Research Computing Cloud

on Google Cloud Platform

Research Computing Cloud (RCC) is now available through Google Cloud Marketplace or Infrastructure-as-Code!

The RCC is a continuously developed and supported Research Computing Cloud for high performance, high throughput and scientific computing workflows. The cloud includes the software and services necessary for functional research computing on Google Cloud.

An RCC combines scalable system services with scalable licensing structures that fit the needs of today's research and science landscape. Teams can rapidly deploy their own systems using Google Cloud Marketplace products or their own infrastructure-as-code using the examples we provide or engage Fluid Numerics to assist them with their deployment and manage RCC services for them.

A primary component for many RCC users is the Research Computing Cluster that is complete with the Slurm job scheduler and research computing software development toolchains. Clusters come with the Spack package manager for installing research computing applications in addition to a suite of compilers (GCC, AOMP, and the Intel OneAPI Compilers) with accompanying builds of OpenMPI and Singularity for deploying containerized applications. This solution integrates with external file systems, including Lustre and Filestore which can provide additional high performance file IO for any research computing workload. Finally, the cluster-services command line utility allows you to easily customize your Research Computing Cluster's configuration on-the-fly giving you the flexibility to to take advantage of the benefits of Google Cloud.

The RCC is compatible with Google Cloud Private Catalog for simple software asset distribution and procurement if necessary.

Research Computing Cluster

Available in four operating system "flavors":

RCC-Centos

RCC-Debian

RCC-Rocky

RCC-Ubuntu

The Research Computing Cluster is a critical component for users that prefer to submit jobs through a resource or workload manager. This system is based on Slurm and our team can help you with various approaches to submitting your research computing workloads to this system. The image is pre-packaged with Singularity and handles package management using Spack. If your team needs an ephemeral research computing cluster rapidly deployed, this is our suggestion. Cloud Monitoring and other features that are helpful for admins and users are enabled out of the box.

If you have batch scripts for your jobs or want to work in an interactive session, this is the cluster for you. We offer software services for customers that need to provision batch scripts for their jobs.

If you are not dependent on a workload or resource manager and your application is completely containerized, you may not need an ephemeral research computing cluster. We have a listing of Applications that are available to run with the Research Computing Cluster and without in our public rcc-apps repository.

Continuous Integration and Continuous Benchmarking with Fluid Run

Fluid Run has been provided to the community as a DevOps tool that enables teams to build and test research computing application images for their RCC using a toolbox and workflow. We have provided this necessary tool for teams in need of a resource that establishes an auditable trail of data that comments on the performance of an application over time and accross various hardware. By using fluid-run as a build step with Google Cloud Build, developers can automate testing and benchmarking on GPU accelerated and multi-VM platforms hosted on Google Cloud. Information about each test, including the system architecture, software version (git sha), build id, and application runtime are recorded and can be saved to Big Query.

Our Google Cloud Marketplace deployments are designed to assist cloud administrators, architects, developers and engineers deploy commonly used applications and infrastructure on Google Cloud Platform. Once you have setup your GCP account, search our Google Marketplace Solutions for what you need and deploy at the click of a button. We are here to support scientific teams with launches and offer a variety of support solutions. Our Help pages are available to the community and we encourage your feedback on how we can make it better. Our team is available for contract support and our product offering is designed to assist beyond.

If your team needs help with the maintenance and support of their Google Cloud Marketplace products, get in touch. We might be able to help.

Google Cloud Academic Research Grants

We are researchers and scientists as well as business owners and operators. We understand the nature of bid & proposal as well as grants & awards. Our team is capable of assisting teams at eligible .edu academic and .org non-profit organizations with answering expected cost questions for applications to the Google Cloud research credit program. Last time we checked(12/9/21 08:38 AM MT) Google offers up to $5000 in credits through the program to qualifying projects. While Fluid Numerics does not receive licensing revenue for RCC resources that are utilized while a research credit is applied, we are available to assist you with your cluster before, during and after your project is complete. Any infrastructure and resources used for a project can be made available and expanded upon within any organization to utilize for future projects or different departments.

This program does request attribution for resource contributions. You can see an example of how this program was utilized and attributed by Dr. Suchetana Gupta for her in silico study, D155Y Substitution of SARS-CoV-2 ORF3a Weakens Binding with Caveolin-1: An in silico Study, posted March 2021. The RCC is ephemeral but it doesn't have to disappear. These self-managed resources are highly available and able to migrate easily. Data storage and management on Google Cloud is very flexible. Your research can virtually continue whenever or wherever you need.

Our click-to-deploy solution is great but you and your team could launch clusters from a python notebook if you prefer.

Image Licensing Discounts for Eligible Organizations

If you are an academic research organization or non-profit organization that is interested in Research Computing Cluster resources but concerned about the costs for licensed support, please contact us. We are able to offer discounts on our licensing rates to organizations which, once accepted, apply directly to the Google Cloud billing account attached to your Research Computing Cluster project. Request a Scientific Research RCC Discount Offer if you feel you are eligible for an offer.

Click-to-Deploy Solutions

Software & Infrastructure-as-Code Solutions

RCC Applications

Create apps with rcc-apps and set cloud triggers to build them.

Fluid Run

Use fluid-run as a build step to test the applications you create.

RCC Terraform

Deploy using rcc Terraform Infrastructure as Code.

Private Catalog Request

Managed RCC Private Catalogs are available by request.

Fluid Numerics and Google Cloud

Fluid Numerics is proud to be a Google Cloud Sales and Services Partner and offer cloud sales and services including architecture, development, consulting and management. Our team has developed within the Google Workspace Suite of collaborative applications and the Google Cloud Platform product base which includes popular engineering systems like Compute Engine, App Engine and Kubernetes Engine. Our team is well versed in Google Cloud and Google Workspace integrations. Our Google-centric approach provides a lean, responsive single source for collaborative and compute resource administration and management. We architect systems on Google's resources for many reasons but security is a big one for us.

Both Google Cloud and Google Workspace are FedRAMP compliant and HITRUST CSF certified further providing confidence that the processing, storing, and transmitting of sensitive data is secure. Additionally, Google's Cloud and Workspace also support HIPAA compliance across dozens of products in all regions, zones, network paths, and points of presence for Google Cloud. We are able to use Google Cloud Security Command Center and Google Cloud Monitoring to keep track of happenings within our organizations infrastructure and beyond.

Fluid Numerics leverages Google Workspace to manage core operational applications and find efficiency within automating workflows and integrating additional infrastructure in order to optimize our abilities at scale. If your organization is considering a cloud based approach for managing business operations like email, office application licensing, collaboration and sharing we are happy to share our successes with you.

Serverless architecture and Infrastructure as Code or IaC are both major components of our planning model. As a team that maintains the SELF-Fluids library and utilizes code at the core of system configuration and management we naturally rely on concepts of continuous integration, continuous development, continuous deployment or continuous delivery in order to maintain course and momentum. Serverless architecture allows resources to run as needed without the need for persistent use. At Fluid Numerics we utilize Cloud Functions and App Engine in order to assist us with these types of operations. We utilize IaC in order to deploy and maintain at our cloud resources at scale. Our working knowledge of Cloud Build, Deployment Manager, Terraform and Packer on GCP is in-depth and aided successful IaC development processes for numerous projects.