Deploying the Open-Source Edition
The open-source edition of qData Data Mid-End (also known as the Community Edition) can be deployed using one of the three methods listed below.
| Deployment Method | Description | Use Case |
|---|---|---|
| Docker Compose Deployment | All components (scheduler, database, message queue, Spark, Flink, etc.) and the qData data mid-end source code are started with a single command via Docker Compose. | Beginner onboarding, feature demonstration, testing environments |
| Local Startup from Source Code | The qData data mid-end source code is run locally by the developer, while dependent components are launched via Docker Compose. | Daily development, integrated debugging |
| Manual Deployment (Fully Hand-Installed) | All dependent components and the qData data mid-end services must be manually installed and configured. | Production environments, large-scale deployment, scenarios requiring deep customization |
1. Docker Compose Deployment
All dependent components (such as the scheduler, database, message queue, Spark, Flink, etc.) and the qData data mid-end services are defined in a Docker Compose configuration file, allowing the entire environment to be launched with a single command.
Characteristics:
- One-click startup with low deployment barrier
- High environment consistency, avoiding the "it works on my machine" problem
- Ideal for quick trials, feature demonstrations, and test validations
- Not suitable for large-scale production environments (limited scalability and operational capabilities)
2. Local Startup from Source Code
The core code of the qData data mid-end is run locally by the developer (via IDE debugging or command line), while dependent components (database, message queue, computing engines, etc.) are launched via Docker Compose to provide supporting services.
Characteristics:
- Retains source code execution, facilitating development and debugging
- Dependencies are containerized, reducing manual configuration
- Suitable for R&D teams during daily development and integration testing
- Faster deployment than fully manual setup, but requires moderate local resources
3. Manual Deployment (Fully Hand-Installed)
All components must be manually installed and configured by the user, including dependencies such as database, message queue, scheduler, Spark, Flink, and the packaging and deployment of the data mid-end platform itself.
Characteristics:
- Highly flexible and customizable—component versions, parameters, and deployment architecture can be tailored to specific needs
- Suitable for production environments or scenarios requiring deep customization
- Requires significant expertise in operations and big data platforms; complex configuration and high deployment cost
- Closer to traditional enterprise operations practices