Materialized View Segmentation for Multi-Tenant Data Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multitenant cloud databases, managing materialized views efficiently to reduce data storage costs and computing overhead is challenging due to the need for frequent updates and synchronization across multiple tenants, leading to increased storage and processing costs.
Innovation Solution
A system that monitors data changes in a primary data store, analyzes their impact on materialized views, and updates shared and independent views accordingly, using a view selector to manage data movement and storage based on a similarity factor to optimize storage and reduce unnecessary updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If materialized views are updated frequently to maintain data consistency across multiple tenants, then data reliability is improved, but storage costs and computing overhead increase
Solution Approach 1:
The patent segments materialized views into shared views (accessible by multiple tenants) and private views (specific to individual tenants). This segmentation allows the system to update only the necessary portions of materialized views when data changes occur, rather than performing full refreshes across all views. The change data feed mechanism further segments updates by identifying only the specific data changes that affect particular materialized views, reducing unnecessary computing overhead while maintaining data consistency.
2Quantity of substance
If shared materialized view stores are used to reduce storage costs, then storage efficiency is improved, but update complexity increases
Solution Approach 1:
The patent introduces a shared materialized view store as an intermediary layer between the primary data store and multiple tenant materialized views. This intermediary consolidates shared data that is common across multiple tenants, reducing overall storage requirements. The system manages update complexity by using a change data feed that monitors the primary data store and automatically determines which shared materialized views need updates based on the specific data changes detected, rather than requiring complex manual update management.
3Reliability
If data changes are propagated to all materialized views, then data freshness is improved, but processing time increases
Solution Approach 1:
The patent implements preliminary action by maintaining a change data feed that continuously monitors the primary data store for changes. When data changes are detected, the system immediately identifies and updates only the affected materialized views rather than waiting for scheduled refresh cycles. This preliminary detection and selective update mechanism ensures data freshness is maintained while minimizing processing time by avoiding unnecessary updates to unaffected views.
Data Source
AI summary
Shared materialized views are maintained during data changes to the primary data and during creation of new materialized views. Shared data stored for use by shared materialized views is distinguished from data stored by an independent materialized view. A view selector manages data updates to shared materialized views and corresponding mapping table. The view selector directs movement of data between a shared materialized view and an independent materialized view through the lifecycle of the materialized views.