Switchless Interconnect Fabric Message Distribution

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current multicomputer interconnect systems face challenges in efficiently managing message distribution and prioritization across clusters, leading to potential bottlenecks and delays in data transmission, especially in high-performance computing environments.

Innovation Solution

The implementation of a message distribution system with end-to-end partitioning of message pathways and multiple priority levels, utilizing data distribution modules (DDMs) and host-bus adapters (HBAs) to ensure non-blocking message delivery and support for cooperative working groups, allowing for efficient and timely transmission of messages across clusters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If message distribution is implemented without end-to-end partitioning, then device complexity is reduced, but message delivery efficiency deteriorates due to bottlenecks and delays

Engineering Contradiction:

Improvemessage delivery efficiencyVSAvoidmessage pathway structure

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The message pathway is segmented into multiple independent partitions from source to destination. Each partition can independently transmit messages without interfering with other partitions, eliminating bottlenecks and improving overall message delivery efficiency while maintaining manageable complexity through modular organization

Inventive Principle:

Principle #1Segmentation

2Reliability

If single priority level is used, then device complexity is reduced, but message transmission reliability deteriorates during busy conditions

Engineering Contradiction:

Improvemessage transmission reliabilityVSAvoidpriority level structure

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

Different priority levels are assigned to different message types or destinations within the system. Critical messages receive higher priority treatment while less urgent messages use lower priority channels, ensuring reliable transmission of important data while maintaining system simplicity through differentiated quality of service

Inventive Principle:

Principle #3Local quality

3Speed

If non-blocking message delivery is implemented, then message delivery speed is improved, but device complexity increases due to additional control mechanisms

Engineering Contradiction:

Improvemessage delivery speedVSAvoidcontrol mechanism structure

Core Design Contradiction:

SpeedVSDevice complexity

Solution Approach 1:

Message pathways are pre-established and reserved between source and destination nodes before actual message transmission. This preliminary setup enables immediate high-speed transmission without blocking, as the path is already confirmed and resources are pre-allocated, reducing the need for complex real-time control mechanisms

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11343197B2Packet-flow message-distribution system

Publication Date: 2022.05.24 LIGHTFLEET CORP

AI summary

Switchless interconnect fabric message distribution includes end-to-end partitioning of message pathways or multiple priority levels with interrupt capability. A switchless interconnect fabric message distribution system includes a data distribution module and at least two host-bus adapters connected to the data distribution module. The data distribution module includes partition first in first out buffers. Each of the host-bus adapters includes an input manager connected to input priority first in first out buffers and an output manager connected to priority first in first out buffers.