H2O at Scale security
H2O at Scale has been deployed to numerous highly regulated enterprises in financial services, insurance, healthcare, and other industries. This typically involves extensive and thorough architecture and security reviews that vary across organizations. Key areas from an architecture standpoint are discussed in the following sections.
Data movement and privacy
Top security points for data movement and privacy are noted here:
- Data moves directly from the enterprise data source to the memory of the user's H2O cluster on the enterprise cluster and does not pass through H2O Enterprise Steam or the data scientist's client.
- Data in concurrent H2O clusters is isolated from other data: users cannot see or access each other's data during model building.
- Data in a user's H2O cluster is removed when the H2O cluster is stopped. Recall that data is partitioned in memory across a user's running H2O cluster nodes.
- Users can...
