Hyper-Converged Storage Volume Migration via Metadata Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In computer systems with hyper-converged infrastructure, traditional data migration methods fail to adequately reduce migration load due to varying metadata amounts, especially when large amounts of data are involved.
Innovation Solution
A computer system that calculates metadata amounts for each volume, determines a migration target volume based on these amounts, and performs volume migration by transferring the computer node providing the volume to a destination node, thereby reducing migration load without transferring actual data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional data migration is performed to distribute load across servers, then load distribution is achieved, but migration load increases due to transferring large amounts of data
Solution Approach 1:
The patent extracts only the metadata (management information) from the complete data set and migrates only this extracted portion to the target server. The actual data remains on the storage device, eliminating the need to transfer terabytes of data while still achieving load distribution by relocating the compute node that accesses the data.
Solution Approach 2:
The patent segments the data access functionality into two separate components: metadata (management information) and actual data. By migrating only the metadata segment to the target server while leaving the data segment on the storage device, the system achieves efficient load distribution without the overhead of migrating the entire data set.
2Quantity of substance
If meta-migration is performed to reduce migration load, then data transfer is eliminated, but migration load cannot be reduced appropriately when metadata amount is large
Solution Approach 1:
The patent changes the migration approach from migrating compute nodes to migrating storage resources (metadata). By changing the migration parameter from node-level to storage-level, the system achieves more granular and efficient load distribution, migrating only the necessary metadata portion rather than entire compute nodes.
3Productivity
If volumes with large metadata amounts are migrated, then load distribution is achieved, but migration time and resources increase
Solution Approach 1:
The patent applies local quality by treating different volumes differently based on their metadata characteristics. Volumes are evaluated individually, and migration decisions are made based on each volume's specific metadata amount and access patterns, allowing the system to optimize migration for each local case rather than applying a uniform migration strategy.
Data Source
AI summary
Each server stores metadata for managing a volume having a logical block mapped to physical blocks of storage devices, and provides a compute host with the volume according to the metadata. A computer system calculates the metadata amount, which is the data amount of the metadata of the volume, for each volume, determines a migration target volume from the volumes based on each metadata amount, and performs volume migration that transfers the computer node that provides the compute host with the migration target volume to a migration destination computer node.