Host Storage Cache Metadata Collaboration for I/O Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage systems, slow metadata access due to limited cache space leads to performance bottlenecks, as the cache cannot hold all metadata for user data from multiple hosts, virtual machines, and containers, resulting in I/O timeouts and reduced system performance.
Innovation Solution
A method that collaborates between host and storage caches to manage metadata, using a hash algorithm to accelerate access and swap metadata between caches, preloading high QoS metadata, and using machine learning to prioritize metadata based on usage, ensuring only one copy exists in either cache and dynamically managing cache space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If metadata is stored in host cache to enable fast access, then metadata access speed is improved, but cache space is limited and cannot hold all metadata from multiple hosts, virtual machines, and containers
Solution Approach 1:
The patent merges the storage cache and host cache into a collaborative metadata management system. The storage cache holds metadata for multiple hosts, virtual machines, and containers, while the host cache provides fast access for its specific host. This combination allows the system to maintain fast metadata access through the host cache while expanding total cache capacity through the storage cache, resolving the contradiction between access speed and cache capacity.
Solution Approach 2:
The storage cache acts as an intermediary between the backend storage and the host cache. It preloads metadata into the storage cache before the host needs it, and the host cache serves as a fast-access buffer for the specific host. This intermediary storage cache resolves the contradiction by providing a middle layer that expands capacity while maintaining fast access through the host cache.
2Speed
If storage cache size is increased to hold all metadata, then metadata access speed is improved, but system cost increases
Solution Approach 1:
Instead of preloading all metadata into the storage cache, the system uses partial action by preloading only the metadata that is likely to be needed based on usage patterns and priorities. The host cache provides additional fast access capacity for its specific host without requiring the storage cache to hold all metadata. This partial approach improves metadata access speed while avoiding the high cost of expanding storage cache to accommodate every possible metadata item.
3Reliability
If metadata is cached in both host cache and storage cache, then availability is improved, but cache space utilization becomes inefficient
Solution Approach 1:
The patent applies local quality by making the host cache host-specific rather than shared. Each host has its own cache that is optimized for its specific workload and access patterns. The storage cache, in contrast, is shared among multiple hosts, virtual machines, and containers. This local quality approach allows each cache to be utilized efficiently for its specific purpose, improving overall cache space utilization while maintaining high metadata availability through the collaborative system.
Data Source
AI summary
In an approach to optimizing metadata management to boost overall system performance, a cache for a storage system is initialized. Responsive to receiving a cache hit from a host cache during a host I/O operation, a first metadata of a plurality of metadata is transferred to a storage cache, where the first metadata is associated with a user data from the host I/O operation, and further wherein the first metadata is deleted from the host cache. Responsive to determining that the storage cache is full, a second metadata of the plurality of metadata is destaged from the storage cache, where the second metadata is destaged by moving the second metadata to the host cache, and further wherein the second metadata is deleted from the storage cache.