Remote Agent System for Scalable Virtual Machine Backup
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data protection systems in computing devices face challenges in efficiently managing data backups across varying data generation rates and network conditions, leading to potential data loss and service degradation.
Innovation Solution
A remote agent system dynamically instantiates and terminates remote agents to provide data protection services to virtual machines, predicting future workloads based on asset information to optimize resource allocation and minimize computing resource consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If remote agents are continuously running to provide data protection services, then service availability is improved, but computing resource consumption increases
Solution Approach 1:
The system dynamically instantiates and terminates remote agents based on predicted workload. The workload predictor analyzes asset information to forecast future backup demands, and the resource manager adjusts the number of active remote agents accordingly, transitioning from static continuous operation to dynamic adaptive operation.
Solution Approach 2:
The workload predictor performs preliminary analysis of asset information to forecast future backup workloads before they occur. This allows the system to proactively instantiate remote agents in advance of predicted high-demand periods, ensuring service availability without keeping agents continuously running.
2Productivity
If more remote agents are instantiated to handle increased backup workload, then backup capacity is improved, but system complexity increases
Solution Approach 1:
The workload predictor acts as an intermediary between the asset information and the remote agent management system. It translates complex asset analysis into simplified workload predictions, which then guide the resource manager in adjusting agent numbers, reducing the complexity burden on the overall system.
Solution Approach 2:
The system implements self-service through automated workload prediction and dynamic resource allocation. The workload predictor automatically analyzes asset information and generates predictions, which the resource manager uses to autonomously instantiate or terminate remote agents without manual intervention, managing complexity through automation.
3Ease of operation
If backups are performed at fixed scheduled times, then resource allocation is simplified, but service degradation occurs during high data generation periods
Solution Approach 1:
The system transitions from static fixed-schedule backup timing to dynamic workload-driven timing. The workload predictor continuously analyzes asset information to forecast backup demands, enabling the system to adaptively schedule backups based on actual conditions rather than rigid predetermined times.
Solution Approach 2:
The system implements feedback through the workload predictor, which continuously monitors asset information and uses it to adjust backup scheduling decisions. This closed-loop feedback mechanism allows the system to respond to changing conditions and maintain service quality by scheduling backups when workloads are appropriate.
Data Source
Figure 1
Figure 2.1
Figure 2.2
AI summary
A remote agent, for providing data protection services to virtual machines, includes persistent storage and a backup manager. The persistent storage is for storing a resource allocation calendar and lifecycle policies associated with the virtual machines. The backup manager obtains a virtual machine data package from a virtual machine of the virtual machines; identifies a computational load change point based on the lifecycle policies; predicts a magnitude of a computational load during a period of time demarcated by the computational load change point using, at least in part, the virtual machine data package; updates the resource allocation calendar to obtain an updated resource allocation calendar, the updating is based on: the identified computational load change point, and the magnitude of the computational load during the period of time; and generates a backup for the virtual machine using the updated resource allocation calendar.