Network Defined Edge Routing for 5G Low Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 5G networks face challenges in achieving low-latency traffic routing due to the lack of 5G radio access network functions and edge computing capabilities in all edge data centers, which hinders the delivery of high-performing services with low end-to-end latency.
Innovation Solution
The implementation of network-defined edge routing using a set of interconnected edge data centers that provide 5G Radio Access Network (RAN) and 5G Core functions, along with edge compute resources, utilizing segment routing and element-aware registration to compute paths that meet traffic class requirements, ensuring efficient traffic forwarding within the MEC domain.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traffic is routed through centralized data centers, then network capacity and throughput are improved, but end-to-end latency increases
Solution Approach 1:
The network is segmented into multiple edge data centers distributed geographically closer to end users, rather than relying on a single centralized data center. This segmentation allows traffic to be routed through the nearest edge data center, reducing latency while maintaining aggregate network capacity across the distributed infrastructure.
Solution Approach 2:
The architecture transitions from a single-dimensional centralized routing model to a multi-dimensional distributed edge routing model. By adding the spatial dimension of multiple geographically distributed edge data centers, the system can simultaneously achieve low latency (by selecting the nearest edge) and high throughput (by utilizing the aggregate capacity of all edge data centers).
2Loss of time
If all edge data centers are equipped with 5G RAN functions and edge compute resources, then low-latency routing is enabled, but device complexity and infrastructure cost increase
Solution Approach 1:
Edge data centers are designed with multi-functionality, capable of operating in different modes: some function as full 5G RAN access points with edge compute resources, while others serve as relay nodes or aggregation points. This universal design allows the infrastructure to provide low-latency routing where needed without requiring every node to have complete 5G RAN functionality, thereby reducing overall complexity.
Solution Approach 2:
The system introduces a path computation element that acts as an intermediary, intelligently determining whether traffic should be routed through a local edge data center with 5G RAN functions or forwarded to a remote data center. This mediator optimizes the balance between enabling low-latency routing and avoiding the complexity of deploying 5G RAN functions at every edge location.
3Productivity
If segment routing is implemented across interconnected edge data centers, then traffic forwarding efficiency is improved, but control plane complexity increases
Solution Approach 1:
The path computation element performs preliminary actions by pre-computing optimal routes through the segment routing domain and installing forwarding state in advance. This allows data planes at edge data centers to simply execute pre-determined segment routing instructions, achieving high forwarding efficiency without requiring complex real-time control plane decisions at each node.
Solution Approach 2:
A dedicated path computation element serves as an intermediary between the control plane and data plane, centralizing the complex segment routing computation functions. This separates the heavy computational burden of path optimization from the forwarding devices, allowing efficient traffic forwarding at edge data centers while concentrating control plane complexity in a specialized management entity.
Data Source
AI summary
Techniques are described for a network providing network defined edge routing for an application workload. For example, a controller receives element registration information for each module hosted on a set of interconnected edge data centers and for one or more network devices that interconnect the modules in the network; obtain one or more routing metrics of the network; compute, based on the element registration information and the one or more routing metrics, one or more paths mapped to respective traffic classes to route traffic via the set of interconnected edge data centers; receive a request to route traffic according to a traffic class; and in response, send, to the set of interconnected edge data centers, a response specifying a path of the one or more paths that is mapped to the traffic class to cause the set of interconnected edge data centers to route traffic according to the traffic class.