In-Memory Database Integration in DBMS
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current database management systems are not designed to support running an entire database fully in-memory without on-disk storage and are not integrated tightly with the database server engine, leading to performance limitations.
Innovation Solution
A system and method for creating an in-memory database fully integrated within a traditional disk-resident database management system, where a specialized database type is specified as a parameter of a native DBMS command, allowing the database to be hosted entirely in-memory without any on-disk storage, using named caches and in-memory devices that provide full run-time transactional support with no I/O to disk.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a traditional disk-resident database management system is used, then the system is well-established and reliable, but data access speed is limited by disk I/O performance
Solution Approach 1:
The invention segments the database into two parts: a small system catalog stored on disk and a large user database stored in-memory. This segmentation allows the majority of data to be accessed at memory speed while maintaining the reliability of disk-based system management, thus improving data access speed without requiring complete architectural complexity change
Solution Approach 2:
The invention introduces a memory-resident database engine as an intermediary layer between the disk-based system catalog and user applications. This intermediary manages the in-memory database using fast memory access while the system catalog on disk provides persistent storage and recovery capabilities, resolving the contradiction between speed and complexity
2Ease of operation
If stand-alone in-memory database products are used, then fast data access is achieved, but integration with existing database server engine is insufficient
Solution Approach 1:
The invention merges the in-memory database engine with the existing database server engine by implementing it as a native component within the server architecture. This integration allows the database server to directly manage memory-resident databases using its existing query processor, transaction manager, and security mechanisms, thus achieving both good integration and high performance
Solution Approach 2:
The invention makes the database server engine universal by enabling it to handle both traditional disk-based databases and new in-memory databases using the same core engine. The server can dynamically switch between storage modes based on user needs, providing multi-functionality that maintains ease of operation while improving productivity
3Productivity
If LRU/MRU cache protocol is used, then memory management is simplified, but data access efficiency is limited
Solution Approach 1:
The invention extracts the memory management complexity from the data access path by implementing a simplified first-fit allocation strategy. Instead of using complex LRU/MRU protocols that track and reorder every memory reference, the system uses straightforward memory allocation where the first available memory block is used, thus improving data access efficiency while reducing management complexity
Data Source
AI summary
System, method, computer program product embodiments and combinations and sub-combinations thereof for managing data storage as an in-memory database in a database management system (DBMS) are provided. In an embodiment, a specialized database type is provided as a parameter of a native DBMS command. A database hosted entirely in-memory of the DBMS is formed when the specialized database type is specified.