CloudyCluster by Omnibond is a comprehensive solution that enables users to create and manage High Performance Computing and High Throughput Computing environments seamlessly on Google Cloud Platform and Amazon Web Services . It offers a familiar HPC environment while leveraging the scalability and flexibility of cloud resources, allowing users to configure jobs with various instance types, including GPU and preemptible instances, and customize memory and CPU configurations to meet specific computational needs.
Key Features and Functionality:
- Automated Cluster Deployment: Quickly set up a fully operational and secure computation cluster within minutes, complete with encrypted storage, compute resources, and HPC schedulers like Torque or SLURM integrated with the CCQ Meta-Scheduler.
- Interactive Research Computing: Utilize a graphical user interface developed in collaboration with the Ohio Supercomputer Center, providing non-computer scientists access to cloud-based HPC tools without the need for command-line interfaces. Features include file management, job script drafting, and launching computing instances with or without GPU acceleration.
- Scalable Storage Solutions: Leverage a variety of storage technologies available through GCP and AWS, configuring data to reside in different storage classes based on age or access frequency, and pulling necessary data to High Performance Parallel Storage for computation.
- Elastic Scaling: Dynamically scale computational resources, leveraging millions of virtual CPUs as needed, ensuring cost-effective and efficient processing for large-scale workloads.
- Security and Compliance: Deploy clusters within a Virtual Private Cloud , utilizing encrypted storage and following industry best practices for security, including multi-factor authentication and SSL certificates.
Primary Value and User Solutions:
CloudyCluster democratizes access to HPC resources by simplifying the creation and management of cloud-based computational environments. It addresses the challenges of setting up and maintaining on-premises HPC infrastructure by providing an on-demand, scalable, and cost-effective solution. Researchers and organizations can focus on their computational tasks without the overhead of hardware management, benefiting from the elasticity of the cloud to handle varying workloads efficiently. This leads to reduced time to discovery and innovation, as users can rapidly deploy and scale resources according to their project requirements.