CIRRUS applications inspire research community by showcasing platform's power

 

The newly launched CIRRUS User Applications page showcases the thriving applications already hosted on NSF NCAR's groundbreaking cloud platform — CIRRUS, or the Cloud Infrastructure for Remote Research, Universities, and Scientists.

     See related CISL News article about CIRRUS.
 

“The examples showcase different ways you can customize CIRRUS to fit your science and your research, whether that's a JupyterHub that's custom-built to go through a tutorial on specific tool sets, an interactive dashboard that has a plot, or a map with sliders,” said Nick Cote, the platform’s Technical Product Owner. 
 

“Seeing the CIRRUS applications out there right now spurs creativity and helps people realize what they can do on the platform.” 

— Nick Cote, CIRRUS Technical Product Owner

The product team is continuously updating the applications page to help researchers explore real-world examples and discover new possibilities for their own scientific services.

 

Launched in June 2025, CIRRUS accelerates the research lifecycle by streamlining how scientists deploy and share their work. From initial code development to live-hosted collaborative projects, the CIRRUS platform offers a seamless path to production. 

 

CIRRUS enriches research papers by allowing users to provide a live link for visualizing their research. Instead of relying on a static image or screenshot, a live, interactive visualization complements the project and better conveys the user's message.

 

A key benefit of the platform is that it simplifies the user workflow by managing everything directly within the codebase. Changes made daily are reflected instantly and in real-time, requiring no extra steps. Users simply integrate this component into their existing process, and it automatically handles the powerful, real-time realization.

 

The speed and ease with which changes are realized are transformative: "I can add a change into my codebase and push it live within five minutes," said Cote. "It's really, really quick."

 

Cote pointed out another "really neat functionality": CIRRUS can deploy applications either internally for UCAR employees, or externally for the general public. 

 

He noted, “Users might want to deploy things that are only available to UCAR employees and not the general public, for security or other reasons. We can do both easily.” 
 

 

“People can bring us their ideas. We're here to help make them reality.” 

— Nick Cote

 

Cote added that this approach frees researchers to focus on their core work. "Scientific researchers get to focus on the code that they care about, and everything else is handled automatically," he said. "Researchers don't have to worry about infrastructure anymore. They can just focus on the actual code and what they want to deliver."

 

The platform has received excellent feedback from users, including a representative from the German Climate Computing Center, DKRZ (Deutsches Klimarechenzentrum).

 

One major benefit of CIRRUS is its support for a wide array of programming languages. While the user applications page currently includes examples written in Python and R, any programming language can be used, provided the code is run within a container. Since the container simulates a Linux environment, users can install and run any required software, increasing accessibility for others.

 

Another significant innovation is the platform's robust GitHub integration. Initially, CIRRUS allowed individual connections to GitHub repositories. Now, this capability is available at the organizational level, greatly simplifying and streamlining the connection of GitHub runners for continuous integration and deployment.

 

Cote explained that CIRRUS eliminates common obstacles. Previously, researchers who wanted to host a website for their work faced significant challenges: "They may have always wanted to run a website for their research, but there was always overhead, or they didn't have a server to run it on, or they didn’t know how to do it, or it was too complicated."

 

This solution, Cote emphasized, "takes out the complexity." The only requirement for the user is to "be able to run your program, get a plot or whatever you're doing." He continued, "We can take care of the rest of the infrastructure pain points that people didn't want to deal with in the past."

 

 

“Cloud-native applications typically provide a better experience for developers.” 

— Kevin Hrpcek, HPC Systems Engineer

CIRRUS opens up a multitude of new possibilities: for example, it simplifies complex application deployment using flexible containerization, combining open-source and custom images. This allows easy integration of services like PostgreSQL for backend APIs (Application Programming Interfaces), or AMQP (Advanced Message Queuing Protocol) for message-based architectures.

CIRRUS also seamlessly connects components to enable powerful workflows, such as automated data ingestion, processing, and delivery. For example, a developer uploads data to CIRRUS S3 storage, which automatically sends an Advanced Message Queuing Protocol (AMQP) message via RabbitMQ upon object addition. Argo Workflows, managed by the CIRRUS team, monitors these queues. Receiving the message triggers a batch processing workflow to download and execute specified processing steps on the new object.

Furthermore, CIRRUS can host services that run behind the scenes, such as an AI model for data processing, which can be exposed via a website or a web-based API. The Mesonet example clearly illustrates this capability. While the map-server link provides a single-view, user-friendly map, it relies on many different components hosted on CIRRUS that aren't directly visible.

 

“The CIRRUS team is here to help with the learning curve.” 

— Kevin Hrpcek

 

These underlying components might include databases or an AI element providing a general forecast. The core concept is that processes capable of running on a laptop or larger system can be broken down into and transferred to run as individual components, or microservices. Instead of a single large application, the process is split into separate, distinct, and independently operating parts that function as a whole.

“The platform itself brings a high-availability solution to every workload on it, removing the overhead of users having to run and manage their own hardware or VMs,” said HPC Systems Engineer Kevin Hrpcek. 

"Cloud native computing and GitOps is a very different way of designing, managing, and deploying applications that many people aren't used to,” added Hrpcek. “One of our goals is to help educate NSF NCAR and university users of the advantages of designing applications and workflows to work in this type of environment. There is a learning curve to designing for a Kubernetes-based platform, but the CIRRUS team is here to help migrate applications and workflows. After an application is designed in a cloud-native fashion it usually provides a better experience for developers to update and manage applications and workflows."

 

"If you have an idea and you're not sure if it's feasible or where to get started, come talk to us." 

— Nick Cote

 

"It's important for us to showcase what's been done and get more people on CIRRUS, because there’s a lot more that can be done," said Cote. "We want to make sure that people understand that this is an option."

 

Cote added that continuing to build the application library is essential, as supporting researchers is a top organizational priority for NSF NCAR. Therefore, researchers are encouraged to submit their own application examples to be featured on the page.

 

To learn more about hosting your project on CIRRUS, or to submit your own application for the example library, please contact the CIRRUS team at cirrus-admin@ucar.edu.