Recovery Broker Database Locks for Message Disaster Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-node computing architectures with replicated databases face challenges in disaster recovery due to restricted network visibility, making existing JMS solutions unusable in secure environments like banking, where message loss can be costly.
Innovation Solution
A system and method that uses recovery brokers to access messages across sites via database locks on a common replicated database, ensuring only one master broker operates concurrently, allowing for disaster recovery even without network visibility between sites.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If network visibility is restricted at the broker level for security reasons, then security is improved, but message recovery capability deteriorates
Solution Approach 1:
The patent introduces database locks as an intermediary mechanism that enables message recovery without requiring broker-level network visibility. The recovery broker uses database locks to detect and recover messages from failed brokers, bypassing the need for direct broker-to-broker communication while maintaining security restrictions.
Solution Approach 2:
The patent replaces the mechanical system of broker-level network communication with a database-based mechanism. Instead of relying on brokers to communicate directly over the network, the system uses database locks and replicated databases to enable message recovery, substituting network visibility requirements with database access capabilities.
2Productivity
If multiple brokers operate concurrently on different sites, then system availability is improved, but message consistency deteriorates
Solution Approach 1:
The patent implements a feedback mechanism through database locks that allow recovery brokers to detect the state of other brokers and adjust their behavior accordingly. The lock mechanism provides feedback about broker availability, enabling the system to maintain consistency by preventing conflicting operations on the same messages while preserving high availability through distributed operation.
3Object-affected harmful factors
If network visibility between sites is eliminated for security reasons, then security is improved, but disaster recovery capability deteriorates
Solution Approach 1:
The patent uses database locks as an intermediary that enables disaster recovery without requiring network visibility between sites. The recovery broker at one site can acquire locks on messages in the replicated database at another site, allowing message recovery while maintaining security restrictions that prevent direct network communication between distributed brokers.
Solution Approach 2:
The patent replaces the mechanical requirement of network visibility with a database-based mechanism. Disaster recovery is achieved through database lock acquisition and message transfer, eliminating the need for network-level communication between sites while preserving the ability to recover from failures.
Data Source
AI summary
A system and method for recovering messages in a multi-node system that incorporates an active-active architecture with replicated databases is disclosed. The system comprises two or more computing nodes where each computing node includes brokers for processing local messages and a recovery broker for recovering messages from the replicated database when another node in the system fails. Each recovery broker competes with brokers on another node to acquire locks on database information. The recovery broker is successful if the network of brokers on the other node are unavailable and fail to obtain or retain locks on the database information. In this situation, the recovery broker can retrieve messages from the database information and pass the messages to other brokers on the same computing node for processing and completing any ongoing transactions.