Shared Database Storage Node Mapping for Elasticity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed database systems using shared storage face challenges in balancing elasticity and query processing performance, as preplacement of data records for efficient operations can affect the system's ability to adjust node assignments and lead to performance issues during reshuffling.
Innovation Solution
Implementing a data-agnostic mechanism with a hash-based mapping that assigns data records to segments of shared storage and maps nodes to segments in a many-to-many manner, allowing for efficient query processing without reshuffling data records, thereby enhancing both elasticity and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data records are preplaced in co-located regions of shared storage for efficient query processing, then query processing performance is improved, but system elasticity deteriorates due to the inability to easily adjust node assignments
Solution Approach 1:
The patent divides the shared storage into multiple segments and implements a many-to-many mapping between nodes and segments. This segmentation allows data records to be strategically grouped in co-located regions for efficient query processing while maintaining the ability to adjust node assignments by remapping nodes to different segments without requiring data reshuffling.
Solution Approach 2:
The patent implements a dynamic mapping mechanism that allows the system to flexibly assign nodes to storage segments based on current workload and performance requirements. This dynamic assignment enables easy adjustment of node assignments while maintaining optimal data placement, thereby preserving both query processing performance and system elasticity.
2Adaptability or versatility
If nodes are removed or added to balance user load, then system adaptability is improved, but query processing performance deteriorates due to required data reshuffling
Solution Approach 1:
The patent pre-establishes a many-to-many mapping structure between nodes and storage segments that anticipates future node additions or removals. This preliminary configuration allows the system to handle node changes without requiring data reshuffling, as the mapping can be dynamically adjusted to accommodate new or removed nodes while maintaining optimal data placement.
Solution Approach 2:
The patent implements a mapping layer that acts as a virtual copy between physical storage segments and logical node assignments. This mapping copy allows the system to reassign nodes to different segments without physically moving or reshuffling the actual data records, thereby maintaining query processing performance during node changes.
3Productivity
If data records are reshuffled for certain query processing operations, then query processing performance is improved, but system complexity increases due to the need for data movement and reassignment
Solution Approach 1:
The patent introduces a mapping mechanism as an intermediary layer between physical storage segments and logical node assignments. This intermediary mapping structure eliminates the need for complex data reshuffling operations by allowing the system to achieve optimal query processing performance through logical reassignment rather than physical data movement, thereby reducing system complexity.
Data Source
AI summary
A technique includes applying a first mapping to associate a plurality of data records for a plurality of database nodes with a plurality of segments of a storage that is shared by the database nodes; and applying a many-to-many second mapping to associate the segments with the database nodes.