Logical Representation for Hybrid Analytic Flow Management

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In heterogeneous analytic environments, managing hybrid analytic flows that comprise sub-flows directed to different execution engines is challenging due to engine-specific requirements and limitations in existing technologies, which hinder optimization and execution across multiple engines.

Innovation Solution

A method is introduced to generate a logical, engine-agnostic representation of physical analytic flows using a dictionary that maps logical operators to various execution engines, allowing for conversion and optimization of hybrid flows into executable forms suitable for multiple engines.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If physical analytic flows are designed using engine-specific programming languages or flow-design tools, then execution compatibility with specific engines is achieved, but manageability and optimization across heterogeneous engines become difficult

Engineering Contradiction:

Improveexecution compatibilityVSAvoidmanageability

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary logical representation layer that sits between the physical flow design and multiple execution engines. This logical representation serves as a mediator that translates engine-specific physical flows into a unified format, enabling cross-engine optimization and management while preserving execution compatibility with heterogeneous engines through dictionary-based mappings.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If engine-specific physical flows are used, then execution efficiency for specific engines is optimized, but portability and reusability across different engines are limited

Engineering Contradiction:

Improveexecution efficiencyVSAvoidportability

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The logical representation is designed to be universal and engine-agnostic, capable of representing analytic flows that can be executed on multiple different engines. Through dictionary mappings, the same logical representation can be translated into various engine-specific formats, enabling a single flow design to serve multiple execution environments while maintaining execution efficiency for each target engine.

Inventive Principle:

Principle #6Universality (Multi-functionality)

3Adaptability or versatility

If hybrid flows comprising sub-flows for different engines are created, then functional versatility is achieved, but optimization and management complexity increase

Engineering Contradiction:

Improvefunctional versatilityVSAvoidoptimization complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the hybrid flow into distinct logical operators that can be independently analyzed and optimized. Each logical operator in the logical representation corresponds to a specific functional unit that can be mapped to appropriate execution engines, allowing for modular optimization strategies and simplified management of complex hybrid flows through the unified logical layer.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentEP3014472B1Generating a logical representation from a physical flow

Publication Date: 2018.04.18 MICRO FOCUS LLC

Figure 1
Figure 2
Figure 3

AI summary

Described herein are techniques for processing analytic flows. A physical flow can be parsed into multiple elements. A logical representation can be generated based on the physical flow using a dictionary. An engine-specific logical representation or second physical flow can then be generated based on the logical representation using the dictionary.