Overview

Why is security important?

  • Having a secure environment allows access to more sensitive and private data
  • Many data use agreements include security compliance
  • Health related data might be subject to HIPAA requirements
  • Data breaches are in general bad

What is this poster about?

  • We built a secure cloud platform for data analysis
  • We feature Jupyter but extend the concept to other applications
  • We try to keep the principlel here simple
    • Security through simplicity
  • The system requires less effort to maintain by leveraging features provided the our cloud provider Amazon Web Services (AWS)

Pitfalls of other approaches

  • Jupyter Hub does not have any well establish solutions for encryption, especially in transit
  • Kubernetes is not by default secure and requires significant effort to secure.
  • Kubernetes is very complex

In a nutshell

  • Elastic Container Service is used
  • Leverage the Application Load Balancer for authentication through an external identity provider
  • Secure attached storage for collaboration
  • Underlying EC2 instances are secured through installation of a variety of agents