GPU-Based Event Matching for Complex Event Processing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current complex event processing (CEP) systems face challenges in achieving high scalability and performance while processing large volumes of events efficiently, as they are complex and expensive, with significant time and resource consumption in the matching and processing engine, especially when using conventional hardware like chip-multiprocessors.

Innovation Solution

Implementing event matching algorithms on graphical processing units (GPUs) to leverage their high parallel processing capabilities, which enables faster and more economical processing of large numbers of events compared to conventional chip multi-processors or grid computing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional chip-multiprocessors or grid computing are used for event matching, then processing capability is improved, but system complexity and cost increase

Engineering Contradiction:

Improveevent processing throughputVSAvoidsystem complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent replaces conventional CPU-based event matching mechanisms with a GPU-based parallel processing mechanism. The GPU's massively parallel architecture naturally suits the event matching workload, eliminating the need for complex multi-threading and synchronization logic required by chip-multiprocessors, thus reducing system complexity while maintaining high throughput

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the fundamental processing parameter from sequential/CPU-based execution to parallel/GPU-based execution. By leveraging the GPU's thousands of cores operating in parallel, the system achieves higher throughput without increasing logical complexity, as the parallelism is handled by the hardware architecture rather than software orchestration

Inventive Principle:

Principle #35Parameter changes

2Productivity

If conventional chip-multiprocessors are used for event matching, then processing speed is improved, but energy consumption increases

Engineering Contradiction:

Improveevent processing speedVSAvoidenergy consumption

Core Design Contradiction:

ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent substitutes energy-inefficient CPU-based processing with energy-efficient GPU-based processing. The GPU's architecture is optimized for parallel workloads like event matching, achieving higher throughput per watt by utilizing its massive parallel core count and efficient memory hierarchy, thereby reducing overall energy consumption while improving processing speed

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If conventional hardware is used for event matching, then implementation simplicity is maintained, but processing efficiency decreases

Engineering Contradiction:

Improveprocessing efficiencyVSAvoidimplementation simplicity

Core Design Contradiction:

ProductivityVSEase of manufacture

Solution Approach 1:

The patent leverages the GPU's universal parallel processing capability that can handle event matching workloads efficiently. The same GPU hardware can be used for various parallel computing tasks, and by utilizing existing GPU programming frameworks and libraries, the implementation remains relatively simple despite the advanced hardware capabilities being employed

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS9317456B2Method and system for performing event-matching with a graphical processing unit

Publication Date: 2016.04.19 INFOSYS LTD

AI summary

A computer-implemented method for event matching in a complex event processing system includes receiving, with a computer processing device, a stream of event data; receiving, with a computer processing device, an event list and an access predicate list, wherein the event list includes one or more event data pairs; and identifying, with a graphical processing device, patterns in the stream of event data.