Hypervisor-Based VM Backup System for Data Center Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data backup methods in virtualized data centers are inefficient as they only backup individual virtual machines (VMs) without considering associations between them, leading to complexity and inconsistency in recovery processes, and dynamic data distribution across multiple nodes complicates the backup process.
Innovation Solution
A method and system that utilize a backup manager and hypervisor to uniformly backup data across multiple computation nodes by exporting data from virtual machines, storing it in a storage apparatus, and rearranging it based on memory segment addresses, allowing simultaneous backup of all VMs and their associations, thereby eliminating the need to worry about recovery orders and reducing system complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If individual VMs are backed up separately through their own OS, then each VM can be backed up independently, but the associations between VMs are lost and recovery complexity increases
Solution Approach 1:
The patent merges the backup operations of multiple VMs into a single unified backup process executed by the physical server. Instead of each VM performing separate backup operations through their own OS, the physical server's backup module consolidates all VM data into one backup set that preserves inter-VM associations, thereby reducing recovery complexity while maintaining backup independence.
Solution Approach 2:
The physical server acts as an intermediary between the VMs and the backup storage. The backup module on the physical server intercepts and centralizes the backup operations, managing the data collection from multiple VMs and storing them with association metadata intact, thus simplifying the recovery process without affecting individual VM backup capabilities.
2Ease of manufacture
If VM data is backed up from within each VM's OS, then the backup process is simple to implement, but the backed up data center is not fully consistent with the original system
Solution Approach 1:
Instead of having each VM's OS perform the backup operation (bottom-up approach), the patent inverts the approach by having the physical server's backup module perform the backup (top-down approach). This inversion ensures that the backup process views the data center as a unified system, capturing all VMs and their associations consistently, thereby improving system consistency while maintaining implementation simplicity through centralized management.
Solution Approach 2:
The backup module on the physical server provides a universal backup function that serves all VMs simultaneously. This multi-functional approach allows a single backup mechanism to handle multiple VMs with their inter-dependencies, ensuring that the backup reflects the true state of the entire data center system rather than isolated VM states.
3Productivity
If dynamic data of each VM is distributed across multiple computation nodes, then resource utilization is optimized, but the backup process becomes more complicated and may influence other VMs
Solution Approach 1:
The patent extracts the backup function from the individual VMs and their distributed storage locations. The backup module on the physical server directly accesses and extracts data from the distributed storage locations of multiple VMs, consolidating them into a unified backup set. This extraction approach maintains the distributed storage architecture for optimal resource utilization while simplifying the backup process by centralizing the data collection and association management at the server level.
Data Source
AI summary
A method for backing up data in a data center, device and system are provided. The data in all the VMs of one data center can be backed up simultaneously, and the associations between the VMs are also backed up during the backup, so it is unnecessary to concern the recovery orders of the VMs when the data of the data center is recovered. Meanwhile, when the stored data is exported from the hypervisor node, it is unnecessary to notify the upper layer OS for a backup, thereby improving the data backup efficiency and reducing the system logical complexity.