Checksum-Based State Change Detection for Storage Path Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage systems face inefficiencies in handling input-output operations due to limited resources and varying workloads, particularly during tasks like volume-migration and array non-disruptive upgrades, where storage arrays fail to provide change notifications at a fine granularity, leading to processing delays and inefficiencies.
Innovation Solution
A host device is configured with checksum computation and comparison logic to analyze storage device access data, computing and comparing checksums for ANA group descriptors to determine state changes, thereby optimizing path selection and reducing unnecessary data mining.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If storage arrays provide detailed change notifications at fine granularity, then measurement precision of state changes is improved, but device complexity increases
Solution Approach 1:
The patent introduces checksums as an intermediary mechanism between the storage array and host device. The storage array computes checksums for changed ANA groups and sends only these checksum values to the host, rather than providing detailed change notifications. This intermediary approach enables precise state change detection while keeping the storage array complexity manageable.
Solution Approach 2:
The patent extracts only the essential change information (checksums of modified ANA groups) from the complete state change data. Instead of providing full detailed notifications, the system extracts and transmits only the checksum values that indicate which ANA groups have changed, reducing the complexity burden on the storage array while maintaining detection precision.
2Measurement precision
If host device performs full data mining to determine state changes, then measurement precision is improved, but loss of time increases
Solution Approach 1:
The host device extracts and compares only the checksum values from the received data outputs, rather than performing full data mining of all ANA group descriptors. This extraction approach maintains detection accuracy by identifying checksum mismatches while significantly reducing processing time.
Solution Approach 2:
The patent changes the parameter being analyzed from complete ANA group descriptor data to simplified checksum values. By transforming the comparison task from analyzing full data structures to comparing compact checksum parameters, the system achieves the same detection accuracy with much lower processing time.
3Adaptability or versatility
If storage system handles multiple host devices with varying workloads, then adaptability is improved, but device complexity increases
Solution Approach 1:
The checksum-based state change notification mechanism serves as a universal solution that works across multiple host devices with different workloads and access patterns. The same simple checksum comparison approach handles diverse scenarios including volume migration, array upgrades, and normal operations, providing adaptability without increasing complexity.
Data Source
AI summary
An apparatus comprises a processing device configured to control delivery of input-output operations from a host device to a storage system over selected ones of a plurality of paths through a network. The processing device is further configured to receive first and second data outputs corresponding to a plurality of groups of the storage system, to compute first and second pluralities of checksums for respective ones of the groups based on the first and second data outputs, and to determine for the respective ones of the groups whether given ones of the second plurality of checksums differ from given ones of the first plurality of checksums. The control of delivery of the input-output operations is based at least in part on the determination. The plurality of groups each correspond to a plurality of storage devices and the second data output is received after the first data output.