Unstructured Data Retrieval via Pre-computed Permission Segments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face performance penalties and user experience issues when retrieving unstructured data due to exponential growth in user permissions combinations, especially in internationalized and localized systems, leading to slow data access and poor user experience.
Innovation Solution
The solution involves dynamic content segmentation and auto-tagging of unstructured data assets, where items are tagged with content permissions upon creation, and user permissions are compared against pre-defined segments at request time, reducing computation complexity by clustering permissions and using hash functions for unique data segment identifiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If permission matching is performed at request time for each user, then user-specific data retrieval accuracy is improved, but system response time deteriorates due to exponential computation complexity
Solution Approach 1:
The patent pre-computes and stores all possible permission combinations during database setup or maintenance periods. When a user requests data, the system performs a simple lookup of the user's permission set against the pre-computed results rather than performing exponential computation at request time. This transforms an O(2^n) request-time operation into a constant-time lookup, resolving the contradiction between accurate permission matching and fast response.
2Adaptability or versatility
If the system supports more user roles and permissions, then system adaptability and user personalization are improved, but the complexity of permission matching grows exponentially
Solution Approach 1:
The patent segments the permission matching problem into two independent parts: (1) pre-computation of all possible permission combinations and their corresponding data access rules, and (2) simple lookup of the user's specific permission set against these pre-computed segments. This segmentation allows the system to support any number of user roles without increasing runtime complexity, as the exponential work is done once during setup and stored for efficient retrieval.
3Quantity of substance
If unstructured data is stored in traditional RDBMS, then data storage capability is improved, but data retrieval performance deteriorates due to irregular record sizes
Solution Approach 1:
The patent changes the data storage model from traditional relational databases with fixed schemas to a document-oriented storage system that natively handles unstructured data with variable sizes. By storing unstructured data (JSON, XML, binary) directly in the database without forcing it into fixed-size relational records, the system eliminates the performance penalty associated with processing irregular record sizes while maintaining full storage capability.
Data Source
AI summary
A system and method provide unstructured data to a client device based on permissions possessed by the device user and required by the data for access. Items of unstructured data stored in a data storage device are organized into data segments based on classifications assigned to them by their creators using a content management system. When a user later requests access to the data via a cloud-based service, such as a search service, the user privileges are converted into data segment identifiers which are then searched, and only the items of unstructured data that correspond to matching identifiers are returned. Data segment identifiers may be provided illustratively as a hash function to facilitate searching and to guarantee non-collision of data segment identifiers.