Sign up for our newsletter and get the latest HPC news and analysis.
Send me information from insideHPC:


CloudyCluster Moves HPC out of the Data Center and Into the Cloud

Sponsored Post

Increasingly, the resources needed for HPC and Big Data computing are emerging in the cloud. The problem is that setting up and running a complex workflow using a remote computing service can often be difficult, time consuming and dependent on expert IT intervention. Omnibond, located in South Carolina near Clemson University, has come up with an answer. The company is already well known for its OrangeFS file system, which supports the high availability, parallel storage, retrieval and management of large data sets running on HPC clusters.

Introducing CloudyCluster

Now Omnibond has introduced CloudyCluster, a new solution that works with OrangeFS to make moving HPC into the cloud even simpler. CloudyCluster allows you to quickly set up and configure a cluster on Amazon Web Services (AWS) to handle the most demanding HPC and Big Data tasks. You don’t need access to a data center and you don’t have to be an expert in the ins and outs of running computationally intensive workloads in a cloud environment.

cloudy clusterBy leveraging the power of OrangeFS, projects that typically required many multicore machines and storage nodes can now be handled in the cloud as part of CloudyCluster. It is designed to handle complex engineering and scientific tasks such as modeling and simulation in manufacturing, or next generation sequencing for life sciences applications. Because it is an object-based Posix compatible file system, OrangeFS easily handles the high level of performance demanded by HPC and Big Data applications

CloudyCluster acts like a “mobile, web-driven data center in the cloud,” as the engineers at Omnibond like to say. It provides an easy-to-use, web-based interface for desktop, tablet and mobile computing devices – replacing the need for a physical data center with cloud-based resources. And unlike most data center models, CloudyCluster and AWS are available around the clock, all day, every day.

The CloudyCluster step-by-step wizard makes setup a straightforward, painless process in your office or on the road. You work with a series of simple options, and the Omnibond solution handles all the details necessary to set up a working cluster based on your specifications.

CloudyCluster also provides simplified management of HPC resources and applications for the cloud with support for schedulers such as Hadoop YARN, Condor or Torque. WebDAV provides native file access. Also included are multiple sets of instances and images associated with popular Open Source HPC software.

Supporting Collaboration
CloudyCluster was designed to optimize collaboration, providing you with easy-to-use tools to add users and grant access permissions to share your AWS compute resources.
Simple options allow you to dynamically provision shared Google groups and folders and box.com folders for enhanced collaboration.

You quickly and easily create your own CloudyCluster Virtual Private Cloud (VPC), which integrates OrangeFS. This combination of the two Omnibond solutions provides you with parallel access to disk storage for applications running in an HPC environment, solving the thorny storage access issues that often accompany Big Data and compute-intensive applications.

OrangeFS provides highly available scratch or working storage. If an OrangeFS instance fails, another instance that is currently running takes its place. The replacement instance automatically re-attaches to the appropriate EBS volumes and continues the operation. You define the parameters for the high availability scratch storage by selecting its size and number of fail over instances.

CloudyCluster provides compute instances with an extensive set of open HPC applications, dynamically configured for fast and comprehensive access to storage. Within a compute group, all instances are based on the same instance types. You control the compute groups – each one can have a different instance type and a different image. You can select a compute scheduler for the entire cluster or a specific compute group. Amazon’s DynamoDB stores the operational meta-data to support the on-demand infrastructure.

Also, the advanced computing infrastructure you have created using the CloudyCluster allows you to compute, pause, resume and delete compute groups at will, maximizing the efficiency of your ASW instance and lowering your costs – you pay only for the actual run time.

Service and Support
CloudyCluster receives the same high level of commercial grade support and services that Omnibond provides for OrangeFS. The company’s highly experienced team of professional software engineers fully supports both CloudyCluster and OrangeFS users, solving problems and providing design, performance and development services. Users also have access to comprehensive CloudyCluster documentation and Amazon Web Services documentation.

Become a Beta Tester

CloudyCluster will be moving into beta testing in the very near future. If you would like to participate, fill out the Beta Request form on the Omnibond CloudyCluster web site by clicking here.

Omnibond is understandably enthusiastic about the combined capabilities of CloudyCluster and OrangeFS, calling it “one of the most efficient and easy to use on-demand HPC cluster solutions available in the cloud.”

CloudyCluster offers a solution that scales to fit your HPC and Big Data needs without the capital investment required by more conventional platforms. With CloudyCluster, you can forget about building an HPC infrastructure. Instead, you can concentrate on solving business and research problems by focusing on your core competencies and leveraging the power of HPC on demand without the hassle.

Resource Links: