In-Memory Database Root Node Hotspot Elimination via Mapping Page Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In in-memory databases, the root node becomes a hotspot during query operations, leading to low concurrent processing efficiency due to the write-once read-many concurrency mechanism, where the root node can only be written by one transaction, preventing other transactions from reading simultaneously.
Innovation Solution
A method and apparatus that modify the root node by using a data area and a general area with mapping pages, allowing the active mapping page to handle queries while the root node is updated, eliminating the need for locks on the root node during modifications, and switching the non-active mapping page to become active after updates are completed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the write-once read-many concurrency mechanism is used on the root node, then data consistency is ensured, but concurrent processing efficiency deteriorates because only one transaction can write while others must wait
Solution Approach 1:
The root node is segmented into multiple mapping pages (first mapping page, second mapping page, etc.), where each mapping page can independently handle query requests. This segmentation allows multiple transactions to access different mapping pages simultaneously, improving concurrent processing efficiency while maintaining data consistency through version control mechanisms.
Solution Approach 2:
Multiple mapping pages are pre-created and maintained in the database before any query operations occur. These mapping pages are prepared in advance with valid root node data, allowing query requests to be immediately routed to appropriate mapping pages without waiting for root node updates to complete, thus eliminating the bottleneck.
2Reliability
If the root node is locked for writing, then data consistency is maintained, but query operations from other transactions are blocked
Solution Approach 1:
Mapping pages serve as intermediary structures between query requests and the root node. Instead of directly accessing the root node, query requests are routed through mapping pages that contain cached root node data. This intermediary mechanism allows queries to proceed without acquiring locks on the root node, eliminating blocking while maintaining data consistency through periodic synchronization.
3Device complexity
If the root node is used as the sole query entrance, then data access is simplified, but the root node becomes a hotspot causing performance degradation
Solution Approach 1:
The single root node entrance is segmented into multiple mapping pages that distribute query traffic. Each mapping page acts as an independent entrance point, allowing query requests to be parallelized across multiple pages. This segmentation reduces the concentration of access operations on a single node, eliminating the hotspot effect while maintaining simplified data access through the mapping page abstraction layer.
Data Source
AI summary
Embodiments of the present invention provide a method for modifying a root node, and a modification apparatus. The method for modifying a root node of the present invention includes: receiving a root node modification request; modifying a root node in a data area of a database according to the modification information; replicating the modified root node to the general area of the database; finding, in the general area of the database, that an identifier of an active mapping page is the identifier of the first mapping page, and determining that the first mapping page is the active mapping page; and replacing the second mapping page with the modified root node according to the modified root node that is replicated to the general area, and modifying the identifier of the active mapping page to the identifier of the second mapping page. The present invention improves concurrent processing efficiency of indexing.