Host Data Interleaving for Storage Quality of Service
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Media applications are sensitive to burst errors and continuous access latency in storage systems, which can degrade quality of service, especially in three-dimensional flash memory systems where weak wordlines cause latency and errors, and existing error correction algorithms may not suffice to handle these issues without increasing write amplification.
Innovation Solution
Data interleaving is implemented by determining a skip length based on factors like frame size, group of pictures, archive status, or failure history to spread out data storage, reducing the impact of burst errors and latency by storing consecutive data units farther apart, thereby minimizing quality of service degradation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is stored sequentially in memory, then write performance is improved, but burst errors and latency impact quality of service
Solution Approach 1:
The patent divides the data stream into smaller units (e.g., 4KB sectors) and interleaves them with skip lengths (e.g., storing every 8th sector), effectively segmenting continuous data into spaced-out portions. This segmentation transforms burst errors affecting contiguous sectors into isolated errors, improving quality of service while maintaining write performance through the host's interleaving mechanism.
2Reliability
If error correction algorithms are enhanced to handle burst errors, then quality of service is improved, but write amplification increases
Solution Approach 1:
The patent applies preliminary action by performing data interleaving at the host before writing to storage. The host calculates skip lengths based on quality of service requirements and pre-spaces out data units, so that when data is written, burst errors are already distributed rather than concentrated. This eliminates the need for intensive post-write error correction, reducing write amplification while maintaining quality of service.
3Reliability
If redundant data storage is increased to protect against burst errors, then quality of service is improved, but storage capacity is reduced
Solution Approach 1:
Instead of storing redundant copies of data, the patent segments the original data stream into spaced-out units using interleaving. The skip length determines how far apart consecutive data units are stored. This segmentation approach provides error protection through distribution rather than replication, maintaining full storage capacity while improving quality of service against burst errors.
Data Source
AI summary
A host and method for interleaving data in a storage system for enhanced quality of service are provided. In one embodiment, a host is provided comprising an interface configured to communicate with a storage system comprising a memory. The processor is configured to determine a skip length for interleaving data to be stored in the storage system; interleave data according to the determined skip length; and send the interleaved data to the storage system for storage. Other embodiments are provided.