Failover Destination Pair Selection Using Performance Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional failover methods in computer systems do not effectively manage the selection of I/O adapters for failover destinations, leading to potential decreases in processing performance and reliability after failure recovery.
Innovation Solution
A management system that includes a processor and storage apparatus to determine failover destination pairs by referencing performance management information, I/O performance management information, and priority management information, ensuring optimal selection of computer and I/O adapter pairs for failover operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional failover methods are used to switch from active to standby server computer, then failover can be performed, but the processing performance and reliability may decrease after failure recovery
Solution Approach 1:
The system performs preliminary actions by pre-establishing performance management information, I/O performance management information, and priority management information before failover occurs. These pre-configured data structures enable the management computer to quickly determine optimal failover destinations without real-time evaluation, thus maintaining high processing performance while ensuring reliable failover.
Solution Approach 2:
The system implements feedback mechanisms where the management computer continuously monitors performance management information and I/O performance management information to determine optimal failover destinations. This feedback loop ensures that failover decisions are based on actual performance data, preventing degradation of processing performance and reliability.
2Ease of operation
If I/O adapter is determined in advance for failover destination, then failover can be executed, but the selection of optimal I/O adapter is not made
Solution Approach 1:
The system transitions from static pre-determined I/O adapter assignment to dynamic selection based on real-time performance monitoring. The management computer dynamically determines the optimal I/O adapter for failover destinations by referencing current performance management information and priority management information, enabling both easy failover execution and optimal performance selection.
Solution Approach 2:
The system changes parameters from fixed I/O adapter determination to variable selection based on performance metrics. By using performance management information and I/O performance management information as decision parameters, the system can adaptively select the best I/O adapter for each failover scenario, improving both ease of operation and reliability.
3Extent of automation
If failover strategy table is used to select failover destination, then failover can be automated, but performance and reliability relationships between server and I/O adapter are not considered
Solution Approach 1:
The management computer performs multiple functions by simultaneously referencing performance management information, I/O performance management information, and priority management information to determine failover destinations. This multi-functional approach enables automated failover while comprehensively considering performance and reliability relationships between servers and I/O adapters.
Solution Approach 2:
The management computer acts as an intermediary that mediates between failover automation requirements and performance optimization needs. By using performance management information and I/O performance management information as intermediary data structures, the system achieves both automated failover execution and intelligent performance-based decision making.
Data Source
AI summary
An exemplary management system stores computer performance management information, I/O performance management information regarding communication by I/O adapters, and priority management information associating a priority of a failover destination pair candidate including failover destination computer and destination I/O adapter candidates with a relationship between a failover source computer and destination computer candidate and with an I/O performance relationship between failover source and destination I/O adapter candidates. The management system determines failover destination pair candidates, each including failover destination computer and I/O adapter candidates. The management system determines a priority of each failover destination pair candidates with reference to the computer and I/O performance management information, and the priority management information based on a performance relationship between the active computer and the failover destination computer candidate and on a performance relationship between the active I/O adapter and the failover destination I/O adapter candidate.