Storage System Cache Control via Local Memory Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage systems face performance degradation due to increased overhead in shared memory access and processor operations when updating control information, especially with high-speed storage media like solid state drives (SSDs), which outweighs the benefits of reduced read times from caching.
Innovation Solution
A storage system architecture that includes a processor managing logical volumes, a cache memory, and a shared memory for caching control information, where the processor determines whether to update cache control information in the shared memory based on conditions such as SSD usage and cache hit rates to minimize overhead and maximize performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data caching is implemented to improve read performance, then read time is reduced, but processing time overhead increases
Solution Approach 1:
The patent extracts the cache control information from shared memory and stores it in local memory, separating the frequently accessed control data from the shared memory space. This reduces the overhead of accessing cache control information while maintaining the performance benefits of data caching.
Solution Approach 2:
The patent implements control caching at the local processor level, where each processor maintains its own cache control information in local memory. This localizes the cache management functionality, reducing contention and overhead in shared memory while improving read performance through faster local access.
2Reliability
If control information is synchronously updated in shared memory, then cache hit rate is maintained, but overhead in shared memory access increases
Solution Approach 1:
The patent extracts cache control information from shared memory and stores it locally, eliminating the need for synchronous updates in shared memory while maintaining cache effectiveness through local tracking of cache status.
Solution Approach 2:
The patent introduces local memory as an intermediary between the processor and shared memory for cache control information. This intermediary layer allows the processor to access cache control information without directly accessing shared memory, reducing overhead while maintaining data consistency.
3Productivity
If data is cached from non-volatile storage to cache memory, then read performance is improved, but overhead in shared memory and processor increases
Solution Approach 1:
The patent extracts cache control information from shared memory and stores it in local memory, reducing the processing overhead associated with shared memory access while maintaining the performance benefits of data caching from non-volatile storage.
Solution Approach 2:
The patent implements local cache control, where each processor manages its own cache control information in local memory. This localizes the overhead to individual processors rather than creating shared memory bottlenecks, improving overall system productivity while reducing processing overhead.
Data Source
AI summary
In an exemplary storage system, a processor assigns an unused process to a read request designating an area of a logical volume. The processor determines whether the data designated by the read request is in a cache memory, based on a first identifier for identifying the area designated by the read request. When the designated data is not in the cache memory and a part of physical volumes providing the logical volume is a first kind of physical volume, the processor stores the first identifier associated with an identifier for identifying an area allocated in the cache memory. When the designated data is not in the cache memory and a part of the physical volumes is a second kind of physical volume, the processor stores a second identifier for identifying the process assigned to the read request associated with an identifier for identifying an area allocated in the cache memory.