Distribution of 5G Core to Network Edge

Thus far in our blog series, we have discussed migrating to an open cloud-native ecosystem, the 5G Core and its architecture, and how Dell Telecom Infrastructure Blocks for Red Hat can help simplify 5G network deployment for Red Hat® OpenShift®. Now, we would like to introduce a key use case for distributing 5G core User Plane functions from the centralized data center to the network edge.

Distributing Core Functions in 5G networks

The evolution of communication technology has brought us to the era of 5G networks, promising faster speeds, lower latency, and the ability to connect billions of devices simultaneously. However, to achieve these ambitious goals, the architecture of 5G networks needs to be more flexible, scalable, and efficient than ever before. With the advent of CUPS, or Control and User Plane Separation, in later LTE releases, the telecommunications industry had high expectations for a prototypical distributed control-user plane architecture. This development was seen as a steppingstone towards the more advanced 5G networks that were on the horizon. CUPS aimed to separate the control plane and user plane functionalities where the Control Plane (specifically the Session Management Function or SMF) is typically centralized while the User Plane Function (UPF) can be located alongside the Control Plane or distributed to other locations in the network as demanded by specific use cases. 

Understanding the need for Distributed UPF 

The UPF is a key component in 5G networks, responsible for handling user data traffic. Distributed User Plane Function (D-UPF) is an advanced network architecture that distributes the UPF functionality across multiple nodes closer to the user and enables local breakout (LBO) to manage use cases that requires lower latency or more privacy, enabling a more scalable and flexible networking environment. With D-UPF, operators can handle increasing data volumes, reduce latency, and enhance overall network performance. By distributing the UPF, operators can effectively manage the increasing data demands across different consumer and enterprise use cases in a cost-effective manner.

Figure 1: Distributed User Plane function in 5G Core Architecture

D-UPF also plays a crucial role in enabling edge computing in 5G networks. By distributing the user plane traffic closer to the network edge, D-UPF reduces the latency associated with data transmission to and from centralized data centers. This opens opportunities for real-time applications, such as autonomous vehicles, augmented reality, and industrial automation, where low latency is critical for their proper functioning. 

Distributed UPF deployment options 

Figure 2: D-UPF deployment and functionality 

The above diagram provides an overview of the different roles D-UPF may play in a 5G architecture. For example:

• Centralized UPF/PSA-UPF: In the simplest scenario, the UPF is centralized and session anchor occurs within the data center and takes care of the long-term stable IP address assignment.  One such example includes VoLTE / NR call where PDU Session Anchor (PSA)-UPF traffic steers to IMS.
• Intermediate UPF (I-UPF):An intermediate UPF (I-UPF) can be inserted on the User Plane path between the RAN and a PSA. Here are two possible reasons to do that:
  • If due to mobility, the UE moves to a new RAN node and the new RAN node cannot support N3 tunnel to the old PSA, then an I-UPF is inserted and this I-UPF will have the N3 interface towards RAN and an N9 interface towards the PSA UPF. This process of linking multiple UPFs is called UPF chaining, which involves directing user data flows through a series of UPFs, each of which is performing specific functions.
  • You might want to deploy UPF within the Local Data Center/Edge for a low latency use case to steer data traffic to a co-deployed MEC for edge services or to break-out traffic to the local data network.

Challenges and considerations for D-UPF deployments 

Now that we have reviewed the need for D-UPF and the different deployment scenarios, let’s consider some of the obstacles you will encounter along the way. As we all know, these network functions have their own needs, especially when it comes to the amount of data being inspected, routed, and forwarded across from the core to the edge, and back again.  Below are four areas for consideration:

