Extended Distribution Unit DMA Data Transfer

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing distributed computer systems face challenges in improving response time behavior due to inefficiencies in data processing and transmission, particularly in time-triggered communication systems where data is buffered and queued, leading to delays.

Innovation Solution

The integration of Direct Memory Access (DMA) units between the switching memory and application memories, controlled by the switching engine, allows for direct and efficient data transfer between memory areas, enabling both time-controlled and event-controlled message processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data is buffered and queued in the switching memory for transmission, then data can be temporarily stored and managed, but the response time behavior deteriorates due to delays in data transmission

Engineering Contradiction:
Improvedata transmission reliabilityVSAvoidresponse time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts the data transfer function from the central switching engine and implements it directly in the application computers via DMA units. This removes the bottleneck of centralized switching memory access and allows application computers to directly access and transfer data from source to destination without queuing delays in the switching memory.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces DMA units as intermediary components between the switching memory and application computer memories. These DMA units enable direct memory-to-memory transfer controlled by the switching engine, bypassing the need for data to be queued in the switching memory and significantly reducing transmission delays.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Ease of operation

If a central switching engine handles all data distribution, then centralized control is achieved, but the data processing speed deteriorates due to sequential handling

Engineering Contradiction:
Improvecentralized controlVSAvoiddata processing speed
Core Design Contradiction:
Ease of operationVSSpeed

Solution Approach 1:

The patent segments the centralized data transfer function into distributed DMA units located in each application computer. Each DMA unit can independently handle data transfers under control of the switching engine, allowing parallel processing of multiple data streams simultaneously while maintaining centralized scheduling control.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent adds a new dimension of parallelism by enabling multiple application computers to perform data transfers simultaneously through their local DMA units. This transforms the single-threaded sequential processing of the central switching engine into multi-threaded parallel processing across the distributed system.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Adaptability or versatility

If data is transferred through the switching memory for all communications, then centralized routing is maintained, but the data transport interval increases due to memory access delays

Engineering Contradiction:
Improverouting flexibilityVSAvoiddata transport interval
Core Design Contradiction:
Adaptability or versatilityVSDuration of action of moving object

Solution Approach 1:

The patent implements preliminary action by having the switching engine pre-configure DMA transfer parameters and initiate DMA operations before actual data transfer is needed. The DMA units are ready to execute transfers immediately when data becomes available, eliminating the delays associated with on-the-fly switching memory access and routing decisions.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3057275B1Extended distribution unit
Publication Date: 2020.08.05 TTTECH COMPUTERTECHNIK AG
  • EP3057275B1 patent drawingFigure 1~2
  • EP3057275B1 patent drawingFigure 3

AI summary

The invention relates to a method for processing real-time data in a distribution unit of a distributed computer system, the computer system comprising a plurality of node computers and distribution units, wherein the distribution unit, in addition to a switching engine (SE) (130) and a switching memory (SS) (140), contains one or more application computers (220, 230) each with one or more application central units (223, 233) and application memories (AS) (222, 232), wherein, upon arrival at one of its ports of a message intended for an application computer (220, 230), the switching engine (130) of the distribution unit forwards this message to the application computer (220, 230) via a direct memory access (DMA) unit (221, 231), which is arranged between the switching memory (140) and the application memory (222, 232) of the addressed application computer (220, 230) and is under the control of the switching engine (130). addressed application computer (220,230). The invention further relates to an extended distribution unit and a computer system with such extended distribution units.