Media Cache Cleaning via Interruptible Merge Operations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data storage devices using shingled magnetic recording (SMR), media cache cleaning operations can be time-consuming and may cause host commands to timeout due to the need to read and merge large numbers of unique user data blocks, leading to performance issues and potential data loss if not completed quickly.
Innovation Solution
Implement a method to efficiently clean the media cache by initiating band-by-band cleaning, interrupting merges based on host activity and command priority, and writing partially merged data to a scratchpad and main storage to free up cache space, allowing for concurrent processing and minimizing timeouts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If media cache cleaning operations are performed by reading and merging large numbers of unique user data blocks, then cache cleaning completeness is improved, but host command timeout risk increases and processing time increases
Solution Approach 1:
The patent segments the media cache into multiple bands, each containing a fixed number of tracks. Instead of cleaning the entire cache at once, the system performs cleaning operations on individual bands separately. This segmentation allows the cleaning process to be divided into smaller, manageable units that can be processed independently, reducing the overall time required and minimizing the risk of host command timeouts while still achieving complete cache cleaning through systematic processing of all bands.
2Reliability
If media cache cleaning operations are performed by reading and merging large numbers of unique user data blocks, then cache cleaning completeness is improved, but processing time increases
Solution Approach 1:
The patent implements preliminary action by checking whether data in the media cache needs to be merged with data on the data storage medium before actually performing the merge operation. The system determines which bands require cleaning and prioritizes them accordingly. This preliminary assessment allows the system to avoid unnecessary merge operations, focusing computational resources only on bands that actually need cleaning, thereby reducing overall processing time while maintaining complete cache cleaning effectiveness.
3Ease of operation
If merge operations are interrupted to allow other commands to be executed, then host command responsiveness is improved, but merge operation completion time increases
Solution Approach 1:
The patent applies dynamics by making the merge operation interruptible and resumeable. When a merge operation is in progress, the system can dynamically interrupt it to allow other host commands to be executed. The merge state is preserved, and the operation can be resumed later without losing progress. This dynamic approach allows the system to balance cache cleaning needs with host command responsiveness, improving overall system flexibility and ease of operation while managing merge operation completion times through prioritization and resumption capabilities.
Data Source
AI summary
This disclosure is related to media cache cleaning based on workload. In some examples, a read-modify-write (RMW) operation may merge new data with existing data and the merge may be interrupted without losing the amount of work already processed. This can be particularly useful for shingled magnetic recording (SMR) systems that utilize a large cache, such as a media cache, that can accumulate many entries that might need to be merged with existing data. An RMW operation can be interrupted such that a host command does not timeout.