Connect Spark Compute
Tecton comes equipped with the AI/ML-native Rift Compute Engine.
You may optionally choose to use Spark Compute with Tecton for running materialization jobs. Spark Compute for materialization can be a good option if your organization already uses Spark widely for ETL and ML, and requires scalability beyond what Rift can provide.
Once Tecton has been connected to a Spark provider, whether to use Spark or Rift compute can be configured at the Feature View level.
To connect Tecton to Spark compute, follow the guide for your Spark provider below.
🗃️ Databricks on AWS
3 items
🗃️ Databricks on GCP
1 item
🗃️ EMR
3 items
🗃️ Dataproc
2 items