SSD L2P Mapping with Multiple Indirection Granularities
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional NAND solid state drives (SSDs) face high read-modify-write overhead and endurance issues due to their fixed indirection unit granularity, which increases storage costs and performance penalties, especially when handling small write operations and random I/Os.
Innovation Solution
Implementing a multiple indirection unit granularity system that allows for multiple Logical-to-Physical (L2P) tables with varying IU granularities, enabling efficient handling of small and large file operations by dynamically selecting the appropriate IU granularity based on operating environments and data types, thereby reducing storage requirements and write amplification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed indirection unit granularity of 4 KB is used in the L2P table, then the mapping structure is simple and consistent, but the DRAM storage requirement becomes excessively high (1 GB per 1 TB SSD capacity) and the read-modify-write overhead increases significantly for small write operations
Solution Approach 1:
The patent divides the L2P table into multiple segments, each with a different indirection unit granularity. Small LBA ranges use fine-grained indirection units (e.g., 512 bytes) while large LBA ranges use coarse-grained indirection units (e.g., 64 KB). This segmentation allows the system to use smaller, more efficient indirection units only where needed, reducing overall DRAM requirements while maintaining simplicity in each segment.
Solution Approach 2:
The patent applies different indirection unit granularities to different local regions (LBA ranges) of the storage device based on their specific access patterns and requirements. This local quality approach ensures that each region uses the optimal indirection unit size for its workload, improving overall system efficiency without requiring a uniform structure throughout.
2Stability of the object's composition
If a fixed indirection unit granularity of 4 KB is used, then the mapping consistency is maintained, but the read-modify-write overhead becomes excessively high (8-128× performance penalty) for small write operations due to the need to read and write entire indirection units even for small data changes
Solution Approach 1:
The patent segments the address space into multiple LBA ranges, each mapped to L2P table segments with different indirection unit granularities. This allows small write operations to affect only small portions of data within fine-grained segments, reducing the read-modify-write overhead while maintaining mapping consistency within each segment.
Solution Approach 2:
The patent introduces dynamic indirection unit granularity that can be selected based on the specific LBA range and access patterns. This dynamic approach allows the system to adapt the indirection unit size to match the operation size, improving write performance by avoiding unnecessary reading and writing of large indirection units for small operations.
3Adaptability or versatility
If conventional file systems designed for hard disk drives are used with SSDs, then compatibility with existing file systems is maintained, but write amplification increases significantly due to the random-write nature of small I/O operations on SSDs
Solution Approach 1:
The patent applies different indirection unit granularities to different local regions based on their access patterns. This allows the system to optimize for small random writes in regions that require fine-grained indirection, thereby reducing write amplification for metadata operations while maintaining file system compatibility.
Solution Approach 2:
The patent changes the indirection unit granularity parameter dynamically based on the LBA range and access patterns. This parameter change allows the system to reduce write amplification by using smaller indirection units for metadata operations and larger units for bulk data operations, improving overall energy efficiency.
4Quantity of substance
If the indirection unit granularity is increased to 64 KB in the near future, then the DRAM storage requirement is reduced, but the read-modify-write overhead and performance penalty for small write operations increases even further
Solution Approach 1:
The patent segments the address space and L2P table into multiple parts with different indirection unit granularities. This allows the system to use larger indirection units (64 KB) in segments where they are appropriate while maintaining smaller indirection units in segments that require fine-grained access, thereby reducing overall DRAM requirements without sacrificing small write performance.
Solution Approach 2:
The patent applies larger indirection unit granularity locally to specific LBA ranges that benefit from it, while maintaining smaller granularities in ranges that require fine-grained access. This local quality approach allows the system to reduce DRAM requirements in appropriate areas without incurring the performance penalty for small writes in areas that need fine-grained control.
Data Source
AI summary
One embodiment provides a memory controller. The memory controller includes logical block address (LBA) section defining logic to define a plurality of LBA sections for a memory device circuitry, each section including a range of LBAs, and each section including a unique indirection-unit (IU) granularity; wherein the IU granularity defines a physical region size of the memory device. The LBA section defining logic also to generate a plurality of logical-to-physical (L2P) tables to map a plurality of LBAs to physical locations of the memory device, each L2P table corresponding to an LBA section. The memory controller also includes LBA section notification logic to notify a file system of the plurality of LBA sections to enable the file system to issue a read and/or write command having an LBA based on an IU granularity associated with an LBA section.