Host-Based Signature Cache for CAS Read Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Content addressable storage (CAS) systems face increased latency due to the need to process read operations across multiple storage nodes, which also leads to higher processor utilization and network congestion, as data is stored based on content-based signatures that do not correlate with logical addresses.
Innovation Solution
Implementing a content-based signature cache that maps logical addresses to storage nodes, allowing read operations to directly target the appropriate storage node, reducing the number of hops and inter-node messaging traffic, and utilizing a multi-path input-output (MPIO) driver to control IO operations efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If data is stored across storage nodes based on content-based signatures, then storage efficiency through deduplication and compression is improved, but read operation latency increases due to multiple hops between storage nodes
Solution Approach 1:
The host device pre-fetches and caches content-based signatures and their corresponding storage node locations before read operations are needed. This preliminary action allows the system to immediately route read operations to the correct storage node without performing multiple hops, thus reducing read latency while preserving storage efficiency through deduplication
2Reliability
If read operations are processed through multiple storage nodes, then content-based signature mapping is achieved, but processor utilization and network congestion increase
Solution Approach 1:
The patent extracts the content-based signature caching functionality from the storage nodes and relocates it to the host device. This extraction eliminates the need for storage nodes to process routing decisions and inter-node messaging, thereby reducing processor utilization and network congestion while maintaining reliable content-based signature mapping
Solution Approach 2:
The host device acts as an intermediary between the host system and storage nodes, maintaining a local cache of content-based signatures and their locations. This intermediary role allows the host to directly route read operations to the correct storage node without requiring multiple hops and inter-node communication, reducing network congestion and processor utilization
Data Source
AI summary
An apparatus in one embodiment comprises at least one processing device comprising a processor coupled to a memory, with the processing device being configured to maintain a content-based signature cache for a plurality of data pages. For each of a plurality of read operations to be directed to a distributed content addressable storage (CAS) system, the processing device determines if a data page targeted by the read operation has a corresponding content-based signature in the content-based signature cache. Responsive to the data page having a content-based signature in the content-based signature cache, the processing device identifies a particular storage node that stores the data page in the distributed CAS system, and directs the read operation to the identified storage node using the content-based signature to specify the data page targeted by the read operation. The processing device illustratively comprises a host device coupled to the CAS system over a network.