Parallel Calculation System Bandwidth Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing parallel calculation systems face issues with high processing overhead due to software layers, unpredictable communication bandwidth, and imperfect separation of virtual machines and processes, leading to performance interference between objects.

Innovation Solution

A parallel calculation system incorporating functional nodes with mathematical and I/O functions, network switches as communication channels, and a bandwidth management unit to ensure direct communication between nodes without an OS or virtualization layer, managing communication bandwidth and node groups for efficient processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a virtualization layer and virtual machines are introduced to manage parallel processing, then resource allocation and process management become more flexible, but processing overhead increases due to the additional software layers

Engineering Contradiction:
Improveresource allocation flexibilityVSAvoidsoftware layer overhead
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent extracts and removes the virtualization layer and virtual machine software components from the parallel processing system. By eliminating these intermediate software layers, the system achieves direct hardware access for processes while maintaining resource allocation capabilities through hardware-based functional node groups, thereby reducing processing overhead while preserving adaptability.

Inventive Principle:
Principle #2Taking out (Extraction)

2Productivity

If hardware is shared by multiple objects through time-sharing allocation, then resource utilization improves, but performance interference occurs between executing objects

Engineering Contradiction:
Improveresource utilizationVSAvoidperformance predictability
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent segments the parallel processing hardware into distinct functional node groups, where each group is dedicated to specific processes or applications. This segmentation allows multiple objects to utilize hardware resources simultaneously without time-sharing conflicts, as each functional node group operates independently with guaranteed communication bandwidth, eliminating performance interference while maintaining high resource utilization.

Inventive Principle:
Principle #1Segmentation

3Ease of operation

If virtual machines are used to isolate processes, then process separation is achieved, but the separation is imperfect due to shared hardware resources

Engineering Contradiction:
Improveprocess isolationVSAvoidseparation completeness
Core Design Contradiction:
Ease of operationVSReliability

Solution Approach 1:

The patent introduces functional node groups as hardware-based intermediaries between processes and physical hardware. These functional node groups act as dedicated communication channels that provide complete separation between processes, ensuring that each process operates within its own isolated hardware environment with guaranteed bandwidth, thereby achieving perfect separation without the imperfections of software-based virtualization.

Inventive Principle:
Principle #24Intermediary (Mediator)

4Adaptability or versatility

If communication bandwidth is not guaranteed for virtual machines, then system flexibility is maintained, but communication performance becomes unpredictable

Engineering Contradiction:
Improvesystem flexibilityVSAvoidcommunication bandwidth
Core Design Contradiction:
Adaptability or versatilityVSSpeed

Solution Approach 1:

The patent implements dynamic bandwidth allocation at the functional node group level, where communication bandwidth is guaranteed for each functional node group while allowing flexible resource distribution within groups. This dynamic approach ensures predictable communication performance between processes by dedicating hardware resources, while maintaining overall system flexibility through configurable functional node group assignments and bandwidth parameters.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS8824317B2Parallel calculation system, and method and program thereof
Publication Date: 2014.09.02 NEC CORP
  • US8824317B2 patent drawing
  • US8824317B2 patent drawing
  • US8824317B2 patent drawing

AI summary

A parallel calculation system includes a plurality of functional nodes and a plurality of network switches, which serve as part of a communication path for communication between the functional nodes and function as communication channels. As for the functional nodes, a functional node group made up of part or all of a plurality of the functional nodes that the parallel calculation system includes is managed as one group. The network switches include a plurality of input/output ports, which are ports for inputting and outputting the communication request, and a switch, which outputs the communication request that the input/output port inputs to the input/output port. The network switches manage a communication bandwidth for each of the communication channels between the functional nodes and control the switch.