Golden Image Management for Distributed System Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed computing systems face challenges in managing software images across computing nodes, particularly during node allocation, software updates, and maintenance, which often require shutting down nodes and can severely impact system performance.
Innovation Solution
A control node generates a 'golden image' for each software type and creates pre-configured 'image instances' for potential nodes, allowing for efficient allocation and management, including updates, without requiring all nodes to be online simultaneously.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If software updates are applied to computing nodes, then software versions are synchronized across the system, but system downtime increases and computing performance deteriorates
Solution Approach 1:
The control node creates image instances from the master software image in advance, before they are needed for deployment. This preliminary creation of image instances allows computing nodes to be updated without real-time software compilation or installation delays, thereby maintaining system productivity while ensuring software version consistency across all nodes.
2Ease of manufacture
If computing nodes are shut down for software installation, then software updates can be applied, but system availability decreases
Solution Approach 1:
Image instances are pre-generated and stored in the image repository before deployment is required. This allows computing nodes to boot directly from pre-prepared image instances without needing to shut down for software installation, thereby maintaining system availability while enabling easy software updates through the control node's image management capability.
Solution Approach 2:
The control node creates copies of the master software image in the form of image instances. These copies can be deployed to multiple computing nodes simultaneously without requiring the nodes to be shut down, as the copying process occurs on the control node rather than on the computing nodes themselves, thus maintaining system availability while enabling software updates.
3Reliability
If multiple computing nodes are updated simultaneously, then software consistency is maintained, but system-wide downtime increases
Solution Approach 1:
The software image is segmented into a master software image stored on the control node and multiple independent image instances stored in the image repository. This segmentation allows the control node to manage and distribute software updates independently to multiple computing nodes without requiring them to be updated simultaneously, thereby reducing total downtime while maintaining software consistency across the distributed system.
Data Source
AI summary
In general, the invention is directed to updating images running on computing nodes within a distributed computing system. For example, a distributed computing system comprises a plurality of application nodes, a software image repository and a control node interconnected via a network. The image repository stores a master software image and a plurality of software image instances generated from the master software image. The control node automatically updates each of the software image instances using an updated version of the master software image.