Unified VM and Storage Snapshots for Live Rollback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage management systems lack efficient online roll-back and roll-forward capabilities for live virtual machines, limiting their ability to revert or restore to past states effectively.
Innovation Solution
A method and system for creating a unified point-in-time image that combines data storage and virtual machine states, allowing for asynchronous storage and replaying actions using redo and undo logs to revert or roll-forward the system to a specific operational state, enabling live VM operation and snapshot management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data storage snapshots and virtual machine snapshots are managed separately, then existing snapshot mechanisms can be used, but unified point-in-time recovery of both data and VM state cannot be achieved
Solution Approach 1:
The patent combines data storage snapshots and virtual machine snapshots into a unified snapshot mechanism. The system creates correlated snapshots where both the data storage state and VM state are captured at the same point in time and managed together, enabling simultaneous restoration of both components to a consistent state.
Solution Approach 2:
The unified snapshot system serves multiple functions: it manages data storage snapshots, VM snapshots, and their correlations. The system can restore data storage independently, restore VMs independently, or restore both together, providing versatile recovery options from a single snapshot management framework.
2Reliability
If synchronous data transfer is used between geographically distributed directors, then data consistency is maintained, but transfer time becomes unacceptable for distant sites
Solution Approach 1:
The system dynamically adjusts the data transfer mode between synchronous and asynchronous based on the geographical distance and latency between directors. For distant sites where synchronous transfer would exceed acceptable thresholds, the system automatically switches to asynchronous transfer while implementing mechanisms to maintain data consistency and order.
Solution Approach 2:
The patent introduces a journal device as an intermediary that records all write operations in chronological order. This journal acts as a mediator that allows asynchronous replication between geographically distributed directors while maintaining write order fidelity, enabling distant sites to catch up on operations without requiring real-time synchronous communication.
3Loss of time
If asynchronous data transfer is used for distant sites, then transfer time is reduced, but maintaining proper write ordering becomes more difficult
Solution Approach 1:
The journal device serves as an intermediary that receives and sequences all write operations before they are replicated to distant directors. By acting as a central ordering point, the journal ensures that even though data is transferred asynchronously to reduce latency, the original write order is preserved and can be reconstructed at the destination.
Solution Approach 2:
The system performs preliminary ordering of write operations in the journal before asynchronous replication occurs. All writes are recorded and sequenced in advance in the journal device, creating a predetermined order that is then followed during asynchronous transfer to distant sites, ensuring write order fidelity is maintained despite the asynchronous nature of the transfer.
4Reliability
If virtual machines are stopped for snapshot restoration, then consistent state recovery is achieved, but operational downtime increases
Solution Approach 1:
The system performs preliminary actions by creating and validating unified snapshots of both data storage and VM state before restoration is needed. The snapshots are prepared and correlated in advance, so when restoration is required, the system can quickly apply the pre-prepared snapshot without needing to stop the VM for extended periods, reducing operational downtime while maintaining consistency.
Data Source
AI summary
The system described herein provides for unifying the streams of I/O, external VM snapshot triggers and VM internal operations into one in connection with providing for snapshotting of both the data storage of a storage system and one or more associated VMs over time. The system described herein thereby allows recording undo and redo journals for all the streams substantially simultaneously. The system described herein unifies snapshots for data storage and all associated VMs in response to one or more triggers. A tag of the unified snapshots may therefore automatically refer to a point-in-time of both the VM and array data.