FPT Data Platform overview
FPT Data Platform gives you managed services to bring in, store, process, query, govern, and visualize data from the FPT Cloud portal. You create a workspace, then add only the services your team needs.
How FPT Data Platform is organized
You start from one of two places in the portal.
| You create | From | What happens next |
|---|---|---|
| A workspace | Data platform → Workspace management | You add services to the workspace from its details page. |
| A Lakehouse | Data platform → Lakehouse | Your team signs in to the Lakehouse Portal to work with data. |
A workspace is an isolated, secure environment for your data work. It runs on worker nodes and on a subnet that you choose when you create it. Every service you add runs inside the workspace, and you open each service from the workspace details page to check its status and settings.
Services by task
Every service on this page except Lakehouse is added to a workspace. Each service has its own section in this guide.
Bring data in
- CDC Service tracks changes in your databases in real time and moves the data into and out of Kafka through connectors that you define.
- Ingestion Service automates the movement of data between systems, and lets you track, monitor, and manage each data flow.
Store and organize data
- Lakehouse stores data of any type at large scale, with SQL queries, data versioning, and ACID consistency.
- Hive Metastore keeps the metadata of your tables, such as schemas, partitions, and data locations, so tools like Apache Spark, Trino, or Presto can find and read the data.
- Nessie helps data teams manage how data is developed, controlled, and deployed in large, distributed data environments.
Process data
- Flink processes continuous data streams with low latency using Apache Flink, and also supports batch and stateful processing.
- Processing Service runs batch or real-time data processing on compute resources that you configure, in CPU or GPU environments.
Query and govern data
- Query Engine runs distributed SQL queries, powered by Trino, across relational databases, data warehouses, and non-relational storage without moving or copying the data.
- Ranger gives you central, fine-grained access control for data you query with Trino, using RBAC or ABAC models.
- Open Metadata collects and organizes metadata from many data sources in one place, with classification, data lineage, and alerts when something changes.
Schedule workflows
Orchestration runs your data processing tasks in sequence or in parallel, on a schedule or when an event occurs, and helps you monitor and troubleshoot them.
Analyze and visualize data
- Apache Superset is an open-source business intelligence (BI) platform for visualizing data, building interactive dashboards, and analyzing data.
- JupyterHub gives your team a multi-user Jupyter Notebook environment for data analysis, data processing, and machine learning model development.
Next steps
- Get started with FPT Data Platform to sign in, create a workspace, and add your first service.
- Lakehouse overview to see what you can do with Lakehouse.