Journey to a Converged Core, A Technical View

Five years have passed since the early drop of 3GPP’s Release 15, launching the industrywide journey in pursuit of 5G.  And today 5G is still evolving, as subsequent releases and pending future releases continue to define the capabilities and services enabled by the 5G architecture.  In hindsight, most marketing teams would admit they jumped the gun on 5G, setting overly high expectations with the promise of new experiences.  And while the biggest promise of 5G has always been speed, the network transformation to get there has been anything but fast.  Migrating from monolithic legacy architectures to a 5G service-based architecture (SBA) has proved to be a daunting task, however this transformation remains critical for the successful monetization of the new experiences and capabilities that 5G has been promising to deliver.

A brief history of 5G and innovation

At the center of 5G innovation is the core.  The 5G Core is a rethinking of how we have done things in the past, yet maintains a backwards compatibility with the 4G LTE era while adding profound benefits for 5G:

• The 5G Core fixes some of the limitations of 4G – for example, Session and Service Continuity (SSC) modes, Access, and Session Management separation
• Control and User Plane separation is designed as part of the architecture from day one, enabling edge computing use cases (MEC), high bandwidth, and LBO to name a few, all of which can be delivered easily and economically
• Protocols like SBI (Service Based Interfaces) enable new development architectures for the 5G Core leading to more rapid innovation
• The 5G Core profiles and rules can be deployed granularly through methods such as network slicing, and network policy

We could go on for pages and pages, bulleting out features and benefits of 5G, typing a list that may never end as future feature releases continue to refine and add enhancements. And while all these new capabilities that 5G is bringing sound great, there is a difference between what has been put to paper and what is being done in the real world today.  For example, many new 5G use cases require devices to support N1 NAS protocols.  With new 5G devices and a brand new 5G SA packet core this is no problem.  But we live in the real world filled with existing 4G services and others previous generations of devices that do not support the new protocol(s). Most of these devices work great today and have no immediate need to be replaced.  Thus, supporting multiple generations of technology has led to a need for maintaining two mobile cores, one that will support 4G legacy NAS protocols and 5G NSA devices with legacy NAS protocols, and a second core for supporting the newer 5G N1 NAS devices.

Two Cores, no problem?

On the surface, adding another core just for 5G needs doesn’t appear to be too challenging and it is not.  In a standard instance, packet cores (e.g., SMF, UPF, PGW, SGW, PCRF, PCF, etc.,) could be deployed and managed in parallel.  Operating two cores in parallel is not optimal and such a configuration presents significant challenges for defining policy and subscriber data management.  Any communication service provider considering this route will need to employ multiple operations teams and some creative engineering for maintaining their offer packages.  With two cores you are effectively doubling the costs and risks without coming close to doubling the reward.  Fortunately, there is an intriguing alternative to running two mobile cores in parallel and that is finding a mechanism for converging the core functions into a unified architecture.

We set out on this path years ago, working closely with service providers, like T-Mobile, to understand the challenges they faced and the value of a converged core.  While there has always been plenty of noise about 5G NSA and 5G SA deployments, with T-Mobile we focused beyond that to define an architecture for a “Converged Core”.  This meant improving operations, customer experience, and overall subscriber quality while maintaining what already existed.   We call it a “Converged Core” because it brings together the past, present, and future of mobile core technologies.  There are other names for this in the industry such as a “Unified Core” or an “Any Generation Core” that describe the same idea.  We leveraged our expertise as the industry leader in 4G packet cores, matching market requirements with technological capabilities to build a smoother path forward to 5G with a converged core.

Addressing the Business Outcomes

The Cisco Converged Core addresses the critical business challenges experienced by our service provider partners, including:

• The ability to define policy at one location and apply the policy across access networks (4G, 5G, Wi-Fi)
• The ability to reduce operations costs with a single, unified platform instead of running a standalone 4G core and a standalone 5G core, in parallel
• The ability to quickly deploy new consumer and enterprises services in a unified manner, regardless of the end users access requirements
• A unified billing and support experience for the end customers

What is the Converged Core?

The Converged Core is the heart of the mobile network.   In contrast to previous generations of technology, 5G Cores are software based and defined by 3GPP’s Service Based Architecture (SBA)–a set of interconnected network functions (NFs).   These network functions, such as the Session Management Function (SMF), Access Management Function (AMF) and the User Plane Function (UPF), to name a few, are containerized software that supports the separation of the control plane from the user plane.  Instead of hardware dependencies and the point-to-point (P2P) limitations of the previous generation’s technology, the Service Based Architecture (SBA) is meant to be a flat, built for portability and scalability with the cloud. Each function is cloud-native and much different than their predecessors. Take the SMF, for example, the SMF provides common session management for all accesses, including WLAN, and handles all session management signaling for user equipment (UE) which is relayed by the Access Management Function (AMF).  In 4G architectures, some of these capabilities are provided by the Packet Data Network Gateway Control Plane Function (PGW-C). The remaining control plane parts including the Serving Gateway Control Plane Function (SGW-C) and Packet Data Network Gateway Control Plane Function (PGW-C) along with the session management functions from the old Mobility Management Entity (MME) were also migrated into the new Session Management Function (SMF).