• Resource Constraints: Edge or remote locations often have limited physical space available for deploying network equipment. The challenge lies in accommodating the necessary hardware, cooling systems, and other infrastructure within these space-constrained environments. Remote locations may also have limited or unreliable power supply infrastructure. Opting for infrastructure with optimal power efficiency, high density, serviceability, ruggedized exterior, and optimized for edge form factors becomes important as UPFs are extended to the edge.
• Performance Requirements: The need for low latency Infrastructure is critical to ensure real-time responsiveness and a seamless user experience when deploying core functions to the edge. Also, by processing data at the network edge with minimal latency, the need for large bandwidth networks to transmit data to centralized core is reduced. This helps in optimizing network bandwidth and lowering the operating costs. This ultimately reduces the CSP’s reliance on centralized core infrastructure for time-sensitive operations.
• Orchestration and Automation: Deploying and managing UPFs distributed across edge locations is a complex challenge. This includes tasks such as workload placement, resource allocation, and automated management of edge infrastructure. Choosing a horizontal telco cloud platform that supports automated distributed core deployment and provides the capability to expand and scale the compute and storage resources to accommodate the varying demands at the edge is a must.
• Lifecycle Management and Operating Cost: Another significant factor is the increased costs associated with first deploying and then operating remote deployments. The large number of these locations coupled with their limited accessibility makes them more expensive to construct and maintain. To address this, zero-touch provisioning at network setup k and sustainable lifecycle management are necessary to optimize the economics of the edge.

The Horizontal Cloud Platform: Dell Telecom Infrastructure Blocks

Figure3: An Implementation View of Dell Telecom Infrastructure Blocks for Red Hat running the UPF 

Dell Technologies is at the forefront of providing cutting-edge cloud-native solutions for the 5G telecom industry. As discussed in our previous blog, Telecom Infrastructure Blocks for Red Hat is one of those solutions, helping operators break down technology silos and empowering them to deploy a common cloud platform from Core to Edge to RAN. These are engineered systems, based on high-performance telecom edge-optimized Dell PowerEdge servers, that have been pre-validated and integrated with Red Hat OpenShift ecosystem software. This makes them a perfect solution for tackling the D-UPF challenges outlined in this blog.

• Resource Constraints:
  • Space-efficient modular designed server options for telecom environments, such as the Dell PowerEdge XR8000 series servers, allow providers to mix and match components based on workload needs.  They can run multiple workloads, such as CU/DU and UPF in the same chassis.
  • Smart cooling designs support harsh edge environments and keep systems at optimal temperatures without using more energy than is needed.[1]
  • Rugged and flexible server options that are less than half the length of traditional servers and offer front or rear connectivity make installation in small enclosures at the base of cell towers easier.[2]
    
    
• Performance Requirements:
  • Provides low-latency processing for edge computing nodes at the network edge.[3]
  • 4^th Gen Intel Xeon Scalable processors with Intel vRAN Boost increase DU capacity (up to 2X in specific scenarios), and increases the packet core UPF and RAN CU performance by 42%.[2]  
    
    
• Orchestration and Automation:
  • Horizontal cloud stack engineered platform based on Red Hat OpenShift allows operators to pool resources to meet changing workload requirements. This is achieved by automating server discovery, creating and maintaining a server inventory, and adding the ability to configure and reconfigure the full hardware and software stack to meet evolving workload requirements.
  • Servers leverage dynamic resource allocation to ensure that computing resources are allocated precisely where and when they are needed. This real-time optimization minimizes resource waste and maximizes network efficiency.[3]
     
• Lifecycle Management and Operating Cost:
  • Include Dell Telecom Infrastructure Automation software to automate the deployment and life-cycle management as its fundamental components.
  • Backed by a unified support model from Dell with options that meet carrier grade SLAs, CSPs do not have to worry about multi-vendor management for the cloud infrastructure support (for both the hardware and cloud platform software), as Dell becomes the single point of contact in the support of telco cloud platform and works with its partners to resolve issues.

Summary

In summary, the need for D-UPF in 5G networks arises from the requirements of handling massive data volumes, improving network efficiency, reducing latency, enabling edge computing, and supporting advanced 5G services. Selecting among the different deployment scenarios possible will require ensuring you have an infrastructure capable of meeting your changing objectives for today and the flexibility and scalability to see you through your long-term goals. For example, you can host and support the deployment and management of content delivery network (CDN) at the network edge, where Dell Telecom Infrastructure Blocks for Red Hat can also serve as a telco cloud building block. By implementing this engineered telco cloud platform solution from Dell, we believe CSPs will be able to streamline the process and reduce costs associated with the deployment and maintenance of UPF across edge locations.

To learn more about Telecom Infrastructure Blocks for Red Hat, visit our website Dell Telecom Multicloud Foundation solutions.