Metadata-Driven Cluster Node Update Orchestration

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing mechanisms for updating computing devices do not consider the relationships between devices in a cluster, leading to inefficiencies and potential negative impacts during the update process.

Innovation Solution

A method and system that utilize metadata from controllers to manage nodes in a cluster by identifying solution groups and processing sequences, enabling infrastructure managers to orchestrate updates while solution managers perform pre- and post-update tasks, thereby minimizing disruption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional update mechanisms are used without considering device relationships, then the update process is simple and quick, but the reliability of the cluster is compromised due to potential negative impacts from ignoring inter-device relationships

Engineering Contradiction:

Improvecluster update reliabilityVSAvoidupdate management complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by obtaining metadata from controllers that describes relationships between computing devices before the update process begins. This metadata is processed to identify solution groups and processing sequences, allowing the system to plan and execute updates in a way that maintains cluster reliability while managing complexity through pre-computed groupings and sequences.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If updates are performed without considering solution groups and processing sequences, then the update process is faster, but the disruption to the cluster increases

Engineering Contradiction:

Improveupdate speedVSAvoidcluster disruption

Core Design Contradiction:

ProductivityVSObject-affected harmful factors

Solution Approach 1:

The system segments the cluster into solution groups based on processed metadata, allowing updates to be applied to specific groups in a controlled sequence rather than uniformly across all devices. This segmentation enables faster updates within each group while minimizing overall cluster disruption through staged deployment.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system implements periodic action by updating solution groups in a defined processing sequence rather than all at once. This allows the cluster to maintain stability between update waves while progressively applying changes, balancing update speed with minimal disruption to ongoing operations.

Inventive Principle:

Principle #19Periodic action

3Ease of operation

If metadata processing is performed to identify solution groups and processing sequences, then the update coordination is improved, but the processing time and computational resources increase

Engineering Contradiction:

Improveupdate coordinationVSAvoidmetadata processing time

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The system performs metadata processing as a preliminary action before the actual update execution. By obtaining and processing metadata early to identify solution groups and processing sequences, the system establishes a clear update roadmap that simplifies subsequent coordination while limiting the time penalty to the initial processing phase rather than extending it through the entire update process.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11216270B2Metadata driven workflow semantics for management operations

Publication Date: 2022.01.04 DELL PROD LP

AI summary

In general, in one aspect, the invention relates to a method for managing nodes in a cluster. The method includes obtaining metadata from controllers on a node, processing the metadata to obtain solution groups and a processing sequence of the solution groups, notifying a solution manager to initiate performance of pre-update tasks associated with the node, and performing, after the notifying, an update of the nodes in the cluster based on the processing sequence of the solution groups.