Industrial Analytics Framework for Automated Kubernetes Deployment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing analytics frameworks for deploying and executing analytics packages on distributed computing systems are inefficient, time-consuming, and complex, making it difficult for inexperienced users to utilize them effectively.

Innovation Solution

An improved analytics framework that automates the deployment and execution of analytics packages by generating a container image, allocating computing resources, and configuring service pods based on deployment parameters, using REST API calls to facilitate seamless integration and execution on a Kubernetes cluster.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual deployment processes are used for analytics packages on distributed computing systems, then users have control over deployment details, but deployment time and complexity increase significantly

Engineering Contradiction:
Improveease of deploymentVSAvoiddeployment time
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The deployment framework enables self-service deployment by automatically generating service pods and allocating computing resources without requiring manual user intervention. The system autonomously processes deployment requests, manages container orchestration, and configures analytics packages, allowing users to simply submit deployment parameters while the framework handles all complex deployment operations automatically.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The framework performs preliminary actions by pre-configuring deployment templates, pre-allocating resource pools, and pre-compiling analytics packages into executable formats. Service pod templates and resource allocation policies are prepared in advance, enabling rapid deployment when requests are received without requiring real-time complex processing.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If automated deployment frameworks are implemented, then deployment speed and ease of use improve, but system complexity increases

Engineering Contradiction:
Improvedeployment efficiencyVSAvoidframework complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The deployment framework acts as an intermediary layer between users and the underlying distributed computing system. It provides abstraction mechanisms that hide complex orchestration logic, resource management details, and container configuration from users. The framework translates simple deployment requests into complex backend operations, shielding users from complexity while maintaining high deployment efficiency.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The framework implements universal deployment capabilities that can handle multiple types of analytics packages, various computing resource types, and different deployment scenarios through a single unified interface. Service pod templates and resource allocation policies are designed to be universally applicable across different analytics workloads, reducing the need for multiple specialized deployment systems and thereby managing complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

3Use of energy by moving object

If computing resources are dynamically allocated based on deployment parameters, then resource utilization efficiency improves, but resource allocation complexity increases

Engineering Contradiction:
Improvecomputing resource utilizationVSAvoidresource allocation complexity
Core Design Contradiction:
Use of energy by moving objectVSDevice complexity

Solution Approach 1:

The framework implements dynamic resource allocation that adapts computing resource distribution based on real-time deployment parameters and workload characteristics. Service pod configurations are dynamically adjusted according to resource availability, analytics package requirements, and system load conditions. This dynamic approach optimizes computing resource utilization by allocating resources precisely where and when they are needed, rather than using static pre-allocation schemes.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4016284B1Framework for industrial analytics
Publication Date: 2025.09.03 BAKER HUGHES CO
  • EP4016284B1 patent drawingFigure 1
  • EP4016284B1 patent drawingFigure 2
  • EP4016284B1 patent drawingFigure 3

AI summary

A method includes receiving data characterizing an analytics package, and generating, by an analytics framework associated with a plurality of compute nodes, a container image associated with the analytics package and a unique identifier indicative of the container image. The container image is saved in a central container registry. The method further includes receiving, from a client, data characterizing deployment parameters associated with the deployment of the container image on the plurality of compute nodes and the unique identifier indicative of the container image. The method also includes generating at least one analytics service pod based on the deployment parameters and the unique identifier. The at least one analytics service pod includes the container image. The at least one analytics service pod is configured to execute the analytics package on one or more compute nodes of the plurality of compute nodes based on the deployment parameters. The deployment parameters include computing resource associated with execution of the at least one analytics service pod on the plurality of compute nodes.