Distributed State Machines for Concurrent Query Throughput
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional approaches for maintaining the integrity of state machines in distributed systems undermine the effectiveness of the architecture by blocking other processes, reducing throughput, and complicating troubleshooting and data synchronization.
Innovation Solution
Implement a primary state machine with a serializable lock and a secondary state machine with a read-committed isolation level, using optimistic locking and messaging-based replay to maintain data integrity and allow concurrent operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a strong lock is maintained for each operation involving the state machine to prevent data corruption, then data integrity is ensured, but system throughput is greatly reduced and other processes are blocked
Solution Approach 1:
The patent divides the state machine system into two separate state machines: a primary state machine that maintains strong locking for data integrity, and a secondary state machine that allows concurrent read operations. This segmentation enables different access patterns for different operational needs, resolving the contradiction between data integrity and system throughput.
Solution Approach 2:
The secondary state machine acts as an intermediary that handles read operations and troubleshooting queries. By directing read-only operations to the secondary state machine, the system allows concurrent access without blocking the primary state machine, thus maintaining throughput while preserving data integrity through the primary machine's locking mechanism.
2Productivity
If a batch data store is used to remove the need for locks on state machines, then query performance improves, but the system becomes more complex to maintain and the data store is disconnected temporally and technologically
Solution Approach 1:
The patent creates a secondary state machine that is a copy of the primary state machine's structure and schema. This copy allows read operations to proceed without locking the primary state machine, improving query performance while maintaining simplicity. The secondary machine is updated through messaging when the primary machine changes state, ensuring temporal and technological consistency without requiring complex batch processing infrastructure.
Data Source
AI summary
Embodiments are disclosed for improving throughput in a distributed transaction processing system. In the context of a method, an example embodiment includes receiving information regarding a change in state of a particular transaction recorded by the distributed transaction processing system. This example embodiment of the method further includes updating, by state machine query circuitry of a networked device within the distributed transaction processing system and based on the received information, a state of a primary state machine associated with the particular transaction. Finally, the example embodiment of the method further includes updating, by state machine query circuitry of a networked device within the distributed transaction processing system and based on the received information, a state of a primary state machine associated with the particular transaction. Corresponding apparatuses and computer program products are also provided.