Federated Database Search with Predictive Prefetching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current procurement systems face delays in responding to search queries due to reliance on static catalogs and time-consuming dynamic data collection from supplier systems, leading to non-current and non-comprehensive search results.
Innovation Solution
A federated database search system that performs predictive searches by anticipating search queries, prefetching results, and simulating user actions to collect product data from supplier systems, allowing for up-to-date results to be stored locally, thereby reducing response time and improving search result quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dynamic product data is collected from each supplier system in response to search queries, then the search result quality is improved, but the response time increases due to the bottleneck of time-consuming data collection
Solution Approach 1:
The system performs predictive searches by anticipating potential search queries and prefetching product data from supplier systems before actual search requests are made. This preliminary data collection and caching mechanism ensures that when a real search query arrives, the data is already available or nearly available, dramatically reducing response time while maintaining comprehensive and current search results.
2Loss of time
If static catalogs are used for product data, then the response time is reduced, but the search results become non-current and non-comprehensive
Solution Approach 1:
The system transitions from static catalogs to a dynamic data collection approach where product data is continuously updated from supplier systems. Predictive searches trigger data prefetching and caching, creating a dynamic balance between data freshness and response speed. The system adaptively determines when to collect data based on predicted search patterns, ensuring both current and comprehensive results without constant real-time queries.
3Measurement precision
If real-time communication with supplier systems is performed for each search query, then the search result currency is improved, but the operational overhead for suppliers increases
Solution Approach 1:
Instead of continuous real-time communication for every search query, the system implements periodic data collection through predictive searches. Data is fetched at intervals based on predicted search patterns and cached for multiple queries. This periodic approach maintains search result currency while dramatically reducing the frequency and volume of communications with supplier systems, thereby lowering their operational overhead.
Data Source
AI summary
A computer-implemented method of performing product data searches is disclosed. The method comprises identifying a list of search terms for one of a plurality of supplier systems; determining a search schedule for the one supplier system; performing a first search of product data hosted by the one supplier system using the list of search terms according to the search schedule, comprising: simulating, upon receiving product data returned by the one supplier system, a user interaction with a graphical user interface that triggers generation of additional product data by the one supplier system, and retrieving dynamic data generated in response to the triggering; storing a result of the first search in a memory coupled with the processor; receiving a search query from a buyer system; identifying one or more matches of the search query from the result of the first search in the memory.