Peer-to-Peer Data Transfer Between Disk Cache and Backing Store
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current computing systems face inefficiencies in data movement between disk cache and backing store due to unnecessary reads and writes involving main memory, which hinder overall system performance despite the use of emerging NVRAM technologies.
Innovation Solution
Implementing a peer-to-peer (P2P) data movement architecture between disk cache and backing store, allowing direct data transfer between SSDs or disk drives without passing through main memory, thereby eliminating wasteful memory operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional caching algorithms are used with main memory involvement, then data can be cached between storage devices, but unnecessary reads and writes to main memory occur which reduce system performance
Solution Approach 1:
The patent extracts main memory from the data movement path between disk cache and backing store. By implementing direct peer-to-peer data transfer between storage devices, the unnecessary reads and writes to main memory are eliminated, resolving the technical contradiction between maintaining caching functionality and avoiding wasteful memory operations
Solution Approach 2:
The patent introduces a storage controller as an intermediary that enables direct data movement between disk cache and backing store without involving main memory. The controller manages the caching operations and data transfer, allowing the system to maintain caching benefits while eliminating the performance bottleneck caused by main memory involvement
2Adaptability or versatility
If data is transferred through main memory between disk cache and backing store, then data caching is enabled, but main memory load increases reducing overall data transfer efficiency
Solution Approach 1:
The patent removes main memory from the data transfer path between storage devices. By enabling direct peer-to-peer data movement, the system maintains full caching capability while eliminating the excessive main memory load that was reducing data transfer efficiency
Solution Approach 2:
The patent segments the storage system into independent components (disk cache, backing store, and storage controller) that can communicate directly. This segmentation allows data to flow between storage devices without being funneled through main memory, reducing memory load while preserving caching functionality
Data Source
AI summary
An apparatus is described. The apparatus includes peer-to-peer intelligence to be integrated into a mass storage system having a cache and a backing store. The peer-to-peer intelligence is to move data between the cache and backing store without the data passing through main memory.