It enables Jupyter to utilize distributed cluster resources by running kernels as Apache Spark applications in YARN cluster mode or as independent pods in a Kubernetes cluster.
It provides end-to-end security such as: encrypted HTTP communication between Jupyter Notebook and Gateway and secured socket communications between Gateway and remote kernels.
It enhances security and sandboxing of all kernels by enabling user impersonation leveraging Kerberos.
All of these, available as free open source software, built by a community for the community.