Query Federator Personalizes Search Scope to Reduce CPU Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Federated search systems experience excessive CPU consumption due to the increasing number of searchable resources, leading to inefficient performance and high costs, despite efforts to parallelize search processes.
Innovation Solution
A query federator is implemented to personalize the scope of federated searches by selecting a subset of searchable resources based on user interaction data, including past search behavior, role, CPU consumption values, and response times, thereby restricting the search to only relevant resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If parallelization is used to maintain search response time independent of the number of searchable resources, then search response time is improved, but CPU consumption increases directly with the number of searchable resources
Solution Approach 1:
The patent extracts and removes irrelevant searchable resources from the search scope by analyzing user interaction data. The query federator identifies and excludes resources that users have not interacted with or found irrelevant, thereby reducing the number of resources that need to be searched while maintaining search quality. This directly reduces CPU consumption without sacrificing response time.
Solution Approach 2:
The patent applies local quality by personalizing the search scope for each user based on their interaction data. Instead of applying a uniform search scope to all users, the system tailors the search to include only those resources relevant to each specific user's needs and preferences. This reduces overall CPU consumption by avoiding searches of irrelevant resources for each user.
2Reliability
If the scope of federated search is expanded to include more searchable resources, then search completeness is improved, but CPU consumption increases
Solution Approach 1:
The patent implements feedback mechanisms by analyzing user interaction data from past search behaviors, clicks, and refinements. This feedback is used to continuously optimize and personalize the search scope, ensuring that only relevant resources are included. The system learns from user interactions to maintain search completeness while excluding irrelevant resources, thereby reducing CPU consumption.
Solution Approach 2:
The patent applies partial action by searching only a subset of available resources that are most relevant to each user, rather than exhaustively searching all possible resources. By using interaction data to identify and search only the most pertinent resources, the system achieves sufficient search completeness without the excessive CPU consumption that would result from searching all available resources.
3Measurement precision
If personalized search scope is implemented using interaction data, then search accuracy is improved, but system complexity increases
Solution Approach 1:
The patent introduces a query federator as an intermediary component that handles the complexity of personalization. This mediator sits between the user and the searchable resources, managing the analysis of interaction data and the dynamic construction of personalized search scopes. By centralizing this functionality in a dedicated intermediary component, the system achieves high search accuracy without distributing complexity throughout the entire system architecture.
Data Source
AI summary
Techniques for reducing CPU consumption in a federated search are disclosed. In some example embodiments, a computer-implemented method comprises determining an initial search scope by selecting a subset of searchable resources from a plurality of searchable resources based on interaction data of a user, with the initial search scope being defined by the selected subset of searchable resources, and the interaction data indicating online activity of the user directed towards past search results of past search queries submitted by the user. A federated search for a current search query is then performed using the initial search scope, with the federated search for the current search query being restricted to only the subset of searchable resources, and current search results for the current search query are generated based on the performing of the federated search.