| Internet-Draft | SRv6mob-arch | February 2024 |
| Kohno, et al. | Expires 18 August 2024 | [Page] |
This document discusses the solution approach and its architectural benefits of translating mobile session information into routing information, applying segment routing capabilities, and operating in the IP routing paradigm.¶
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 18 August 2024.¶
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.¶
The current mobile user plane is defined as an overlay tunnel session to a mobile anchor point (UPF: User Plane Function in 5G context).¶
While this approach may be well suited for the use cases which require frequent mobile handover and per-session per-usage charging, it is difficult to cost-effectively and scalably address the high traffic volumes of the 5G/Beyond 5G era and more distributed data and computing demands in future.¶
The requirements for wireless systems, such as IoT and FWA (Fixed Wireless Access) applications, are becoming more diverse, and there are cases where the frequent mobile handover and per-session per-usage charging is not necessarily mandatory.¶
This document discusses the solution approach and its architectural benefits of translating mobile session information into routing information, applying segment routing capabilities, and operating in the IP routing paradigm.¶
The current tunnel session based mobile user plane has the following limitations and is getting hard to support new application requirements.¶
Mobile session information is a function of M,N (GTP-U start point and end point), whereas routing information is a function of N (destination). Therefore, for any-to-any communications, session based paradigm yields O(N^2), whereas IP routing paradigm yields O(N).¶
Edge/distributed computing can be seen as a subset of any-to-any communication. IP Routing paradigm naturally supports ubiquitous computing.¶
As for FMC/WWC, there is currently a coordinated standardization effort between 3GPP WWC [TS.23316] and BBF [BBF407]. However, the idea is to anchor even wireline traffic in the mobile packet core, which compromises simplicity and scalability.¶
In addition, the anchor point that terminates tunnel sessions becomes a scaling bottleneck.¶
The IP routing paradigm naturally removes these tunnel session based restrictions. Segment Routing enables fast protection, policy, multi-tenancy, and provide reliability and SLA differentiation.¶
This section describes the advantages of applying SRv6 mobile user plane for 5G use cases.¶
Network slicing enables network segmentation, isolation, and SLA differentiation in terms of latency and availability. End-to-end slicing will be achieved by mapping and coordinating IP network slicing, RAN and mobile packet core slicing.¶
But existing mobile user plane which is overlay tunnel does not have underlying IP network awareness, which could lead to the inability in meeting SLAs. Removing the tunnel and treating it with a IP routing paradigm simplifies the problem.¶
Segment Routing has a comprehensive set of slice engineering technologies. How to build network slicing using the Segment Routing technology is described in [I-D.ali-teas-spring-ns-building-blocks].¶
Moreover, the stateless slice identifier encoding [I-D.filsfils-spring-srv6-stateless-slice-id] can be applicable to enable per-slice forwarding policy using the IPv6 header.¶
Edge computing, where the computing workloads and datastores are placed closer to users, is recognized as one of the key pillars to meet 5G's demanding requirements, with regard to low latency, bandwidth efficiency, data locality and privacy.¶
Edge computing is more important than ever. This is because no matter how much 5G New Radio improves access speeds, it won't improve end-to-end throughput because it's largely bound to round trip delay.¶
Even with existing mobile architectures, it is possible to place UPFs in a multi-tier, or to distribute UPFs, to achieve Edge Computing. [TS.23548] and [ETSI-MEC] describes how to properly select the UPF of adequate proximity. However, complicated and signaling-heavy mechanisms are required to branch traffic or properly use different UPFs. Also, if the UPF is distributed, seamless handover has to be compromised to some extent.¶
IP Routing paradigm simply supports ubiquitous computing.¶
3GPP [TR.23725] investigates the key issues for meeting the URLLC requirements on latency, jitter and reliability in the 5G System. The solutions provided in the document are focused at improving the overlay protocol (GTP-U) and limits to provide a few hints into how to map such tight-SLA into the transport network. These hints are based on static configuration or static mapping for steering the overlay packet into the right transport SLA. Such solutions do not scale and hinder network economics.¶
Another issue that deserves special mention is the ultra-reliability issue. In order to support ultra-reliability with the tunnel session paradigm, redundant user planes paths based on dual connectivity has been proposed. The proposal has two main options.¶
In the case of the former, UE and hosts have RHF(Redundancy Handling Function). In sending, RFH is to replicate the traffic onto two GTP-U tunnels, and in receiving, RHF is to merge the traffic.¶
In the case of the latter, traffic are to be replicated and merged with the same sequence for specific QoS flow, which requires further enhancements.¶
And in either cases, the bigger problem is the lack of a reliable way for the redundant sessions to get through the disjoint path: even with the redundant sessions, if it ends up using the same infrastructure at some points, the redundancy is meaningless.¶
These issues can be solved more simply without GTP-U tunnel.¶
In addition, Segment routing has some advantages for URLLC traffic. First, traffic can be mapped to a disjoint path or low latency path as needed. Second, Segment routing provides an automated reliability protection mechanism known as TI-LFA, which is a sub-50ms FRR mechanism that provides protection regardless of the topology through the optimal backup path. It can be provisioned slice-aware.¶
Mobile networks are composed of radio, mobile packet core, and IP networks (access and backbone), with separate standard organizations and communities. Therefore, in the steady state, it is difficult to innovate to a new architecture and requires coexistence and incremental deployment.¶
[RFC9433] defines the user plane convergence between GTP-U and SRv6, so that it can co-exist with the existing user plane as needed.¶
[I-D.mhkk-dmm-srv6mup-architecture] defines the MUP architecture for Distributed Mobility Management, which can be plugged into the existing mobile service architecture. In the architecture, mobile session information is transformed to routing information, and operated in L3VPN scheme.¶
The deployment of this architecture is targeted in an administrative domain, and the functionality aimes to be domain specific.¶
This memo includes no request to IANA.¶
Authors would like to thank Satoru Matsushima, Shunsuke Homma,Yuji Tochio and Jeffrey Zhang, for their insights and comments.¶