IBM DataStage Architecture

 The architecture contains three major layers:

---

1. Client Layer

This is where developers work.

• DataStage Designer → Build ETL jobs

• DataStage Director → Run and monitor jobs

• DataStage Administrator → Manage projects, permissions

These tools are installed on the developer’s machine.

---

2. Server Layer

This is the engine where ETL jobs run.

• DataStage Engine

• Parallel Engine

• Job execution components

• Runtime metadata

• Job logs

This layer handles:

• Job execution

• Resource allocation

• Parallel processing (using multiple nodes)

---

3. Repository Layer

Stores all DataStage metadata.

• Job designs

• Schema definitions

• Shared containers

• Table definitions

• Parameter sets

The metadata is stored in a repository database (XMETA).

---

✔️ Architecture Flow

1. Developer creates job in Designer

2. Job is compiled → converts to OSH (parallel job code)

3. Job runs on the engine layer using available resources

4. Logs and results are stored in repository

5. Director monitors job output