Unified Domain Model View via Synchronized Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional information retrieval systems face challenges in efficiently updating indexes without compromising search engine performance, particularly in full-text searching, and often return different object instances for the same data entry when using different query methods, leading to inconsistent views of data.
Innovation Solution
A unified application framework with a master search node and slave search nodes maintains a synchronized index across multiple databases, allowing asynchronous updates and replication, ensuring consistent indexing and unified object views through Object-Relational Mapping (ORM) and full-text search integration, using APIs like JDBC and JMS for efficient data management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional full-text search indexing is used, then document retrieval capability is improved, but index update performance deteriorates
Solution Approach 1:
The system divides the indexing workload into separate full-text indexing operations and direct query operations. The full-text index is maintained independently from the database tables, allowing updates to proceed without blocking direct queries. This segmentation enables parallel processing where index updates and query operations can occur simultaneously, resolving the performance conflict between retrieval capability and update speed.
Solution Approach 2:
The system performs preliminary indexing of document contents into the full-text index before they are needed for retrieval. By pre-processing and storing inverted indexes in advance, the system ensures that document retrieval can proceed efficiently without performing intensive processing at query time, while index updates can be batched and applied without compromising ongoing query performance.
2Adaptability or versatility
If different query methods are used, then search flexibility is improved, but data view consistency deteriorates
Solution Approach 1:
The system introduces an intermediary layer that translates both full-text queries and direct queries into a unified object model. This intermediary ensures that regardless of the query method used, the application receives consistent object representations from the database. The intermediary layer harmonizes different query approaches by normalizing their results through a common data access interface, thereby maintaining view consistency while preserving search flexibility.
Solution Approach 2:
The system implements a universal query interface that handles both full-text search and direct queries through a single unified mechanism. This multi-functional approach allows the same data access layer to serve multiple query types, ensuring that all query methods return objects from the same underlying data model. By making the query system universal, the patent eliminates the inconsistency between different query methods while maintaining the flexibility to support various search approaches.
3Reliability
If index updates are performed frequently, then data freshness is improved, but system performance deteriorates
Solution Approach 1:
The system implements periodic batch updates of the full-text index rather than continuous real-time updates. Documents are indexed in periodic batches, allowing the system to maintain data freshness at acceptable intervals while avoiding the performance degradation associated with continuous update operations. This periodic approach enables the system to balance between keeping index data fresh and maintaining overall system performance by consolidating update operations.
Solution Approach 2:
The system creates copies of document data for indexing purposes, separating the indexing copy from the original database storage. By copying document contents into the full-text index structure, the system can update the index without modifying the original database tables, thereby maintaining data freshness in the index while preserving system performance. The copying mechanism allows independent index maintenance without impacting the performance of direct database operations.
Data Source
AI summary
Techniques for providing a unified view of a domain model to a user are described herein. In one embodiment, in response to a first search query received from a client via a first search mechanism (e.g., outside of the relational DB) for a list of persistent objects representing data entries of a relational database, it is determined whether the persistent objects have been accessed via a second search query via a second search mechanism based on an object identifier of the persistent object. If the requested persistent object has been accessed via a second search query, an identical instance of the persistent object is returned to the client as a result of the first search query. Other methods and apparatuses are also described.