This architecture uses data science and machine learning features embedded in Oracle Autonomous Data Warehouse (ADW) to analyze data from a broad range of enterprise data resources for business analysis and machine learning.

The following components apply to the entire Oracle Cloud Infrastructure region:

• Governance (Data Catalog)
• Infrastructure, Network Security. Identity and Access Management (IAM)

The region is divided into functional layers that house physical or functional components:

• Data Sources: Data sources can and will be anything. Those shown include end users, devices, enterprise applications, any digital asset, events, sensors, and machine learning (ML) models.

• Ingest, Transform: Data moves into a block labeled Batch Ingestion with DB Tools (provided by ADW) listed as the recommended option and Oracle Data Integration as the secondary option. Refined data passes to a block labeled Serving Data Store in the Persist, Curate, Create layer.

  Metadata either bypasses DB Tools and is processed in a block labeled Batch Ingestion (Oracle Cloud Data Integration) or is fed into cloud storage and then to batch ingestion using DB Tools provided by ADW. Metadata from object storage moves into the Governance (Data Catalog) block.

• Persist, Curate, Create: Data moves into the block labeled Serving Data Store (Autonomous Data Warehouse) to blocks in the Analyze, Learn, Predict layer. Metadata from ADW moves into the Governance (Data Catalog) block.

• Analyze, Learn, Predict: Data moves from ADW to the following blocks:

  • Analytics and Visualization (Oracle Analytics Cloud). Metadata also moves into the Governance (Data Catalog) block.
  • Learn with Oracle Machine Learning Notebooks (provided by ADW) listed as the recommended option and Oracle Machine Learning AutoML UI, 3rd Party UI listed as secondary options. The results are passed to the Predict block and to people and partners in the Measure, Act layer.
  • Predict with Oracle Machine Learning services (provided by ADW) listed as the recommended option and In-Database, Oracle Machine Learning Notebooks) listed as secondary options.
  • Data moves from the Predict block to the Data Products, APIs block with API Gateway listed as the recommended option and Oracle Integration Cloud and Oracle Functions listed as secondary options. The results are passed to data consumers (things, applications, people and partners) in the Measure, Act layer
• Measure, Act: Includes data consumers such as things, applications, and people and partners.