Storage Switch Failure Detection via Data Volume and Error Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional failure detection methods in storage systems are time-consuming and inaccurate, as they individually check each switch in the I/O path, leading to false positives and difficulties in identifying the actual failing switch due to software, link, or random errors.
Innovation Solution
A method and apparatus that determine data amounts and check errors across multiple switches within a predetermined time window, calculating a failure risk by considering the data transmission relationships between switches, thereby identifying failing switches more accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional individual switch checking method is used, then failure detection can be performed, but the detection accuracy is low and false positives occur frequently
Solution Approach 1:
The patent merges multiple switches into a single I/O path context for collective analysis. Instead of checking switches independently, the system analyzes the entire I/O path as an integrated unit, combining data from all switches in the path to determine failure sources. This approach resolves the contradiction by improving detection accuracy through holistic analysis while reducing false positives that occur when switches are evaluated in isolation.
Solution Approach 2:
The patent segments the I/O path into individual switch components for separate data collection, then recombines them for unified analysis. Each switch's data amount and check error counts are individually tracked, but the failure determination is made by analyzing the segmented data collectively. This segmentation enables precise localization of failures within the I/O path while maintaining the benefits of integrated analysis.
2Reliability
If all switches in I/O path are checked individually, then potential failing devices can be identified, but the process is time-consuming and troublesome
Solution Approach 1:
The patent performs preliminary data collection from all switches in the I/O path simultaneously, gathering data amounts and check error counts before conducting the failure analysis. This preliminary action prepares all necessary information in advance, allowing the actual failure determination to be made quickly by comparing pre-collected data. This eliminates the time-consuming sequential checking process while maintaining complete coverage of all potential failure sources.
Solution Approach 2:
The system implements feedback mechanisms where check error counts and data amounts from each switch are continuously monitored and fed back into the failure determination process. This feedback loop enables the system to quickly identify patterns indicating failures without manual intervention, reducing the time required for comprehensive failure detection while maintaining high reliability through continuous monitoring of all switches.
3Productivity
If check errors are used to determine failing switches, then failure detection can be performed, but the method has accuracy deficiency due to false error sources
Solution Approach 1:
The patent applies local quality analysis by examining the specific characteristics of data transmission at each switch location within the I/O path. Instead of treating all check errors uniformly, the system analyzes the local context of each switch, considering its position in the data path and the specific patterns of errors it detects. This localized analysis improves precision by identifying the actual source of errors rather than simply flagging all switches with errors.
Solution Approach 2:
The patent changes the parameters used for failure detection from simple presence/absence of check errors to a more nuanced analysis combining data amounts and error rates. By transforming the detection parameters to include quantitative measurements of data flow and error frequency, the system achieves higher productivity in detecting failures while improving precision through multi-parameter analysis that distinguishes true failures from false error sources.
Data Source
AI summary
The present disclosure relates to a method and an apparatus for failure detection in a storage system. One embodiment provides a method for failure detection in a storage system. The method includes determining an amount of data received by a plurality of switches in the storage system within a predetermined time window so as to obtain a plurality of data amounts; determining a count of check errors detected by the plurality of switches in the data to obtain a plurality of check error counts; and calculating a failure risk for the plurality of switches based on the plurality of data amounts and the plurality of check error counts. Another embodiment of the present invention provides a corresponding apparatus.