Selective Block Migration Notifications for Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data migration methods in storage systems, such as early and late cutover techniques, face challenges in efficiently managing resources and ensuring reliable data transfer, especially during heavy I/O operations, leading to performance issues and potential data loss or corruption.
Innovation Solution
Implementing a system with selective migration notifications that includes a host system with a processor, memory, and storage interface to communicate with source and destination storage nodes, using a migration table and logic to selectively identify and transfer data blocks, and manage storage requests between nodes, ensuring seamless data migration without interrupting I/O operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If early cutover technique is used to migrate data blocks, then data migration can begin immediately, but storage performance deteriorates because the destination system must handle all storage requests and manage bookkeeping on transferred data blocks
Solution Approach 1:
The patent introduces a cutover manager as an intermediary component that coordinates between source and destination storage systems. The cutover manager receives notifications from the destination system about migrated data blocks and manages the switching of storage requests, preventing the destination system from being overwhelmed by all storage requests while enabling progressive migration.
Solution Approach 2:
The patent implements preliminary actions by having the destination storage system notify the cutover manager about migrated data blocks before the host system needs to access them. This allows the system to prepare for cutover in advance, tracking which data blocks have been migrated and are ready for access at the destination, thus avoiding performance degradation during actual migration operations.
2Productivity
If late cutover technique is used to determine convergence window, then storage performance is maintained, but data migration reliability deteriorates due to frequent modifications during heavy I/O operations making it difficult to hit the convergence window
Solution Approach 1:
The patent implements a feedback mechanism where the destination storage system sends notifications to the cutover manager about successfully migrated data blocks. This feedback allows the system to track migration progress in real-time and determine appropriate cutover points based on actual migration status rather than relying on predicted convergence windows, thereby improving reliability during heavy I/O operations.
Solution Approach 2:
The patent makes the cutover process dynamic by allowing the cutover manager to adaptively determine when to switch storage requests based on received notifications about migrated data blocks. Instead of using a fixed convergence window, the system dynamically adjusts the cutover timing based on actual migration progress and system conditions, ensuring reliability while maintaining performance.
3Productivity
If destination system handles all storage requests during migration, then data migration can proceed, but device complexity increases due to the destination system managing bookkeeping on transferred data blocks
Solution Approach 1:
The patent extracts the bookkeeping functionality from the destination storage system and places it in a separate cutover manager component. The destination system only needs to notify the cutover manager about migrated data blocks, while the cutover manager handles the complex task of tracking and managing which data blocks have been transferred and are ready for access, thereby reducing destination system complexity.
Data Source
AI summary
Systems and methods for storage operation processing during data migration using selective block migrated notifications are disclosed. A host system may be configured with connections to a source storage node and a destination storage node while a data migration is moving data blocks from the source to the destination. The host may send a storage request to the source storage node and receive a block migrated notification from the source storage node. The host may then store a migrated indicator for that data block in a migration table and direct future storage requests to the destination storage node.