Technical realization of multi-tenant architecture and data isolation mechanism
Base Chat enables multi-tenant support through a layered design that ensures strict isolation of data across organizations:
Architecture Implementation Principles
- global service layer (GSL): Shared chatbot core engine and authentication system
- tenant context layer: Associate users and organizations with Google Authentication information
- Knowledge Base Routing Layer: Automatically maps to the corresponding Ragie Connect end node based on tenant IDs
Data Segregation Strategy
- database design: Using tenant_id as a foreign key for all tables in PostgreSQL
- session isolation: Inject tenant context information into the JWT token after user authentication is passed
- Knowledge Base Connection: Each organization needs to configure a separate instance of Ragie Connect to physically isolate knowledge data
- Query Filtering: All SQL queries are automatically appended with the WHERE tenant_id=xxx condition.
This design ensures shared utilization of system resources while ensuring privacy of data across organizations through logical segregation. Administrators can manage tenant configurations, including knowledge base connection parameters, user permission group settings, etc., through a proprietary interface.
This answer comes from the articleBase Chat: A Multi-Tenant Chatbot for Building Knowledge Bases Based on RagieThe
May not be reproduced without permission:AI productivity tools " How exactly is Base Chat's multi-tenant architecture implemented? How is data segregated across different organizations?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil