Hash-Based Storage Volume Migration Bandwidth Conservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data migration approaches in information processing systems often result in excessive data transfer over network links, straining bandwidth resources due to the need to migrate all data, even if some data already exists in the target storage system.
Innovation Solution
Implement a hash-based migration method where the target storage system identifies and requests only data pages not already present, using content-based signatures to determine existing data and update reference counts, thereby avoiding unnecessary data transfers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional migration approaches transfer all data from source to target storage system, then complete data migration is achieved, but network bandwidth resources are excessively consumed
Solution Approach 1:
The patent applies preliminary action by computing hash digests of data pages before migration and storing them in a mapping table at the target storage system. This allows the target system to identify which data pages already exist locally before actual data transfer occurs, preventing redundant network transmissions while ensuring complete migration of only necessary data.
Solution Approach 2:
The patent extracts the essential identification function from the migration process by separating hash digest computation and existence checking from the actual data transfer. The mapping table stores only hash digests and logical-to-physical address mappings, not the actual data, allowing the system to determine migration needs without transferring unnecessary data pages.
2Loss of energy
If hash-based migration is implemented to reduce data transfer, then network bandwidth is conserved, but system complexity increases due to hash computation and mapping table management
Solution Approach 1:
The patent introduces a mapping table as an intermediary structure that simplifies the migration process. This mapping table stores hash digests and address mappings, acting as a mediator between the source and target storage systems. Instead of complex real-time comparisons during migration, the system uses this pre-computed mapping table to quickly determine which data pages need transfer, reducing overall system complexity.
Solution Approach 2:
The patent replaces mechanical data comparison and transfer operations with hash-based digital identification. Instead of comparing actual data content byte-by-byte or transferring all data to verify existence, the system uses hash digest computation and comparison, which is computationally more efficient and simplifies the migration mechanics.
3Productivity
If content addressable storage functionality is added to source storage system to enable hash generation, then migration efficiency improves, but source storage system complexity and cost increase
Solution Approach 1:
The patent applies universality by making the hash computation capability optional and multi-functional. The source storage system can either have native hash computation capability or rely on the target system to perform it. This universal approach allows the same migration framework to work with both simple storage systems and advanced content-addressable systems, accommodating different complexity levels without requiring modifications to the core migration process.
Data Source
AI summary
An apparatus in one 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 for each of a plurality of logical addresses of the storage volume, to send a command requesting a content-based signature for the logical address to at least one of the source storage system and a host device and to receive the content-based signature in response to the command. Responsive to a first one of the received content-based signatures having a corresponding data page already stored in the target storage system, the processing device updates an associated reference count in place of requesting the corresponding data page. Responsive to a second one of the received content-based signatures not having a corresponding data page already stored in the target storage system, the processing device requests the corresponding data page.