Flash Storage Stream Categorization for QOS Protection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flash storage devices, such as SSDs, face performance issues due to block operations like garbage collection, which can negatively impact container processes sharing the same storage blocks, leading to quality of service (QOS) degradation, as these operations are not differentiated by QOS requirements.
Innovation Solution
The method categorizes output data streams from computing processes using stream identifiers (IDs), allowing flash storage devices to group and treat data streams with the same IDs similarly, thereby minimizing block operations that impact QOS, without modifying the container processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If flash storage devices perform block operations like garbage collection, then storage management is maintained, but performance of container processes sharing the same blocks degrades
Solution Approach 1:
The patent segments storage blocks into different categories based on QOS requirements. High QOS blocks are separated from low QOS blocks, allowing differential treatment during garbage collection operations. This segmentation enables the storage system to prioritize cleanup operations on low QOS blocks while protecting high QOS blocks from performance-degrading operations.
Solution Approach 2:
The patent applies local quality by implementing QOS-specific characteristics for different storage blocks. Each block is tagged with QOS metadata that determines its treatment during block operations. This allows the system to apply different quality levels of service to different portions of the storage device, ensuring high QOS processes receive preferential treatment during garbage collection and other block operations.
2Ease of operation
If block operations are performed without QOS differentiation, then storage operations are simplified, but QOS requirements of different container processes cannot be met
Solution Approach 1:
The patent introduces QOS metadata as an intermediary layer between the block operation logic and the actual storage blocks. This metadata acts as a mediator that carries QOS information without requiring changes to the underlying block operation mechanisms. The filter driver and storage driver use this intermediary metadata to make QOS-aware decisions during garbage collection, maintaining simple block operations while achieving QOS differentiation.
3Productivity
If stream identifiers are used to categorize data streams, then QOS-based performance differentiation is achieved, but system complexity increases due to filter drivers and stream ID management
Solution Approach 1:
The patent implements self-service by having the filter driver automatically capture stream identifiers from container processes and associate them with the appropriate QOS categories. The system self-configures the QOS metadata based on the stream ID without requiring manual intervention or complex configuration. This automation reduces the operational complexity of managing QOS differentiation while maintaining the productivity benefits.
Data Source
AI summary
Categorizing computing process output data streams for flash storage devices is disclosed. A first computing process characteristic of a first computing process that generates a first output data stream is determined. A structure that correlates the first computing process characteristic to a first stream identifier is accessed. A first filter driver is associated with the first computing process to configure the first filter driver to receive the first output data stream. The first filter driver is associated with a flash storage device. The first stream identifier is sent to the first filter driver.