Parallel LUN Migration via Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data migration methods for large logical unit numbers (LUNs) are inefficient, taking up to 20 hours to migrate source arrays of 1 terabyte or larger, due to sequential scanning and copying processes.
Innovation Solution
Divide a large LUN into smaller, manageable partitions and migrate these partitions in parallel using multiple sessions, allowing for simultaneous scanning and copying, thereby reducing the overall migration time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sequential scanning and copying process is used for large LUNs, then data migration can be performed using a single session, but migration time becomes excessively long (up to 20 hours for 1TB or larger arrays)
Solution Approach 1:
The patent divides a large LUN into multiple smaller partitions and creates multiple sessions to migrate these partitions in parallel. This segmentation allows the migration process to simultaneously copy data from different portions of the source array, dramatically reducing overall migration time compared to sequential processing.
Solution Approach 2:
The patent transitions from a single-dimensional sequential migration approach to a multi-dimensional parallel migration approach by creating multiple sessions that operate simultaneously. This dimensional change in the migration process enables concurrent data transfer across multiple partitions, improving productivity while managing the complexity of large LUN migrations.
2Productivity
If large LUNs are migrated as a single unit, then session management is simplified, but resource utilization becomes inefficient and migration performance deteriorates
Solution Approach 1:
The patent segments a large LUN into multiple smaller partitions, each handled by a separate session. This segmentation improves resource utilization and migration performance by enabling parallel processing, while the system manages the increased complexity through automated session creation and coordination mechanisms.
Solution Approach 2:
The patent creates multiple sessions that can operate independently yet concurrently, allowing each session to utilize available resources efficiently. This multi-session approach provides universal applicability for migrating large LUNs by adapting the number of sessions based on LUN size and available resources, optimizing both performance and resource utilization.
Data Source
AI summary
An example methodology includes, by a computing device, receiving a request to migrate a logical unit number (LUN) from a source array to a target array and, responsive to a determination that the LUN is a large LUN, dividing the large LUN into a plurality of partitions, creating a number of sessions corresponding to the plurality of partitions, wherein each session of the created sessions is operable to migrate data from a partition of the plurality of partitions, and initiating the created sessions to start copying data from the respective partitions in parallel. The method may also include, by the computing device, responsive to a determination that the LUN is not a large LUN, creating a single session operable to migrate data from the LUN and initiating the single session to start copying data from the LUN.