Primary patterns

There are three basic patterns for deployment related to containerization / serverless platforms:

• Customer Managed - Docker image

  Publish Talend Jobs (executable jars) and Data services/Mediation Routes (Microservice - Sprint Boot) as a Talend Docker image, then deploy the image into container management platform for orchestration and compute. The standard Docker image setting can be customized if needed (for example the image name, tag, and so on). For more information, see Customizing Docker images build settings

• Customer Managed – Binary

  Publish Talend Jobs (executable jars) and Data services/Mediation Route (Microservice - Sprint Boot) as standalone executables to artifact repositories (such as Nexus or Artifactory) and to an on-premise containerization or Cloud serverless platform (AWS Fargate, Azure Container Instances and so on) for execution. The relevant environmental setup will be required on the serverless platform itself, such as security, network, logging, monitoring, reporting and orchestration capabilities according to your project requirements.

• Talend Managed – Server as Docker image

  Deploy a Talend infrastructure application image (such as Talend Remote Engine) to container/serverless architecture. With this approach you can leverage Talend infrastructure services APIs to achieve infrastructure auto-scaling, auto key-pairing, automatic code deployment, and so on. The general architecture blueprint with this pattern will still follow the Talend-managed Reference Architecture guideline.

Talend encourages the first two approach patterns as the third pattern brings server applications back to the containerization/serverless architecture. Talend would also recommend customers use Talend Professional Services to better support and provide the most suitable implementation approach based on each customer’s data management requirements, environment readiness and IT containerization/serverless maturity.