Interconnection Network Counter Extraction for Task-Specific Failure Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current monitoring tools for interconnection networks are unable to quickly identify equipment failures during calculation tasks, as they require manual exploration of node counters and lack the ability to selectively analyze performance indicators specific to the task, leading to inefficiencies in troubleshooting.

Innovation Solution

A device and method that acquire and extract counter values associated with a calculation task by correlating node lists with routing tables, filtering relevant data, and generating performance analysis reports, allowing for rapid identification of malfunctioning equipment within the network.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual exploration of node counters is used to identify equipment failures, then comprehensive monitoring coverage is achieved, but troubleshooting time increases significantly

Engineering Contradiction:
Improveequipment failure detection capabilityVSAvoidtroubleshooting time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts and isolates only the counter values relevant to the specific calculation task from the entire interconnection network. By identifying the subset of ports involved in the task and extracting only their counter values, the system eliminates the need to manually explore all node counters, thereby maintaining comprehensive monitoring coverage while dramatically reducing troubleshooting time.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent segments the monitoring task by dividing the interconnection network into task-related ports and unrelated ports. The counter extraction module separately processes and extracts counter values only for task-related ports, allowing focused analysis on equipment failures impacting the calculation task without being overwhelmed by the entire network's data.

Inventive Principle:
Principle #1Segmentation

2Loss of information

If all counter values across the interconnection network are monitored, then complete performance visibility is achieved, but data processing complexity increases

Engineering Contradiction:
Improveperformance information completenessVSAvoiddata processing complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent extracts only the necessary performance information related to the calculation task by identifying task-related ports through routing table analysis and node involvement information. This extraction process maintains complete performance visibility for the task while eliminating unnecessary data processing for unrelated ports, thereby reducing data processing complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary identification of task-related ports using routing table analysis and node involvement information before extracting counter values. This preliminary action filters the data set in advance, so that only relevant ports are included in the counter extraction process, significantly reducing the complexity of subsequent data processing while maintaining information completeness for the calculation task.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If current monitoring tools are used to track all ports, then network-wide monitoring capability is maintained, but the ability to quickly identify task-specific failures is lost

Engineering Contradiction:
Improvenetwork monitoring capabilityVSAvoidfailure identification speed
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments the monitoring function into network-wide monitoring capability and task-specific failure identification. By maintaining the ability to monitor all ports while adding a filtering layer that identifies and extracts only task-related port counters, the system preserves network-wide monitoring capability while dramatically improving failure identification speed for calculation tasks.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary filtering mechanism that sits between the network monitoring system and the failure identification process. This intermediary uses routing table analysis and node involvement information to filter and extract only the counter values relevant to the calculation task, enabling quick task-specific failure identification while maintaining the underlying network-wide monitoring capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3471356B1Device and method for acquiring meter values associated with a calculation task
Publication Date: 2020.06.17 BULL SA
  • EP3471356B1 patent drawingFigure 1
  • EP3471356B1 patent drawingFigure 2
  • EP3471356B1 patent drawingFigure 3A~3B

AI summary

The invention relates to a device for acquiring counter values ​​associated with a computational task implemented on an interconnection network, said interconnection network comprising a plurality of nodes interconnected via switches, said nodes and switches comprising a plurality of ports associated with counters, the values ​​of said counters being recorded over time, said computational task involving a subset of ports associated with subsets of nodes and switches, and said device comprising: - a routing acquisition module, capable of acquiring a routing table containing all the connections of the interconnection network, - a node involved acquisition module, capable of acquiring a list of the nodes involved in the computational task, - a port extraction module.capable of extracting a list of ports involved in the calculation task from the list of involved nodes and the routing table, and - a counter extraction module, capable of extracting the counter values ​​of the ports involved in the calculation task from the list of involved ports and the recorded values ​​of said counters.