Cross-Shard Read Request Management via Delayed Action Queue
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed systems face challenges in managing data accessibility and resource efficiency due to limited computing resources, where read requests are delayed by ongoing write processes, leading to reduced user experience and increased cognitive burden.
Innovation Solution
A data management system that employs a state model with writing, resolving, metadata update, and reading states, using a delayed action queue to manage cross-shard read requests, ensuring that read requests are processed in parallel with writes by transitioning them to a resolving state and converting them into read requests once the last delayed action is processed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If read requests are processed during ongoing write processes, then data consistency is compromised, but if read requests are delayed until writes complete, then user experience and system responsiveness deteriorate
Solution Approach 1:
The patent segments the write operation into distinct phases (writing state, resolving state, metadata update state) and uses state machines to manage transitions. Read requests are routed to appropriate data versions based on the current write state, allowing concurrent access without compromising consistency. The delayed action queue segments read request handling into controlled batches that respect write operation boundaries.
Solution Approach 2:
The patent introduces a delayed action queue as an intermediary mechanism between read requests and the data storage system. This queue buffers read requests during critical write operations and releases them in an ordered manner after writes complete, mediating the conflict between immediate read access and write completion requirements while maintaining data consistency.
2Productivity
If computing resources are allocated to handle both read and write requests simultaneously, then resource utilization improves, but system complexity and coordination overhead increase
Solution Approach 1:
The patent employs dynamic state machines that automatically transition between writing state, resolving state, and metadata update state based on operational context. This dynamic approach allows the system to adapt resource allocation and request handling strategies in real-time without manual intervention, optimizing productivity while managing complexity through automated state-driven logic.
Solution Approach 2:
The system changes operational parameters (state transitions) to manage the complexity of simultaneous read-write operations. By defining specific states and transition conditions, the patent transforms a complex coordination problem into a manageable state machine model where resource allocation and request routing are determined by current system state rather than complex real-time negotiations.
3Ease of operation
If cross-shard read requests are processed immediately, then user experience improves, but data availability and consistency across distributed nodes cannot be ensured
Solution Approach 1:
The patent performs preliminary actions by pre-processing and validating cross-shard read requests before they are fully executed. The delayed action queue pre-buffers these requests and performs preliminary checks to ensure data availability across nodes before committing to processing, thereby maintaining good user experience while ensuring reliability through advance validation.
Data Source
AI summary
A node includes storage that stores a portion of high accessibility data corresponding to a shard that is divided into revisions. The portion of the high accessibility data is managed based on an activity state model including a writing state, a resolving state, a metadata update state, and a reading state. The node further includes a repository manager that obtains a cross-shard read request for the high accessibility data; makes a determination that the portion of the high accessibility data is not in the reading state; in response to the determination, adds the cross-shard read request to a delayed action queue at the revision of a last delayed action in the delayed action queue; and after processing the last delayed action in the delayed action queue and while in the resolving state: converts the cross-shard read request to a read request; and processes the read request.