Data Migration Interception for CAS Deduplication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data migration processes between content addressable storage (CAS) and non-CAS systems are inefficient, particularly when performing deduplication, as they often involve sequential read requests and wasteful data retrieval of all-zero pages.
Innovation Solution
Implementing a method that intercepts I/O operations on the target system, uses a hash-to-address table to identify and migrate data pages, and updates structures to track migration status, allowing for efficient data transfer and deduplication by pushing data from the source to the target system based on the hash-to-address table, and handling read and write operations to ensure data integrity and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sequential read requests are used for data migration, then data can be transferred from source to target system, but data migration efficiency deteriorates due to unnecessary data retrieval of all-zero pages
Solution Approach 1:
The patent applies preliminary action by performing deduplication operations before data migration. The source system identifies and removes duplicate data pages (including all-zero pages) prior to migration, so that only unique data pages are transferred to the target system. This prevents the waste of time retrieving and transferring redundant data, directly resolving the contradiction between data migration efficiency and time consumption.
Solution Approach 2:
The patent extracts and removes duplicate data pages from the source system before migration. By identifying duplicate pages through hashing and removing them in advance, the system ensures that only necessary unique data is migrated, eliminating the inefficiency of transferring redundant all-zero pages and improving overall migration efficiency.
2Reliability
If all data pages are retrieved during migration, then complete data transfer is achieved, but data retrieval becomes wasteful due to retrieval of duplicate pages
Solution Approach 1:
The system performs preliminary deduplication on the source system before migration, identifying and removing duplicate data pages in advance. This ensures that when data is migrated to the target system, only unique pages are transferred, maintaining data completeness while eliminating the energy waste associated with retrieving and transferring duplicate pages.
Solution Approach 2:
The patent implements a feedback mechanism where the source system provides information to the target system about which data pages have already been migrated. The target system uses this feedback to avoid requesting duplicate pages, thereby reducing energy consumption while ensuring complete and reliable data transfer.
3Loss of substance
If deduplication is performed during migration, then data redundancy is reduced, but processing time increases due to sequential operations
Solution Approach 1:
The patent performs deduplication as a preliminary action before data migration begins. By removing duplicate data pages from the source system in advance, the system reduces data redundancy without adding processing time during the actual migration phase. This separates the deduplication process from the migration process, avoiding time penalties during data transfer.
4Reliability
If I/O operations are intercepted on target system, then data integrity is ensured, but system complexity increases
Solution Approach 1:
The patent introduces an intermediary component in the form of an I/O interceptor on the target system. This intermediary captures and manages I/O operations during migration, ensuring data integrity by controlling how data is written to the target system. While this adds some complexity, it provides a structured approach to maintaining reliability during the migration process.
Data Source
AI summary
Techniques for data migration may include: copying data of a source logical device of a source system to a target logical device of a target system; during said copying, receiving at the target system an I/O operation directed to a logical address of the target logical device and intercepting the I/O operation on the target system; determining, on the target system, to request from the source system a data page stored at the logical address; responsive to determining to request the data page stored, performing processing including: issuing a request to the source system for the data page stored at the logical address; and responsive to receiving said request, sending information from the source system to the target system, wherein the information includes the data page stored at the logical address and additional logical addresses of the source logical device at which the data page is stored.