Data Flow Offloading via Switch Bypass in 5G UPF Nodes

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In 5G mobile access architectures, network function modules in UPF nodes handle high data rates inefficiently due to repeated packet processing, unscalable SDN programming, and security restrictions on NFVI access, leading to performance penalties and latency issues.

Innovation Solution

Offload data processing tasks from network function modules to the NFVI forwarder or network interface, utilizing intelligent server adapters to bypass the switch and enable direct packet handling, thereby optimizing data flow management.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If packets are processed by both the switch and the network function module, then security policies are followed and proper authorization is maintained, but processing latency increases and throughput decreases

Engineering Contradiction:
Improvesecurity complianceVSAvoidprocessing latency
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent introduces an intermediary mechanism (flow table with offload indicators) that sits between the switch and network function module. This intermediary enables the switch to handle certain packet processing tasks independently based on pre-configured flow rules, while maintaining the ability to forward packets to the network function module when needed. This resolves the contradiction by allowing security-compliant processing paths to be established in advance, reducing real-time latency without compromising security authorization.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary actions by pre-configuring flow tables with offload indicators and processing rules before packets arrive. The switch is pre-programmed with knowledge of which packets can be processed independently and which require network function module involvement. This preliminary configuration enables fast path processing for eligible packets without real-time security checks, reducing latency while maintaining security compliance through pre-validated rules.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If the SDN controller programs the switch to handle IP flows, then flow processing capability is improved, but scalability deteriorates due to control signal latency

Engineering Contradiction:
Improveflow processing capabilityVSAvoidscalability
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent segments the control and data planes by enabling the switch to operate with pre-configured flow rules independently of the SDN controller. The flow table is divided into regular entries and offload indicator entries, allowing the switch to handle different packet types through different paths without continuous controller intervention. This segmentation improves flow processing capability while enhancing scalability by reducing dependency on controller-switch communication for every flow decision.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The switch is empowered with self-service capabilities through the offload indicator mechanism. When the switch encounters packets matching offload indicator entries in the flow table, it can independently determine to process them through fast paths without contacting the SDN controller. This self-service approach improves productivity by eliminating control signal latency for routine flows while maintaining scalability, as the switch autonomously handles the majority of processing decisions.

Inventive Principle:
Principle #25Self-service

3Reliability

If the network function module processes all packets, then security and authorization are maintained, but processing efficiency and throughput decrease

Engineering Contradiction:
Improvesecurity authorizationVSAvoidprocessing efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies local quality by differentiating processing paths based on packet characteristics and flow rules. Not all packets are treated uniformly; instead, packets matching specific criteria (offload indicator entries) are routed through optimized fast paths in the switch, while other packets continue to the network function module. This local differentiation maintains security authorization for packets requiring it while improving overall processing efficiency through selective fast-path handling.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system dynamically adapts packet processing paths based on real-time conditions and pre-configured rules. The flow table can be updated to add or remove offload indicator entries, allowing the system to dynamically adjust which packets receive fast-path treatment. This dynamic approach maintains security authorization by preserving the ability to route packets to the network function module when needed, while improving processing efficiency by enabling fast paths for eligible packets during high-throughput scenarios.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4131014B1Method and device for offloading processing of data flows
Publication Date: 2026.01.07 CISCO TECHNOLOGY INC
  • EP4131014B1 patent drawingFigure 1
  • EP4131014B1 patent drawingFigure 2
  • EP4131014B1 patent drawingFigure 3

AI summary

In accordance with various implementations, a method is performed at a data plane node with one or more processors, non-transitory memory, and a control interface between a network function module associated with the data plane node and a switch associated with the data plane node. The method includes determining whether an offload capability is available for a data flow received at an ingress network interface of the data plane node. The method also includes determining whether the data flow satisfies offload criteria in response to determining that the offload capability is available. The method includes bypassing the network function module associated with the data plane node and providing the data flow to at least one of the switch associated with the data plane node or an egress network interface associated with the data plane node in response to determining the offload capability is available and the offload criteria is satisfied.