Network Element Configurable Forwarding Circuitry Data Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing network computing systems face challenges in efficiently performing multiple complex calculations over high-bandwidth data streams in parallel with low latency, particularly in implementing flexible data reduction operations across network elements without bandwidth sharing.

Innovation Solution

A network element with multiple ports and computational modules, connected via configurable forwarding circuitry, allows for flexible interconnection topologies to perform data reduction operations. The central block derives and applies topologies to interconnect child and parent ports and modules, enabling parallel execution of data reduction operations without sharing port bandwidth.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If network elements perform multiple complex calculations over high-bandwidth data streams in parallel, then productivity is improved, but device complexity increases

Engineering Contradiction:
Improvethroughput of data reduction operationsVSAvoidcomplexity of forwarding circuitry and topology management
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The network element is divided into multiple independent computational modules (e.g., ALUs) that can be individually configured and interconnected. Each module handles a portion of the data reduction operation, allowing parallel processing while maintaining modular complexity management through the configurable forwarding circuitry.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The forwarding circuitry is made dynamically configurable to establish different interconnection topologies among computational modules based on the specific data reduction operation required. This dynamic reconfiguration allows the system to adapt to various computational patterns without permanent hardware commitments, balancing flexibility with manageable complexity.

Inventive Principle:
Principle #15Dynamics

2Adaptability or versatility

If flexible interconnection topologies are implemented among computational modules, then adaptability is improved, but device complexity increases

Engineering Contradiction:
Improveflexibility of data reduction operationsVSAvoidcomplexity of configurable forwarding circuitry
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The configurable forwarding circuitry serves multiple functions: establishing interconnections among computational modules, routing data between ports and modules, and supporting various data reduction topologies. This multi-functional approach provides high adaptability without requiring separate dedicated hardware for each function, thereby controlling overall device complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Productivity

If data reduction operations are performed in parallel at full port bandwidth, then productivity is improved, but loss of time increases due to topology configuration

Engineering Contradiction:
Improveparallel processing capabilityVSAvoidlatency for topology derivation and configuration
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary configuration of computational modules and forwarding circuitry based on pre-derived topologies before actual data reduction operations begin. By preparing the interconnection structure in advance according to the operation type, the system minimizes configuration latency during execution and enables immediate parallel processing at full bandwidth.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3855690B1Network element supporting flexible data reduction operations
Publication Date: 2023.11.22 MELLANOX TECHNOLOGIES LTD(IL)
  • EP3855690B1 patent drawingFigure 1
  • EP3855690B1 patent drawingFigure 2
  • EP3855690B1 patent drawingFigure 3A~3B

AI summary

A network element includes a plurality of ports, multiple computational modules, configurable forwarding circuitry and a central block. The ports include child ports coupled to child network elements or network nodes and parent ports coupled to parent network elements. The computational modules collectively perform a data reduction operation of a data reduction protocol. The forwarding circuitry interconnects among ports and computational modules. The central block receives a request indicative of child ports, a parent port, and computational modules required for performing reduction operations on data received via the child ports, for producing reduced data destined to the parent port, to derive from the request a topology that interconnects among the child ports, parent port and computational modules for performing the data reduction operations and to forward the reduced data for transmission to the selected parent port, and to configure the forwarding circuitry to apply the topology.