The Zone

The Zone Team

Summer 2025

Featuring Anray Liu and Justin Zhang!

Co-op student developers from uOttawa and Carleton

Developers (IT-02):

• Wendy Gaultier, Mathis Marcotte, Jose Matsuda, Souheil Yazji

Team Lead (Acting IT-03):

• Bryan Paget

Our Story

One Platform, Two Zones

• The Advanced Analytics Workspace (AAW)
  was the first Zone
• The Zone is its Protected B counterpart
• Built on the same foundation:
  Kubeflow on Kubernetes

Same core. Same capabilities.
Higher security.

What Is The Zone?

The Zone is a data science platform based on Kubeflow, designed to orchestrate notebooks, jobs, and machine learning workflows.

Featuring:

• JupyterLab supporting Python, R, SAS, Julia
• Kubeflow for scalable infrastructure for data science and automation

The Zone is a platform that runs on Azure AKS
but can run on any Kubernetes cluster.

Everyone Is Welcome

Which Language? No Barrier.

• Python, R, Julia, SAS are all supported
• Migrate at your pace, in your language
• The only platform at StatCan where
  SAS and Python/R coexist

Everyone Is Welcome

Which Organization? No Barrier.

• Open source by design: cloneable, shareable, federatable
• We've already done it once: we can do it again
• Ready to support other teams, departments, and levels of government

Why Was The Zone Created?

To deliver a secure, modern, and independent data platform.

• Provide a Protected B compliant environment
• Reduce reliance on proprietary tools
• Long-term cost reduction through reusable, open infrastructure

This is about sovereignty, sustainability, and
self-reliance.

What We've Built

• Over 2,200 onboarded users
• 130+ daily notebook sessions
• CronJobs powering real production workflows
• A stable, proven platform used across divisions

This is not a prototype. This is production grade.

Summer 2025

• MKL acceleration for faster numerical computing
• Tesseract OCR to extract text from scanned documents
• Volume and namespace cleaners for automatic resource cleanup
• CronJobs for scheduling our pipeline-ready infrastructure
• Readiness Probe: no more loading errors
• Optimized Docker images for faster load times

From CronJobs to Pipelines

Today: CronJobs run in isolation.

Tomorrow: Connected, observable workflows powered by Argo Workflows and Kubeflow Pipelines.

• Argo Workflows orchestrates complex job sequences on Kubernetes
• Kubeflow Pipelines enable end-to-end ML workflows with UI, caching, versioning
• Full logging, retry logic, error handling, triggers

Automation, evolved.

Kubernetes Is Built for Orchestration

At its core, Kubernetes is designed to orchestrate workloads: scaling, scheduling, and managing containers across clusters.

• Powers modern cloud-native apps
• Handles complex workflows reliably
• Already runs The Zone's notebooks and jobs

Orchestration isn't the future.
It's already here, under the hood.

Kubeflow Brings Orchestration to Data Scientists

Kubeflow was built for pipelines. It brings Kubernetes' power to data scientists through an intuitive interface.

• The AAW had Kubeflow Pipelines, Argo Workflows
• We can re-enable them, no reinvention needed, for end-to-end workflows: versioned, reproducible, monitored

We're not starting from scratch. We're restoring what works and making it secure, scalable, standard.

Challenges Remain

• Database connections (workload intake forms)
• Legacy filer performance (data migration)
• Linux and Kubernetes concepts (education)
• Secrets management (Keycloak)
• VS Code extension repo (org wide)

We're solving these, with your feedback.

A Federated Future

The Zone is open source and designed to scale. It can be deployed by:

• Other teams at StatCan
• Federal departments
• Provincial and municipal governments

Same foundation. Same security.
Deployed where you need it.

Enter The Zone

You are invited to:

• Access the platform: https://zone.statcan.ca
• Attend training and workshops
• Help shape the future of data science at StatCan
• Host the next Zone in your department

The Zone is a movement toward openness, sovereignty,
and shared capability.