Server Backup Method for Virtual Machine Fault Tolerance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The virtual fault tolerance (FT) mechanism in server backup systems introduces additional delay time due to the sequential completion of phases before outputting user requests, which is problematic for time-sensitive workloads like online games or real-time trading, leading to potential internet disconnection or transaction failures.
Innovation Solution
A server backup method that dynamically adjusts the backup start time based on the quantity of collected dirty pages, allowing for continuous collection and suspension of operations to execute backup snapshots and transmissions, thereby controlling delay time within a predetermined range.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a virtual FT mechanism is applied to ensure fault tolerance and data safety, then system reliability is improved, but additional delay time is introduced that causes response time issues for time-sensitive workloads
Solution Approach 1:
The patent implements dynamic adjustment of backup cycle parameters based on real-time system state. The backup start time point is determined dynamically according to the quantity of collected dirty pages, allowing the system to adapt between frequent backups (for reliability) and less frequent backups (for reduced delay) based on current workload characteristics and time sensitivity requirements
Solution Approach 2:
The patent changes the parameter of backup timing by determining the backup start time point based on the quantity of dirty pages collected. This parameter change allows the system to optimize between data safety (more frequent backups) and response time (less frequent backups) by adjusting when backups are initiated based on actual data change accumulation
2Reliability
If the backup snapshot operation and transmission operation are executed sequentially before outputting user requests, then data consistency is ensured, but the delay time increases causing internet disconnection or transaction failure
Solution Approach 1:
The patent applies preliminary action by continuously collecting dirty pages in the background during the running operation phase before the backup is actually needed. This allows the system to prepare backup data incrementally without interrupting user requests, so that when a backup is triggered, the data is already prepared and can be transmitted quickly
Solution Approach 2:
The patent segments the backup process into distinct phases: continuous dirty page collection during running operation, snapshot generation at the determined start time point, and transmission after suspension. This segmentation allows different phases to operate independently, with the collection phase not blocking user requests while the snapshot and transmission phases ensure data consistency
Data Source
AI summary
A server backup method and a backup system using the server backup method are provided. The server backup method includes continuously collecting a plurality of dirty pages during a running operation and determining a backup start time point according to a quantity of the collected dirty pages. The server backup method also includes suspending the running operation according to the backup start time point and executing a backup snapshot operation to generate a data backup snapshot corresponding to the dirty pages, and executing a backup transmission operation to transmit the data backup snapshot.