Multi-Tenant Database Segmentation for Scalable Cloud Infrastructure
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional multi-client server systems face challenges in scalability, database management, content deployment, and data recovery, leading to increased downtime and reliability issues, while single-client systems require significant client effort and expertise for maintenance.
Innovation Solution
A system where a provider stores shared and tenant-specific data structures, allowing tenants to access shared metadata for efficient data retrieval and management, enabling separate storage and management of each client's applications and data without impacting others.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a multi-client system shares a common database for all clients, then resource utilization is improved, but database maintenance becomes problematic and upgrade time increases
Solution Approach 1:
The patent segments the common database into separate physical databases for each client (tenant). Each client's data is isolated in its own database instance, allowing independent maintenance and upgrades without affecting other clients. This segmentation resolves the contradiction by enabling parallel database operations - while maintaining resource sharing at the application level, the physical separation allows each database to be upgraded independently, significantly reducing total upgrade time compared to a monolithic common database approach.
2Productivity
If a multi-client system uses a common database, then resource sharing is improved, but data recovery for a specific client becomes difficult
Solution Approach 1:
The patent implements segmentation by creating separate physical databases for each client while maintaining shared access to application services. This allows standard database tools to be used for recovering specific client data from their dedicated databases, dramatically improving data recovery reliability compared to extracting data from a mixed common database. The resource sharing is maintained at the application server level, while data isolation at the database level enables reliable client-specific recovery operations.
Solution Approach 2:
The patent introduces an intermediary layer (the database management system and application servers) that mediates between the shared resource pool and individual client access. This intermediary manages the separation between physical databases and logical resource sharing, allowing clients to access shared applications while their data remains isolated in dedicated databases, thus enabling both resource sharing and reliable data recovery.
3Ease of operation
If a single-client system is used for each client, then data management independence is improved, but client effort and expertise requirements increase
Solution Approach 1:
The patent merges the benefits of single-client database isolation with multi-client resource sharing by implementing a multi-tenant architecture where multiple clients share application services but have dedicated database instances. This combination allows each client to have independent data management capabilities (like single-client systems) while the provider manages all databases centrally (reducing client effort). The application servers provide unified management interfaces, so clients gain data independence without bearing the full complexity of managing their own complete systems.
Solution Approach 2:
The patent creates a universal provider-managed platform that handles multiple client databases through a single set of application servers and management tools. The provider's system performs multiple functions: managing diverse client databases, providing unified application access, handling backups and recovery for all clients, and enabling independent data management for each client. This multi-functionality reduces the burden on individual clients while maintaining data independence.
Data Source
AI summary
The present invention discloses methods and systems for hosting tenants in a computer-based environment in which a provider stores a shared data structure. Each of the tenants may store shared-metadata referencing the shared data structure, while a first tenant may store a tenant-specific data structure specific to the first tenant for access by the first tenant. Based on the shared-metadata and in response to a data request from the first tenant, the system may the query the provider or the first tenant for requested data and provide the requested data based on the querying.