Cache Destination Selection for Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In scale-out type storage systems, loosely coupled storage apparatuses lead to inefficient data transfer, resulting in increased bandwidth consumption and performance degradation due to two-stage data transfer between storage apparatuses, which cannot be effectively resolved by existing cache size control methods.
Innovation Solution
Determine the cache destination storage apparatus based on the target I/O pattern and coupling mode, allowing I/O data to be cached in a CM area of the appropriate storage apparatus, thereby reducing unnecessary data transfers and optimizing bandwidth usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If storage apparatuses are loosely coupled to maintain system scalability and ease of operation, then the system can be easily expanded by adding nodes, but I/O data must be stored in both the source and destination cache memory areas, doubling cache requirements and increasing data transfer overhead
Solution Approach 1:
The patent merges the cache memory resources of multiple storage apparatuses into a unified cache pool. By allowing storage apparatuses to share and manage each other's cache memory areas, the system transitions from isolated cache management to collaborative cache management, thereby reducing total cache requirements while maintaining scalability.
Solution Approach 2:
Each cache memory area is designed to serve multiple functions: it acts as both local cache for its own storage apparatus and as shared cache for other storage apparatuses in the system. This multi-functionality allows the same physical cache resources to be utilized by multiple nodes, reducing overall cache demand.
2Quantity of substance
If storage apparatuses are tightly coupled to reduce cache memory usage by sharing cache segments, then cache efficiency improves, but the complexity of managing cross-apparatus cache memory increases
Solution Approach 1:
The patent introduces a cache management unit as an intermediary component that handles the complexity of cache allocation, tracking, and management across multiple storage apparatuses. This mediator abstracts the complex cross-apparatus cache management operations, allowing individual storage nodes to operate independently while benefiting from shared cache resources.
Solution Approach 2:
The system implements feedback mechanisms where cache management units continuously monitor cache usage patterns, hit rates, and data access frequencies across the distributed cache pool. Based on this feedback, the system dynamically adjusts cache allocation and data placement decisions to optimize performance while managing complexity through data-driven automation.
3Reliability
If I/O data is cached in both source and destination storage apparatuses to ensure data availability, then data reliability improves, but bandwidth consumption increases due to redundant data transfer
Solution Approach 1:
Instead of fully duplicating I/O data in both source and destination cache areas, the patent implements selective copying where only necessary data segments are replicated based on access patterns and reliability requirements. The cache management unit determines which data needs to be copied to which location, reducing redundant bandwidth consumption while maintaining data availability.
Solution Approach 2:
The patent applies different caching strategies to different data segments based on their specific characteristics and access requirements. Hot data that requires high availability is replicated across multiple cache locations, while cold data is stored in a single location. This localized quality approach optimizes the balance between reliability and bandwidth usage for different data types.
Data Source
AI summary
When one of a plurality of storage apparatuses receives an input/output (IO) request in which the address of a logical volume is designated, a cache destination storage device in which I/O data conforming to the received I/O request is to be cached is determined on the basis of a target I/O pattern and/or a coupling mode. The I/O data is cached in the CM area of the cache destination storage device. The target I/O pattern is the one among a plurality of I/O patterns to which an I/O conforming to the received I/O request belongs. Each of the plurality of I/O patterns pertains to whether an I/O destination address in the logical volume is random or sequential. The coupling mode indicates whether or not a storage device that receives an I/O request in which the same address as that designated in the received I/O request is designated has been determined.