Storage Management Data Verification CPU Load Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional preprocessing methods for dirty data in storage management are inefficient due to the need for separate calculations of CRC and non-encrypted hash values, requiring multiple cycles and increased CPU loads.
Innovation Solution
Simultaneously determining check information and identification information during data loading cycles, reducing the need for cyclic traversal and halving the data load, thereby improving CPU efficiency and IO performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If separate calculations of CRC and non-encrypted hash values are performed in conventional preprocessing methods, then data verification and identification can be achieved, but CPU loads increase and processing efficiency decreases
Solution Approach 1:
The patent combines CRC calculation and non-encrypted hash value calculation into a single unified processing step. Instead of performing separate calculations for each type of check information, the system computes both types of information simultaneously during data loading, thereby reducing CPU load and improving preprocessing efficiency while maintaining data verification reliability
2Reliability
If multiple cycles are used for separate check information calculations, then comprehensive data validation is achieved, but processing time increases and IO performance deteriorates
Solution Approach 1:
The patent performs preliminary calculation of both CRC and non-encrypted hash values during the initial data loading cycle from persistent storage to memory. By computing all necessary check information in advance during this first pass, the system eliminates the need for subsequent re-traversal cycles, thereby reducing total preprocessing time while ensuring comprehensive data validation
3Measurement precision
If redundant data loading cycles are performed, then all check information can be calculated, but CPU efficiency decreases and processing overhead increases
Solution Approach 1:
The patent merges the calculation of CRC check information and non-encrypted hash identification information into a single processing operation. By computing both types of information simultaneously during one data loading cycle rather than performing separate calculation passes, the system maintains measurement precision for all check information while significantly improving CPU efficiency and reducing processing overhead
Data Source
AI summary
Storage management techniques involve: acquiring target data in a target storage page in a memory; determining, based on the target data, check information and identification information associated with the target data, the check information being used to verify whether the target data is correct and the identification information being used to identify the target data; and determining, based on the identification information, storage information associated with the target data and the check information, the storage information indicating whether to store the target data and the check information to a persistent storage device. Therefore, the processing efficiency can be improved, and the input/output (I/O) performance can be improved.