Operation Aware Striping in Parallel Flash Drives
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage devices, such as flash memory, experience delays in read, write, and erase operations due to wear leveling, which can reduce the speed and efficiency of data distribution across segments, leading to a shorter service life.
Innovation Solution
A parallel flash drive architecture that uses a processor to send signals indicating occupied flash controllers, allowing data to be striped across multiple flash drives, thereby skipping occupied controllers and maintaining even data distribution for improved throughput and service life.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Duration of action of stationary object
If wear leveling operations are performed on flash memory segments, then the service life of the storage media is prolonged, but read, write, and erase operations experience delays reducing operational speed
Solution Approach 1:
The system divides the flash storage system into multiple independent flash drives, each with its own controller. This segmentation allows wear leveling operations to be performed on individual drives without blocking operations on other drives, thereby maintaining overall system speed while extending service life through distributed wear leveling.
Solution Approach 2:
A processor acts as an intermediary between the host system and multiple flash controllers. The processor receives data, determines which controllers are occupied, and routes data to available controllers, thereby bypassing the delays caused by occupied controllers performing wear leveling operations.
2Productivity
If data is written to multiple flash drives in parallel, then operational throughput is increased, but occupied controllers cause delays that reduce the effectiveness of parallel operations
Solution Approach 1:
The system dynamically adjusts data routing based on the real-time status of flash controllers. The processor continuously monitors which controllers are occupied and dynamically redirects data writes to available controllers, ensuring that parallel operations maintain high throughput without being blocked by occupied controllers.
Solution Approach 2:
When a flash controller is occupied with wear leveling or other operations, the processor skips that controller and routes data to alternative available controllers. This skipping mechanism allows the system to rush through the bottleneck caused by occupied controllers and maintain overall productivity.
3Device complexity
If a simple striping approach is used across all flash drives, then data distribution is simplified, but occupied controllers still cause delays and reduce service life
Solution Approach 1:
The system implements feedback mechanisms where flash controllers signal their occupancy status to the processor. The processor uses this feedback information to make intelligent routing decisions, directing data away from occupied controllers to available ones. This feedback-based approach maintains relatively simple data distribution while ensuring even wear across all drives, thereby extending service life.
Data Source
AI summary
Embodiments of the present invention provide a method comprising performing an operation on a first flash drive of a plurality of flash drives configured in a parallel flash drive architecture, wherein the operation occupies a flash controller corresponding to the first flash drive, sending a signal to a processor coupled with the parallel flash drive architecture to indicate that the flash controller is occupied, and writing data to two or more of the plurality of flash drives, other than the first flash drive, by striping the data amongst the two or more of the plurality of flash drives in response to the signal to the processor. Other embodiments may be described and/or claimed.