Disk Array Data Migration Using Accumulated Write Commands
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data migration methods in disk arrays are inefficient due to the limited maximum data size per writing command, leading to a high number of writing controls and increased system resource usage.
Innovation Solution
The method involves sending reading commands to accumulate data into sub-data sizes that match the maximum write size, allowing for simultaneous writing commands that maximize data transfer per operation, thereby reducing the number of writing controls and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is written using maximum size commands (128K bytes) to reduce the number of writing controls, then writing efficiency is improved, but the data must be accumulated in memory first which increases memory usage
Solution Approach 1:
The patent accumulates read data in memory buffer before issuing write commands, performing the accumulation action in advance. This preliminary action allows subsequent write commands to operate at maximum efficiency with full 128K byte transfers, resolving the contradiction by accepting temporary memory usage in exchange for improved writing efficiency.
Solution Approach 2:
The patent introduces a time dimension to the data transfer process by accumulating data over multiple read operations before writing. This transforms the problem from a single-dimension (immediate write) to multi-dimension (accumulate then write), allowing optimization of write efficiency while managing memory resources across time.
2Ease of operation
If data is striped into sub-data smaller than maximum write size to match disk array configuration, then data distribution is optimized, but the number of writing commands increases reducing efficiency
Solution Approach 1:
The patent changes the parameter of write command size from the default smaller striped portions to the maximum 128K byte size. By adjusting this parameter, the system achieves optimal writing efficiency while still maintaining proper data distribution across the disk array through the striping mechanism operating on the accumulated data blocks.
3Reliability
If multiple small writing commands are used to transfer data according to disk array striping, then data is properly distributed across disks, but system resource usage increases
Solution Approach 1:
The patent merges multiple small read operations into a single large write operation by accumulating data in memory buffer. This combining approach reduces the total number of write commands issued to the disk array, thereby reducing system resource usage while maintaining proper data distribution through the underlying striping mechanism.
Data Source
AI summary
While performing data transfer for reading data from a source disk or disks and writing the read data to disks of a disk array, data reading is repeated until accumulated read data can be split into several sub-data with a data size of each sub-data not smaller than a write maximum. Then, these sub-data are written into the disks of the disk array. In this way, the invention can take full advantage of each writing control to write as much data as possible, and the number of writing controls can be reduced to reduce the demands on the computer system.