Cluster Application Manager for Automated Software Installation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Manual software installation in big data clusters is complex, tedious, and error-prone due to the lack of standardized installation commands and procedures across different software applications, making it difficult for cluster administrators to manage and maintain large numbers of cluster machines efficiently.

Innovation Solution

A cluster application manager system that includes an input interface for receiving management requests, a command determiner to identify base commands, a node determiner to select appropriate cluster nodes, and an output interface to provide management commands, utilizing software wrappers to translate base commands into application-specific commands and manage software installation, uninstallation, upgrades, and status checks across multiple nodes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If manual installation commands are issued on each cluster machine, then software can be installed on individual nodes, but the process becomes complex, tedious, slow and error-prone

Engineering Contradiction:
Improveease of software installationVSAvoidinstallation time
Core Design Contradiction:
Ease of operationVSLoss of time

Solution Approach 1:

The system enables self-service automated installation where the cluster administrator defines installation parameters once at the cluster level, and the system automatically distributes and executes installation commands across all relevant nodes without manual intervention on each machine

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The installation process is segmented into distinct phases: parameter definition at cluster level, automatic command generation, distributed execution on target nodes, and result aggregation. This segmentation allows each phase to be handled independently and automatically

Inventive Principle:
Principle #1Segmentation

2Productivity

If standardized installation commands are implemented across all software, then management efficiency improves, but flexibility to handle application-specific installation procedures is reduced

Engineering Contradiction:
Improvemanagement efficiencyVSAvoidadaptability to different software
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The system provides a universal installation framework that can handle multiple types of software installations through a common interface. The cluster-level parameter definition and automated command distribution mechanism works across different applications while allowing application-specific customization through configuration files or scripts

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

Solution Approach 2:

An intermediary layer is introduced between the standardized command interface and application-specific installation procedures. This intermediary translates generic installation requests into application-specific commands while maintaining the benefits of standardization at the management level

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If manual installation procedures are used for each application, then specific application requirements can be met, but the complexity and error rate increase significantly

Engineering Contradiction:
Improveinstallation reliabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

Multiple installation procedures for different applications are merged into a unified automated process. The system combines cluster-level parameter management with node-level execution, merging the benefits of centralized control with distributed implementation to reduce errors while managing complexity

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS10833915B2Application plugin framework for big-data clusters
Publication Date: 2020.11.10 EMC IP HLDG CO LLC
  • US10833915B2 patent drawing
  • US10833915B2 patent drawing
  • US10833915B2 patent drawing

AI summary

A system for a cluster application manager includes an input interface, a command determiner, and a node determiner. The input interface is for receiving a first management request for a management function for a first application of a set of different applications. The command determiner is for determining a first management base command for the first management request. A second management base command is the same as the first management base command for a second management request for the management function for a second application of the set of different applications that is different from the first application of the set of different applications. The command determiner is for determining a management command, wherein the management command is based at least in part on the first management base command and the first application. The node determiner is for determining a node of a cluster to send the management command.