Technology

Bridging 5G and IP Networks: Intel Details SRv6 MUP Technology

Author: Babu Peddu, Marketing Manager, Switch & Fabric Group, Intel Corporation

Intel

Published in

Intel Tech

4 min readJul 20, 2022

Through collaborative efforts and technology advancements, Intel’s Network Builders are working to make the 5G Stand Alone (SA) transformation a practical reality. As part of our endeavors, I recently presented at an Intel on-demand webinar with Arrcus and SoftBank called “SRv6 Mobile User Plane — Breaking Barriers Between Mobile Network and Internet.” During the hour-long seminar on June 16th, I detailed our Intel Intelligent Fabric architecture, Programmable networks, and 5G innovation alongside Technical Meisters Satoru Matsushima and Katsuhiro Horiba from SoftBank; and Arrcus Founder and CTO, Keyur Patel.

From analog voice transfer to packet data transfer, rapid mobile network advancements over the last 40 years have shaped everyday life. Still, the shift towards 5G SA and edge computing may be the most transformative yet. The ever-increasing level at which internet users are generating data means that next-generation networks must be smarter and have increased bandwidth to handle real-time analysis, transmission, security and reconstruction. By providing an ultra-reliable network with low latency and enhanced mobile broadband, 5G SA will provide stronger support for the incredible volume of machine-to-machine communications required in modern society. Due to the complex nature of legacy networks, programmable networks are appealing for their increased flexibility and deployability, and reduced costs.

While communications service providers (CoSPs) and cloud service providers (CSPs) are moving toward cloud-native architectures, the transition to 5G SA can be daunting. Despite the benefits, new service rollouts, complex network architectures and deployments, broad coverage and CAPEX/OPEX investments produce challenges for CoSPs on many fronts. Additional considerations must also be made for end-to-end security and operational difficulties resulting in network slowdowns.

SRv6 Mobile User Plane Architecture and Approach

SRv6 Mobile User Plane (MUP) is an innovative keystone architecture led by SoftBank that addresses the specific challenges of simplicity, scalability and end-to-end network slicing by leveraging network programmability for the future of computing. Our collaboration with Arrcus and SoftBank has produced an approach to embody the SRv6 MUP architecture as a plug-in for 5G through a masterful combination of Intel P4 programmable switches, Arrcus’ programmable SRv6 fabric and SoftBank’s vision.

Taking a step toward IP-based packet routing, SRv6 is segment routing (SR) that works within the IPv6 address scheme. This makes it easy to deploy because utilizing the IPv6 forwarding plane allows it to bypass complex multi-protocol label switching (MPLS) and hardware support for forwarding. MUP architectures subsequently enable the deployment of SRv6 in mobile networks. Sitting at the heart of the SRv6 network domain is the state-of-the-art ArcOS. The 64-bit internet-scale network operating system runs on a switch using the Intel® Tofino™ programmable Ethernet Switch ASIC to provide MUP enabled PE router GTP-U to SRv6 translation for packet processing and forwarding to the 5G core.

In addition to the P4-programmable pipeline, intelligent packet processing for accelerating AI/ML workloads, and expandable table and buffer sizes with Intel FPGAs contribute to overall network intelligence. The incorporation of Intel Intelligent Fabrics also provides enhanced performance, visibility, and management by elevating congestion control, identifying delays or hotspots with real-time in-band network telemetry (INT), and analyzing packet flows with Intel Deep Insight Network Analytics Software. Ultimately, SRv6 MUP acts as a plug-in to bridge 5G and IP networks by simplifying data routing and making the network more flexible, efficient and robust.

The implementation of SRv6 MUP will allow mobile networks to employ an internet-based way of computing. This method utilizes commercial IP routers rather than expensive custom hardware, thus reducing costs. Simultaneously, SRv6 MUP facilitates connectivity with other new applications such as multi-access edge computing (MEC). MEC brings cloud computing services to customers at the “edge” of a network.

The network architecture keeps client data more localized rather than sending it across networks for processing. Edge computing provides enhanced security and data management, lower latency, and improved processing performance. SRv6 MUP also enables network slicing techniques, which divide a network into distinct virtual connections. This allows operators to “slice” sections of the network for consumer use.

The SRv6 MUP architecture provides the bridge to make these and other emerging technologies a practical reality.

We showcased our proof of concept (PoC) earlier this year at Mobile World Congress Barcelona. The PoC demonstration showed that the SRv6 MUP can successfully bridge IP routing and 5G networks and add flexibility and efficiency to the network as promised. The webinar, followed by a Q&A session, features a similar demo regarding the impact of SRv6 MUP on network traffic, along with a closer look at the construction of the SRv6 framework and results from SoftBank’s performance evaluation research. Revisit the webinar recording and supplemental resources to learn more about the SRv6 MUP approach, its foundational components, deeper applications for edge computing, and other cost-reduction benefits.