Multi-tenancy Database Content Patching via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-tenancy database systems face challenges in efficiently deploying content and managing changes across shared and tenant containers, particularly in maintaining data consistency and reducing downtime during upgrades and patches, while ensuring compatibility with existing applications.
Innovation Solution
The system employs a method of deploying content by identifying shared and tenant-specific data, using a shared database container for common data and tenant database containers for tenant-specific data, and exchanging the shared database container during upgrades, allowing individual tenant upgrades with minimal downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If content is deployed to both shared and tenant containers during upgrades, then data consistency is maintained, but deployment complexity and downtime increase
Solution Approach 1:
The patent segments content into three distinct categories: shared content (deployed only to shared container), local content (deployed only to tenant containers), and partially-shared content (deployed to both). This segmentation allows independent deployment strategies for each category, reducing overall deployment complexity while maintaining data consistency through selective synchronization.
Solution Approach 2:
The patent introduces a deployment system that acts as an intermediary, automatically determining content classification and deployment targets. This intermediary layer manages the complexity of coordinating between shared and tenant containers, shielding users from the underlying complexity while ensuring consistent data propagation.
2Reliability
If the entire database container is exchanged during upgrades, then data consistency is ensured, but downtime increases
Solution Approach 1:
The patent extracts and separates shared content from tenant-specific content, allowing the shared container to be updated independently. By taking out only the necessary shared content for deployment and leaving tenant content untouched, the system achieves data consistency for shared resources without requiring full container exchange, thereby minimizing downtime.
Solution Approach 2:
The patent performs preliminary classification of content into shared, local, and partially-shared categories before deployment begins. This preliminary action enables the deployment system to prepare only the necessary shared content for exchange, avoiding the need to exchange entire containers and reducing the time systems remain unavailable.
3Adaptability or versatility
If all content is deployed to each tenant container, then tenant-specific functionality is preserved, but deployment time and resource usage increase
Solution Approach 1:
The patent applies local quality by deploying content selectively based on its classification: shared content goes to shared containers, local content goes only to tenant containers, and partially-shared content goes to both. This targeted approach ensures each tenant container receives only the specific content it needs, preserving tenant-specific functionality while dramatically reducing deployment time and resource consumption compared to deploying all content universally.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure involves systems, software, and computer implemented methods for patching content in database systems. One example method includes receiving a deployment package for an upgrade of a database system to a new version. Content is deployed to a shared container. Shared objects are identified that are completely stored in the shared container. First shared content is determined for shared objects. Partially-shared objects are identified that have a shared portion in the shared container and a tenant portion in one or more tenant containers. Second shared content is determined for partially-shared objects. The first shared and second shared content are deployed to the shared container. Local content is determined for the partially-shared objects and deployed to each tenant container. Local objects that have no data in the shared container are identified. Local content is identified for the local objects and deployed to each respective tenant container.