Controlled Data Page Ordering for Storage Migration Deduplication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data migration approaches negatively impact deduplication ratios in target storage systems with cache-based deduplication, especially when the target system has a larger native page size than the source system, leading to reduced deduplication performance.
Innovation Solution
Implementing a controlled ordering of data page transfers during migration, where data pages with common hash values are grouped together, facilitated by computing hash values based on the target storage system's native page size, allowing for enhanced cache-based deduplication in the target system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional data migration approaches are used, then data can be transferred from source to target storage system, but the deduplication ratio in the target storage system is adversely impacted
Solution Approach 1:
The source storage system performs preliminary actions by computing hash values for data pages using the target storage system's native page size before migration. This allows the source system to pre-group duplicate pages together, ensuring that when data is transferred to the target system, duplicate pages arrive in sequences that maximize cache-based deduplication effectiveness, thereby improving the deduplication ratio without compromising migration efficiency
Solution Approach 2:
The invention changes the parameter of hash value computation by using the target storage system's native page size instead of the source system's page size. This parameter change ensures that hash values are computed in a manner that aligns with the target system's deduplication cache structure, allowing duplicate pages to be properly identified and grouped even when page sizes differ between source and target systems
2Volume of stationary object
If the target storage system uses a larger native page size than the source system, then storage capacity is improved, but the deduplication performance is significantly worse
Solution Approach 1:
Before migration, the source storage system performs preliminary grouping of data pages by computing hash values based on the target system's larger page size. This preliminary action ensures that duplicate pages are identified and grouped together in advance, so that when transferred to the target system with larger page size, the duplicate sequences can be effectively processed by the cache-based deduplication mechanism, thereby maintaining deduplication performance despite the page size difference
Solution Approach 2:
The invention applies parameter changes by adjusting the hash value computation to use the target storage system's larger native page size. This parameter adjustment allows the source system to generate hash values that correspond to the target system's page structure, enabling proper duplicate identification and grouping even though the physical page sizes differ between source and target systems
Data Source
AI summary
An apparatus in an illustrative embodiment comprises at least one processing device comprising a processor coupled to a memory. The processing device is configured to identify a storage volume to be migrated from a source storage system to a target storage system, and to issue one or more commands to at least one of the source storage system and a host device as part of a migration session for the storage volume. Responsive to the one or more commands, a plurality of data pages are received, with the data pages having respective hash values that are computed based at least in part on a native page size of the target storage system as specified in at least one of the one or more commands. The one or more commands illustratively comprise a plurality of read next page commands, with a given one of the read next page commands including information identifying the storage volume, the native page size of the target storage system, and the migration session.