Dynamic Metro Storage Configuration via Host IO Response Time Hints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-site storage systems, existing methods for identifying a host computer's proximity to storage appliances in a metro cluster are static and unchanging, leading to inefficiencies in selecting optimal IO paths due to network latency and physical location changes, making it difficult to adapt to changes in storage area network congestion or nondisruptive upgrades.
Innovation Solution
The system dynamically configures IO paths by receiving IO response time hints from the host computer, modifying initial local or remote identifications, and changing IO path states based on these hints, providing notifications to the host computer to select more appropriate paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual configuration of local and remote identifications is used, then initial IO path states can be set, but the identifications remain static and unchanging until modified by administrator
Solution Approach 1:
The patent transforms the static local/remote identification system into a dynamic one by introducing IO response time measurements. The system continuously monitors response times and automatically updates host identification status based on current performance data, allowing the configuration to adapt automatically to physical location changes without requiring administrator intervention.
Solution Approach 2:
The system implements feedback by measuring IO response times continuously and using this information to dynamically adjust host identification. The response time measurements provide feedback about actual performance conditions, which triggers automatic updates to the local/remote status and IO path states, creating a closed-loop system that self-corrects based on observed conditions.
2Stability of the object's composition
If static local and remote identifications are maintained, then configuration stability is achieved, but IO response times increase due to SAN congestion and network latency
Solution Approach 1:
The system maintains stability through a structured approach while introducing dynamics in the identification logic. It uses threshold-based triggers and controlled update mechanisms to change identifications only when performance degradation is detected, balancing stability with the need to respond to changing conditions.
Solution Approach 2:
The system performs self-service by automatically detecting performance degradation through response time measurements and autonomously updating host identification and IO path states without requiring administrator intervention. This allows the system to self-optimize in response to SAN congestion and network latency conditions.
3Measurement precision
If administrator manually updates identifications after moves, then accuracy is maintained, but system downtime and operational disruption occur
Solution Approach 1:
The system eliminates the need for administrator intervention by implementing self-service automatic detection and update mechanisms. The system continuously monitors IO response times and automatically updates host identification accuracy based on measured performance data, maintaining precision without requiring manual updates that would cause downtime.
Solution Approach 2:
The system uses continuous feedback from IO response time measurements to maintain identification accuracy. Rather than relying on periodic manual updates, the feedback loop continuously validates and updates the host's local/remote status based on actual performance conditions, ensuring accuracy while maintaining system availability.
4Adaptability or versatility
If dynamic updates based on IO response times are implemented, then adaptability to network conditions improves, but system complexity increases
Solution Approach 1:
The patent manages complexity by implementing a focused feedback mechanism that monitors only IO response times as the key performance indicator. This single metric drives the dynamic updates, providing adaptability to network conditions while avoiding the complexity of monitoring multiple parameters or implementing complex decision logic.
Solution Approach 2:
The system manages complexity by changing a single critical parameter - the host's local/remote identification status - based on IO response time measurements. Rather than dynamically adjusting multiple configuration parameters simultaneously, the system focuses on updating the identification status, which then cascades to update IO path states, simplifying the overall management complexity.
Data Source
AI summary
Techniques for dynamically configuring a multi-site storage system such as a metro cluster using input/output (IO) response time (RT) hints from a host computer. The techniques include receiving IO RT hints at each storage appliance of the multi-site storage system from the host computer, which is initially identified as “local” or “remote” relative to a physical location of the storage appliance. The techniques further include modifying, by the storage appliance, an initial local or remote identification of the host computer relative to the physical location of the storage appliance based on the received IO RT hints, dynamically changing, by the storage appliance, states of IO paths between the host computer and nodes of the storage appliance based on the modified local or remote identification of the host computer, and providing, by the storage appliance, notification of the changed states of the IO paths to the host computer.