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


ClusterVision Trinity Private Cloud Moves to UberCloud HPC Application Containers

trinityToday ClusterVision and UberCloud formally announced the integration of UberCloud Application Containers with Trinity, ClusterVision’s private cloud solution. Trinity provides a complete HPC cluster manager on top of a cloud platform capable of managing HPC, general IT, and Big Data, bringing the flexibility of Platform-as-a-Service to HPC.

During the Docker User Group at CIUK 2015, ClusterVision and UberCloud demonstrated Trinity by managing and running UberCloud application containers in a virtual HPC cluster environment. Real-time access and use of an UberCloud container were shown loaded with a complete virtual desktop, a fluid dynamics application based on OpenFOAM software, and the application data, and resulting data analyzed interactively through remote visualization.

“Docker presents a new and exciting paradigm for building and deploying ICT services,” said Hans Then, lead developer at ClusterVision. “Many within the HPC community have looked at Docker as a lightweight virtualization method without the performance loss associated with Virtual Machines. Trinity was designed with this in mind. However, Docker is also a way to simplify the build, versioning, distribution, and management of HPC applications. It obviates the need to compile your HPC applications. This is one of the strengths of the UberCloud application containers. Combining the two allows HPC sites to offer a standard set of HPC applications which can be easily and securely run on any Trinity HPC cluster.”

Over the past several years, ClusterVision has developed Trinity, a unified management environment for both Cloud and HPC. Trinity was a response to customer requests for OpenStack to manage their HPC clusters without any loss of performance. Docker containers enable the lightest weight virtualization for a private HPC cloud that matches bare metal performance. Now, Trinity is able to allocate the best time slot to start UberCloud application containers, and to run distributed instances with MPI over an Infiniband layer.

At UberCloud we are convinced that production-ready HPC application containers are how software will be distributed and used in the near future,” said Burak Yenier, CEO of UberCloud, “As tools like ClusterVision Trinity for managing containers mature, containerization of sophisticated HPC application software will become more and more popular. UberCloud’s contribution to this trend will be in offering containerized application images for engineering and scientific applications in HPC, Big Data Analytics, IoT, and other compute and data intensive areas. This will be a game changer for the users; they no longer would worry about creating a cluster (taken over by Trinity) and they no longer would worry about installing, configuring, managing their s/w applications (taken over by UberCloud containers). UberCloud would also be in a position to provide professional services for customers who want to containerize applications.”

Sign up for our insideHPC Newsletter

Resource Links: