Fortune 500 NOV achieves flexible, consistent data science workspaces
Headquartered in Houston, Texas, the heart of the world’s oil and gas industry, NOV (formerly National Oilwell Varco) is a Fortune 500 independent provider of oilfield equipment and technologies. Originally founded in 1862, NOV and its predecessor companies have spent 159 years helping transform oil and gas field development and improving its cost-effectiveness, efficiency, safety, and environmental impact.
Joshua Cluff is the Data Science Operations Manager at NOV. As such, he is responsible for defining and leading the adoption of best practices in model lifecycle (development, deployment, maintenance) across the data science organization of the entire company.
The problem: Existing platform failed to meet the growing needs of data science teams
With continual growth and development in data science, NOV knew that the traditional method of developing on local workstations was not sustainable and that moving development to the cloud was essential.
As an AWS customer, the team made a strong effort to utilize existing AWS services, specifically Cloud9—their cloud-based integrated development environment (IDE).
However, NOV’s various data science teams quickly pushed the boundaries of what Cloud9 could provide. As they transitioned from on-premises to cloud development, the team quickly grew increasingly frustrated with not being able to use PyCharm, RStudio, and VS Code.
They needed an adaptive solution that would meet the requirements and preferences of various data science teams. It was Joshua Cluff’s responsibility to find that solution.
Challenge: Must support multiple IDEs
The evolution of IDE adoption at NOV has followed a pattern often seen across the industry. The organization began with a large amount of legacy code in R. A collective decision was made to move new projects to Python. As a result, data scientists began to prefer working in JetBrains PyCharm and VS Code, neither of which is supported by Cloud9. The NOV data scientists also found the debugging tools available in Cloud9 to be lacking compared to those offered in PyCharm and VS Code.
Any solution would require support for legacy IDEs like RStudio as well as newer IDEs like VS Code and PyCharm with their robust debugging features.
Challenge: Must satisfy the security group’s requirements
A new solution would also have to be approved by NOV’s information security group. When scientists developed on local machines, the security group required they use VPN when connecting to resources within the firewall, resulting in a poor experience for the scientists as well as creating additional work for the sysadmins responsible for the VPN.
NOV sysadmins saw clear advantages to browser-based remote development platforms that required fewer VPN configurations, which would also mean fewer support tickets to wade through.
The solution: Coder
When Joshua Cluff discovered Coder, he knew that it had the potential to meet all of NOV’s varied needs.
Second, Coder supports multiple editors and IDEs, including RStudio, VS Code, and PyCharm (as well as other JetBrains IDEs). So whether working on legacy code or newer code in Python, NOV’s data scientists can use the tools they need.
Finally, Coder lets NOV’s data scientists connect to their workspaces with an encrypted TLS browser session, and each workspace resides securely in an isolated AWS cloud Kubernetes pod.
“Coder opens the door for data scientists to use the IDEs they prefer like RStudio, PyCharm, and VS Code,” says Joshua Cluff. “It simplifies environment management, removes the need for multiple VPNs, increases the security of data usage by restricting data to the cloud, and all without increasing the load on sysadmins.”
Learn about other use cases and see a demonstration of Coder
If you would like to see a live demonstration of Coder for your data science, front-end, and back-end software development needs, schedule a demo. You can also evaluate Coder with a free trial running on Docker or Kubernetes.