Data Separation
Machine Learning provides a multi-tenant, cloud-based, SaaS solution where clients share underlying infrastructure with other tenants. We store all tenant data in logically discrete storage segments.
This document describes the different layers of security and separation we employ to keep data safely compartmentalized.
Application-level Separation
Each user owns the data uploaded to Machine Learning and any results derived from that data. At the time of publishing this document, the ability to share data between users isn’t available. Users are only able to see the data they upload.
Network-level Partitioning
Networks are physically and logically separated to keep the development environments distinct from the production environments.
Network Security Groups are utilized to implement Access Control List allowlisting.