NIC Packet Duplication for C-RAN Kernel Bypass Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional packet capture techniques are infeasible in cloud radio access networks (C-RAN) due to the lack of separate servers and network components, and existing tools like tcpdump are ineffective with kernel bypass protocols, leading to inefficiencies in network packet analysis.

Innovation Solution

Utilizing a packet duplication function within a network interface card (NIC) to mirror incoming network traffic, exploiting embedded switches for selective packet duplication based on conditions like MAC addresses, and implementing a loopback function to streamline analysis by leveraging preexisting hardware and software infrastructure.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional packet capture techniques (top-of-rack switch port mirroring) are used, then network packet analysis is achieved, but additional dedicated hardware servers and network components are required which are infeasible in C-RAN environments

Engineering Contradiction:
Improvenetwork packet analysis capabilityVSAvoidhardware infrastructure requirements
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent merges the packet capture functionality into the existing network interface card (NIC) by utilizing its embedded switch, eliminating the need for separate capture servers and top-of-rack switches. The NIC's embedded switch is configured with port mirroring rules to duplicate packets directly at the network edge where they originate.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The NIC performs packet duplication and capture operations autonomously using its own embedded switch resources, without requiring external dedicated capture infrastructure. The system serves its own packet analysis needs through self-contained functionality within the NIC.

Inventive Principle:
Principle #25Self-service

2Loss of information

If packet mirroring is implemented using top-of-rack switch, then complete network traffic duplication is achieved, but significant data volume and processing overhead are generated

Engineering Contradiction:
Improvenetwork traffic visibilityVSAvoiddata volume for analysis
Core Design Contradiction:
Loss of informationVSQuantity of substance

Solution Approach 1:

The patent applies selective packet mirroring where only specific traffic flows matching defined criteria are duplicated. The embedded switch uses port mirroring rules with filters based on source/destination MAC addresses, IP addresses, and protocol types to mirror only relevant packets rather than all traffic.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

Instead of mirroring all network traffic, the system implements partial mirroring of only those packets that match specified conditions. This reduces the quantity of duplicated data while maintaining sufficient visibility for network analysis and troubleshooting.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If standard packet capture tools (tcpdump) are used with kernel bypass protocols, then packet analysis is attempted, but the tools are ineffective due to inability to handle kernel bypass traffic

Engineering Contradiction:
Improvepacket analysis accuracyVSAvoidcompatibility with kernel bypass protocols
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent introduces the NIC's embedded switch as an intermediary that performs packet duplication before traffic reaches the kernel bypass path. This allows standard capture tools to analyze the mirrored packets copied by the embedded switch, while the main traffic flow continues through the kernel bypass protocol without interruption.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP4679792A1Network packet mirroring for communications network analysis
Publication Date: 2026.01.14 MICROSOFT TECHNOLOGY LICENSING LLC
  • EP4679792A1 patent drawingFigure 1
  • EP4679792A1 patent drawingFigure 2
  • EP4679792A1 patent drawingFigure 3

AI summary

The techniques disclosed herein provide a system for mirroring network traffic by way of network packet duplication using a network interface card in a communications environment. More specifically the present techniques are directed to retrieving data that is transmitted over a communications network (e.g., network packets) in a cloud radio access network (C-RAN) environment, also referred to as packet capture. For instance, the proposed system can mirror network traffic between a radio unit (RU) and a virtual radio access network (vRAN) within a distributed unit (DU) server of a telecommunications network (e.g., a 5G network). As such, the disclosed techniques exploit preexisting hardware and software features present in many network interface cards. In this way, the present techniques enable granular packet capture without requiring additional specialized hardware and/or software. Moreover, the network interface card can be configured to loopback scheduling information to streamline packet analysis operations.