Storage Node Connection Unit for Dynamic Write Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage systems with multiple networked nodes, delays in processing by one node can significantly degrade the performance of the entire system, especially when garbage collection processes are ongoing, leading to inefficiencies in data input/output operations.
Innovation Solution
Implementing a storage system with a connection unit that directly communicates with nodes, segments data for optimal writing, and dynamically reroutes write commands to nodes with faster processing times, while also detecting and mitigating delays by using a response history to identify and redirect data to less busy nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is written to a node performing garbage collection, then data can be stored, but processing speed decreases significantly
Solution Approach 1:
The system dynamically selects write destinations based on real-time node status. The connection unit monitors response times from multiple nodes and adaptively redirects write commands to nodes with faster response times, transforming the static write destination selection into a dynamic decision-making process that optimizes processing speed while maintaining storage capacity.
Solution Approach 2:
The system changes the parameter of node selection by monitoring response time thresholds. When a node's response time exceeds a predetermined threshold (indicating garbage collection activity), the system changes the write destination parameter to select an alternative node, thereby maintaining high processing speed without sacrificing data storage capability.
2Speed
If write commands are redirected to faster nodes, then processing speed improves, but system complexity increases
Solution Approach 1:
The connection unit implements a feedback mechanism by monitoring response times from storage nodes and using this information to make intelligent routing decisions. The system continuously receives feedback about node performance and adjusts write command distribution accordingly, achieving high processing speed through adaptive control rather than complex predetermined routing logic.
Solution Approach 2:
The system employs a self-service approach where storage nodes automatically indicate their availability status through response time characteristics. Nodes performing garbage collection naturally exhibit slower response times, which the connection unit detects and uses to automatically redirect traffic, eliminating the need for explicit status signaling or complex coordination protocols.
3Productivity
If multiple nodes are monitored for response time, then data distribution efficiency improves, but detection and measurement difficulty increases
Solution Approach 1:
The system leverages the natural self-indicating behavior of storage nodes during garbage collection. Instead of requiring complex monitoring infrastructure, the connection unit simply measures response times to write commands, which automatically reveal node status. Fast responses indicate availability, while slow responses indicate garbage collection activity, making detection straightforward without additional measurement complexity.
Data Source
AI summary
A storage system includes a plurality of nodes, each of the nodes including a processor and a nonvolatile storage device, and a connection unit directly connected to at least one of the nodes and configured to issue commands directed to a target node. The processor of the target node is configured to access the nonvolatile storage device of the target node in accordance with a command received from the connection unit, notify completion of processing for the command to the connection unit when a processing result responsive to the command is returned from the nonvolatile storage device of the target node within a reference time period, and notify a busy state of the target node to the connection unit when the processing result responsive to the command is not returned from the nonvolatile storage device of the target node within the reference time period.