HPB Cache Hit Table for Multi-Region L2P Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing host performance booster (HPB) feature in storage devices leads to frequent activation and inactivation of HPB sub-regions when hot data is spread across multiple HPB sub-regions, resulting in inefficient data retrieval.
Innovation Solution
The storage device supports a multiple HPB sub-region mode and maintains a hit table to identify frequently accessed logical block address pages across HPB sub-regions. This information is transmitted to the host to be stored in the HPB cache, allowing for optimized read commands and minimizing HPB entry inactivation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the host stores L2P entries for multiple HPB sub-regions in the HPB cache, then data retrieval efficiency is improved, but the host must frequently activate and inactivate sub-regions causing HPB entry inactivation
Solution Approach 1:
The storage device proactively identifies hot data across multiple HPB sub-regions and preloads their L2P entries into the host's HPB cache before the host needs to access them. This preliminary action allows the host to retrieve data without frequently activating and inactivating sub-regions, as the necessary mapping information is already cached.
Solution Approach 2:
The storage device monitors read requests and uses this feedback to identify hot data patterns. By tracking which L2P entries are frequently accessed, the storage device can dynamically determine which sub-regions contain hot data and prioritize caching those entries in the host's HPB cache, thereby reducing unnecessary activation and inactivation cycles.
2Measurement precision
If the storage device tracks hit counts across multiple HPB sub-regions, then frequently accessed data can be identified, but the complexity of managing multiple sub-regions increases
Solution Approach 1:
The storage device merges the hit count tracking mechanism across all HPB sub-regions into a unified approach. Instead of managing separate tracking systems for each sub-region, the device aggregates hit count information globally, allowing it to identify hot data regardless of which sub-region it resides in. This simplifies the overall management complexity while maintaining precise hot data identification.
Solution Approach 2:
The hit count tracking mechanism serves multiple functions: it monitors data access patterns, identifies hot data across sub-regions, and provides feedback for optimizing cache management. This multi-functional approach reduces the need for separate specialized mechanisms, thereby reducing overall system complexity while achieving precise hot data identification.
3Quantity of substance
If the host uses HPB read buffer commands to retrieve L2P entries, then data can be cached in the HPB cache, but the host must send multiple commands to retrieve entries from multiple sub-regions
Solution Approach 1:
The storage device proactively identifies hot data across multiple HPB sub-regions and preloads their L2P entries into the host's HPB cache in advance. This preliminary action allows the host to retrieve multiple L2P entries with fewer commands, as the storage device has already prepared and transmitted the necessary mapping information before the host needs to access the data.
Solution Approach 2:
The storage device uses feedback from read requests to identify which L2P entries are most frequently accessed. Based on this feedback, the storage device prioritizes retrieving and caching those specific entries, allowing the host to obtain the most useful L2P mappings with fewer commands by focusing on high-value entries rather than uniformly distributing commands across all sub-regions.
Data Source
AI summary
A storage device minimizes HPB entry inactivation resulting from data associated with hot reads being retrieved from multiple HPB sub-regions covering a logical-to-physical table. The storage device may support the HPB feature and a multiple HPB sub-region mode. The storage device includes a controller that tracks a hit count associated with a logical block address in a read command. The controller determines that the hit count has reached a hit threshold and updates a hit table to identify logical block address pages associated with hit counts that have reached the hit threshold across HPB sub-regions covering a logical-to-physical table. The controller transmits the hit table to a host device to be stored in an HPB cache on the host device and to be used by the host device for read commands sent from the host device to the storage device.