Multi-tenant Inverted Index Hosting for Text Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face challenges in efficiently allocating computing resources for text searches across multiple user accounts while maintaining data confidentiality and preventing resource waste, particularly in scenarios with varying data sizes and search workloads.
Innovation Solution
Implementing multi-tenant hosting of inverted indexes with a request router, common data store, and caching strategies to dynamically allocate resources and ensure encryption across index nodes, allowing for efficient text search and index updates while preserving user data confidentiality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If separate computing resources are allocated for each user account's inverted index, then data confidentiality is maintained, but resource waste occurs when some accounts have small data sizes
Solution Approach 1:
The patent combines multiple inverted indexes from different user accounts into a single shared index structure stored in a common data store. This merging approach allows multiple tenants to share the same computing resources and storage infrastructure, eliminating resource waste while maintaining data confidentiality through logical separation and encryption mechanisms.
Solution Approach 2:
The patent introduces a request router as an intermediary component that receives search requests, determines the appropriate user account and inverted index, and routes the request to the correct data. This mediator enables efficient resource sharing while preserving data isolation and confidentiality across multiple user accounts.
2Productivity
If a shared inverted index is used across multiple user accounts, then resource allocation efficiency improves, but data confidentiality may be compromised
Solution Approach 1:
The patent applies local quality by implementing user-specific encryption keys and security policies for different portions or entries within the shared inverted index. Each user account's data within the shared index structure is protected with localized security measures, ensuring that while resources are shared, data confidentiality is maintained at the user-specific level.
Solution Approach 2:
The patent segments the shared inverted index into user-specific portions or logical sections, each accessible only to its corresponding user account. This segmentation allows efficient resource sharing through a unified index structure while maintaining data confidentiality by preventing unauthorized access to other users' data segments.
3Loss of energy
If computing resources are dynamically allocated based on varying search workloads, then resource waste is reduced, but system complexity increases
Solution Approach 1:
The patent implements dynamic resource allocation by allowing the system to adjust computing resource allocation based on varying search workloads and data access patterns. The shared inverted index structure enables flexible resource utilization where resources are allocated to handle peak loads efficiently while reducing allocation during low-demand periods, optimizing resource waste reduction.
Data Source
AI summary
Multi-tenant hosting of inverted indexes for text searches is implemented. Text search requests are routed to different index nodes that cache inverted indexes for different user accounts. Updates to inverted indexes are routed to index nodes that have acquired a lock on an inverted index. The index nodes have access to a common data store that persistently stores the inverted indexes.