Relative to the Session Management Function (SMF), the User Plane Function (UPF) supports a set of operations such as forwarding to other functions, encapsulation/decapsulation of data, bitrate enforcement, and application detection. The Session Management Function (SMF) can dynamically activate and configure the User Plane Function (UPF), or a subset of User Plane Functions, to provide the traffic handling functionality needed for a session.  Within the 5G architecture, one or multiple User Plane Functions (UPF) can be activated and configured by the Session Management Function (SMF) per session, or as needed per application. At a high level and unlike 4G, which has distinct user plane entities (SGW-U or PGW-U), the 5G SBA consolidates these entities with the User Plane Function (UPF).  Effectively establishing a general version of the 4G core’s user plane parts, inclusive of the Serving Gateway User Plane Function (SGW-U), Packet Data Network Gateway User Plane Function (PGW-U) and the Traffic Detection Function (TDF).

The decoupling and reshuffling of network functions’ roles and responsibilities between a 4G core architecture and the 5G Service Based Architecture (SBA) has been both a challenge and opportunity.  Rewriting and implementing network functions to be backwards compatible and forwardly compliant was slow and difficult, but worth it. Like the fable we learn at a young age, slow and steady wins the race, especially when trust and reliability are of utmost importance.  Our approach to the business problems and technical challenges of supporting multiple access technologies (4G, 5G, and Wi-Fi) was to build a composable 5G converged core that communication service providers like T-Mobile can use to deploy capabilities as needed.  Fundamental to the design of Cisco’s Converged Core is being cloud native, a containerized solution supporting both 5G devices and legacy 4G devices, on the same platform, thereby eliminating the need for a separate 4G Packet Data Network Gateway (PGW), two sets of policies deployed in two different locations, and a future full of headaches.

The Cisco Converged Core is aligned with 3GPP’s TS 23.501 and TS 23.502 definitions for non-roaming architecture interworking between 5G and EPC/E-UTRAN. To simplify the deployment, SMF+PGW-C as defined by 3GPP is interpreted as SMF+S5-C, without a full PGW-C but with the ability to support the S5 Control interface on the SMF. Similarly, UPF+PGW-U, as defined by 3GPP, is also interpreted as UPF+S5-U.  This approach, , allows communications service providers to migrate e and leverage a Service Based Architecture by using the new South Bound Interfaces (SBI) defined in 3GPP.  The result is focused investments and a smooth migration path for a communication service provider’s 5G journey.

With a Converged Core communication service providers can support 4G devices connecting via the Mobility Management Entity (MME) and the Serving Gateway (SGW) in 4G or 5G NSA architectures to be anchored on the same Session Management Function (SMF) and User Plane Function (UPF) as a device that would connect via the Access Management Function (AMF) in a 5G SA architecture.  All capabilities northbound such as Policy via Policy Charging Function (PCF) and Charging via Charging Function (CHF) are supported using new service-based architectures which communication service providers can use to unify offerings and focus critical investments.  The architecture also allows for ease of mobility across coverage domains i.e., 4G LTE, 5G NSA and 5G SA creating a much more seamless experience for end users.

Leading with this architecture, the Cisco team developed a single converged core platform that can be deployed and configured, as needed.  Taking the deployment architecture one step further, the Cisco Converged Core can be deployed in different variations depending on operator use cases:

1. Standalone 5G SA Session Management Function (SMF) and 5G User Plane Function (UPF) only
   1. Typically used for 5G SA only use cases such as Fixed Wireless Access (FWA) with 5G SA devices
   2. Device Types supported: 5G SA Capable

2. SMF+S5C and UPF+S5-U mode to enable external Serving Gateway (SGW) integration
   1. Typically used for migrating in a multivendor environment where existing investments in the Serving Gateway (SGW) are leveraged for extended period or when using specific roaming architectures
   2. Device Types supported: 4G, 5G NSA, 5G SA

3. Session Management Function (SMF) + Serving Gateway Control Plane Function (SGW-C) and the User Plane Function (UPF) + Serving Gateway User Plane Function (SGW-U) as a complete converged core
   1. An optimized footprint deployment for supporting a single unified operations
   2. Device Types supported: 4G, 5G NSA, 5G SA

Summary

By leading with a unified product architecture that enables in-field decisions on how to configure and deploy various use case specific models, the Cisco Converged Core:

• Reduces learning curve for communication service provider operations teams
• Simplifies management and operations by unifying policy and service offerings
• Improves end-user experiences
• Provides service parity for communication service providers migrating from existing Cisco 4G deployments
• Economizes investments in the architecture of the future

Cisco has long been an industry leader in mobile cores and our Converged Core solution continues that leadership by delivering outcomes tightly aligned with our communication service provider customer needs.   Migrating from 4G services to the 5G future is a journey best made with partners you can trust.  The Converged Core is continuously improving, adding new capabilities, and optimizing itself as it evolves.  With new use cases for communication service providers and enterprises alike on the horizon, the opportunities for innovation seem endless.

Resources

Visit our product pages to learn more about how the Cisco Cloud Core and Packet Core portfolio

To learn more about T-Mobile and Cisco Launch of the World’s Largest Cloud Native Converged Core Gateway, read the December 2022 press release

Share: