NoC Branched Routing for Low-Latency Multicast Delivery

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing network-on-chip (NoC) technologies lack efficient methods for branched communication and multicast transmission, particularly in statically routed and time-sensitive environments, leading to suboptimal performance and resource utilization.

Innovation Solution

A method for branched communication in a statically routed and time-sensitive NoC, where router nodes exchange messages with each other by sending flow control data containing branch markers to define branch routes, allowing messages to be forwarded along both main and branch routes, and a flow control data structure with flits that include branch markers to guide routing decisions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If sequential routing through all destination nodes is used for multicast transmission, then all destinations can be reached, but transmission latency increases and communication efficiency decreases

Engineering Contradiction:
Improvemulticast transmission completenessVSAvoidtransmission latency
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent segments the multicast transmission into parallel path segments by introducing branch markers that divide the message flow into multiple independent routes simultaneously. Instead of sequential routing through all destinations, the message is divided into parallel transmission paths that reach different destination nodes concurrently, thereby reducing overall transmission latency while maintaining complete delivery to all destinations.

Inventive Principle:
Principle #1Segmentation

2Adaptability or versatility

If traditional P2P or multicast routing is used, then point-to-point or point-to-multipoint connections are established, but branched communication to multiple destinations simultaneously is not supported

Engineering Contradiction:
Improvecommunication pattern flexibilityVSAvoidrouting complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces dynamic routing capabilities by embedding branch markers within the message flow that enable router nodes to dynamically create branched communication paths. The routing structure transitions from static P2P or traditional multicast to dynamic branched routing, where the message can adaptively split into multiple destinations based on the branch marker instructions, enhancing communication pattern flexibility without requiring complex pre-configured routing tables.

Inventive Principle:
Principle #15Dynamics

3Manufacturing precision

If flow control data is modified at each hop to remove current hop information, then routing information is kept updated, but processing overhead increases

Engineering Contradiction:
Improverouting information accuracyVSAvoidmessage forwarding efficiency
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent applies preliminary action by pre-configuring branch markers and associated hop information within the flow control data before message transmission begins. Router nodes perform bulk modifications of flow control data in advance, preparing multiple destination routes simultaneously. This preliminary preparation reduces the processing overhead during actual message forwarding, as routers can efficiently distribute messages along pre-established branched paths without performing complex real-time modifications at each hop.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4064644B1Communication method in a network-on-a-chip
Publication Date: 2025.11.12 AIRBUS DEFENCE & SPACE GMBH
  • EP4064644B1 patent drawingFigure 1
  • EP4064644B1 patent drawingFigure 2
  • EP4064644B1 patent drawingFigure 3

AI summary

The measures described herein can extend the functionality of NoCs (22). In particular, a method for multi-source communication for Triple Modular Redundancy (TMR), a method for branched communication, and a method for virtual buses are proposed.