Selective Indexing for Multi-Tiered Database Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Creating and maintaining multiple indexes in multi-tiered databases that span both memory and disk storage requires significant processing power and memory resources, which can be wasteful when not all data is frequently accessed, leading to reduced system throughput.
Innovation Solution
A database manager system that allows users to selectively create and maintain indexes on specific portions of data stored in memory or disk, optimizing resource usage by enabling users to choose which data to index based on access patterns, and using flags to prevent duplicate index creation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If multiple indexes are created for each storage location in multi-tiered databases, then data access efficiency is improved, but system resource consumption (memory and processing power) increases
Solution Approach 1:
The patent applies local quality by creating indexes selectively on specific storage locations based on data access patterns. Instead of uniformly indexing all data in memory and disk, the system identifies which portions of data are frequently accessed and creates indexes only on those specific locations, making the indexing strategy adaptive to local access characteristics rather than applying a blanket approach across all storage tiers.
Solution Approach 2:
The patent implements partial action by creating indexes on only a subset of data locations rather than all possible storage locations. The system determines which indexes are necessary based on query patterns and creates only those indexes that provide actual benefit, avoiding the excessive creation of indexes on all storage locations regardless of their utility.
2Speed
If multiple indexes are maintained across memory and disk storage, then query performance is improved, but device complexity increases
Solution Approach 1:
The patent applies self-service by implementing automatic index management capabilities within the database system. The database manager automatically monitors query patterns, identifies which indexes would be beneficial, creates appropriate indexes, and maintains them without requiring manual intervention. This automation reduces the operational complexity of managing multiple indexes across different storage locations.
Solution Approach 2:
The system uses feedback mechanisms by monitoring query patterns and access statistics to dynamically determine which indexes should be created or maintained. The index creation strategy is informed by feedback from actual database usage patterns, allowing the system to adapt its indexing structure based on observed query behavior rather than using static predetermined rules.
3Speed
If indexes are created on all data portions, then data retrieval speed is improved, but processing power consumption increases
Solution Approach 1:
The patent implements partial action by creating indexes only on portions of data that are actually frequently accessed, rather than indexing all data portions uniformly. The system analyzes query patterns to identify which data locations benefit from indexing and applies indexes selectively, avoiding the waste of processing power on creating and maintaining indexes for data that is rarely accessed.
Solution Approach 2:
The patent applies parameter changes by dynamically adjusting the indexing strategy based on observed data access patterns. The system monitors query characteristics and modifies which indexes are created or maintained according to changing access patterns, allowing the processing power investment in indexing to be optimized based on actual usage parameters rather than using a fixed indexing approach.
Data Source
AI summary
Disclosed herein are system, method, and computer program product embodiments for providing index and storage management for multi-tiered databases. An embodiment operates by receiving a request to create an index on a multi-tiered database including both an in-memory store and a disk store. A multi-store table associated with the index is determined, wherein the multi-store table includes both a first set of data stored on the memory store and a second set of data stored on the disk store. Either the first set of data or the second set of data on which to create the index is selected based on the request. The index for the selected set of data of the multi-store table is generated. The index is stored on either the disk store or the memory store as corresponding to the selected set of data for which the index was generated.