Read Temperature Identification via Read Disturb
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional read temperature identification systems in information handling systems are costly, complex, and inaccurate, requiring dedicated host memory and power-fail safety measures, and often introduce performance issues due to host processor involvement, with limited granularity and resilience to changing workloads.
Innovation Solution
A storage device with a local read temperature identification engine that determines read disturb information to generate a local logical storage element read temperature map, identifying rows with higher read temperatures and storing it, allowing for efficient data placement and load balancing without relying on host processors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional read temperature identification systems use host processors and host memory to track read temperatures, then read temperature identification can be performed, but device complexity and costs increase due to dedicated host memory requirements and power-fail safety measures
Solution Approach 1:
The storage device performs read temperature identification autonomously using its own processing system and memory, eliminating the need for host processor intervention. The storage device self-monitors read disturb information and generates read temperature maps independently, thereby reducing host memory requirements and system complexity while maintaining measurement precision.
Solution Approach 2:
The read temperature identification function is extracted from the host processor and host memory system and relocated to the storage device itself. By taking out this functionality from the host system, the patent eliminates the need for dedicated host memory and reduces overall system complexity while preserving the ability to accurately identify read temperatures.
2Reliability
If conventional systems use host processors to generate and synchronize read temperature maps across multiple storage devices, then read temperature tracking is achieved, but system complexity increases due to synchronization requirements and host processor involvement
Solution Approach 1:
Each storage device independently generates and maintains its own read temperature map using its local processing system, eliminating the need for host processor coordination. Storage devices self-manage their temperature maps without requiring synchronization protocols, thereby reducing host processor involvement and system complexity while ensuring reliable local temperature tracking.
3Reliability
If conventional read temperature identification systems implement power-fail safety measures with battery backup or Storage Class Memory, then read temperature data persistence is ensured, but costs increase
Solution Approach 1:
The storage device uses its own non-volatile memory and processing system to persistently store read temperature information without requiring external power-fail safety infrastructure. The device self-preserves temperature data using its inherent storage capabilities, eliminating the need for battery backups or additional SCM devices while maintaining data persistence reliability.
4Measurement precision
If host processors perform read temperature identification operations, then read temperature data can be collected, but performance impact increases due to host processor bus utilization and cache thrashing
Solution Approach 1:
The storage device performs read temperature identification operations locally using its own processing system and memory resources, completely eliminating the need for host processor involvement. By handling temperature monitoring autonomously, the storage device avoids consuming host processor bus resources and preventing cache thrashing, thereby maintaining high storage system performance while accurately collecting read temperature data.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This solution reduces costs and complexity by leveraging the read disturb effect to accurately identify read temperatures at the storage device level, enabling efficient data placement and improved performance across varying workloads without host processor intervention.
Implementation Method 1
determine read disturb information; identify, using the read disturb information, a subset of rows in at least one block in the storage device that have a higher read temperature
Data Source
AI summary
A read-disturb-based read temperature identification system includes storage device(s) that each determine read disturb information for each block in that storage device, use that read disturb information to identify a subset of rows in at least one block in that storage device that have a higher read temperature than the other rows in the at least one block in that storage device and, based on that identification, generate and store a local logical storage element read temperature map that identifies a subset of logical storage elements associated with that storage device that have a higher read temperature than the other logical storage elements associated with that storage device. A global read temperature identification subsystem coupled to the storage device(s) may then retrieve at least a portion of the local logical storage element read temperature map(s) and use them to generate a global logical storage element read temperature map.