 10. References 10.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1997, . [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, May 2017, . 10.2. Informative References [I-D.zzhang-dmm-mup-evolution] Zhang, Z. J., Patel, K., Contreras, L. M., Islam, K., Mutikainen, J., Jiang, T., Jalil, L., Sejati, O. P., and S. Zadok, "Mobile User Plane Evolution", Work in Progress, Internet-Draft, draft-zzhang-dmm-mup-evolution-09, 2 July 2024, . Liebsch Expires 9 January 2025 [Page 13] Internet-Draft Mobile Traffic Steering July 2024 [I-D.zzhang-dmm-5gdn-end-marker] Zhang, Z. J., Liebsch, M., and T. Jiang, "End Marker in 5G Data Network", Work in Progress, Internet-Draft, draft- zzhang-dmm-5gdn-end-marker-01, 6 February 2024, . Author's Address Marco Liebsch NEC Laboratories Europe GmbH Kurfuersten-Anlage 36 D-69115 Heidelberg Germany Email: marco.liebsch@neclab.eu Liebsch Expires 9 January 2025 [Page 14]