Nimbix Launches JARVICE 3.0 HPC Hybrid Cloud Platform with Kubernetes

Print Friendly, PDF & Email

Today cloud provider Nimbix announced the Technology Preview of JARVICE 3.0, a robust HPC hybrid cloud software platform fully integrated with Kubernetes. The new container platform is designed to provide the enterprise with access to the entire Nimbix self-service catalog of HPC, AI and machine/deep learning applications. Packaged as a new enterprise software tool, JARVICE 3.0 optimally runs supercomputing and HPC workloads on-premises or remotely on the public cloud.

The concept of using containers to run cloud workloads has been in practice at Nimbix since our inception, so naturally as Kubernetes has gained traction in the enterprise for web services, we wanted to extend the industry-leading compute capabilities of JARVICE for HPC and AI into an infrastructure platform like Kubernetes,” explains Nimbix CEO, Steve Hebert. “We’ve been working to democratize bare-metal cloud supercomputing technology over the last several years. With JARVICE 3.0, our customers can bring Nimbix technology into their own data centers to support their fast-emerging
advanced computing workflows.”

JARVICE 3.0 is designed to provide users with instant access to HPC workflows as well as open-sourced applications in a click-to-run catalog. Upon installation of JARVICE 3.0, customers will have access to the entire Nimbix suite of containerized HPC applications as well as the ability to install their own platform apps in a modern continuous integration and deployment methodology. This new solution runs on heterogeneous machines, operating on any CPU, GPU, FPGA, any fabric (ethernet or InfiniBand), bare-metal or virtualized systems and clusters. Users may also deploy JARVICE 3.0 on their private
infrastructure or in the public cloud. JARVICE 3.0 is innovating the hybrid cloud space by establishing a true infrastructure agnostic platform for HPC anywhere.

JARVICE 3.0 will target organizations who are standardizing and modernizing their data centers with containerized applications while retaining the option to burst to the public cloud. This complete platform will provide customers with a hybrid software solution that will run large scale HPC and AI
workflows side by side with existing web service and container-native software, delivering the same seamless user experience of the Nimbix Cloud. As situations occur where compute resources are unavailable on-premises, users can then use the unified processing API to burst to the Nimbix Cloud with
a simple click of the mouse. This provides the ability to manage a private cloud infrastructure on-premises while simultaneously having the ability to extend compute to the Nimbix Cloud, taking advantage of a software platform that will intuitively run HPC workflows across a diverse resource base.

The JARVICE 3.0 service catalog will be delivered as a set of containerized turn-key workflows, ready to run at scale on existing Kubernetes infrastructure using Nimbix’s advanced cloud technology. In this model, Kubernetes acts as the underlying infrastructure layer while JARVICE 3.0 optimizes large-scale applications and delivers them with point and click ease. This includes tightly coupled HPC solvers and distributed deep learning algorithms, utilizing advanced resources such as low latency fabrics and computational accelerators.

According to the newest HPC market data from Intersect360 Research, 2017 was a breakout year for the growth of HPC in the public cloud. Cloud-based spending exceeded $1 Billion, with an increase of 44% over 2016. The sudden expansion was concentrated in application hosting (software-as-a-service, SaaS), which grew 125% year-over-year. The forecast from Intersect360 Research projects continued high growth through 2018 and 2019, with spending on HPC in public cloud eventually reaching nearly $3 Billion by 2022.

We have been anticipating double-digit growth in cloud adoption for HPC for multiple years, and 2017 was the year it finally took off,” said Addison Snell, CEO of Intersect360 Research. “The biggest challenge for an enterprise is managing an expanding set of high-performance workloads, including analytics and deep learning, against a diversifying set of technologies, such as new processors, storage, and interconnect architectures, while budgets have stayed near the same level. The container-based cloud approach in the JARVICE 3.0 platform gives users an option for scaling these workloads while managing both the cost of deployment and the risk of new technologies.”

In related news, Nimbix also announced that the United States Patent and Trademark Office has issued US Patent 9,973,566, entitled “Dynamic Creation and Execution of Containerized Applications in Cloud Computing”. The technology describes the methods of using containers to run HPC workloads and workflows in the cloud, which takes containers as the packaging mechanism, deploys those for the computational or processing path and then destroys the containers once that processing is complete. That environment is then recreated within a given container and is ready for execution as the next
secure, containerized workload.

Leo Reiter, CTO, Nimbix

We first brought dynamic containerization of large-scale applications to market in 2013”, explains Nimbix CTO, Leo Reiter. “This novel technology increases efficiency, improving time-to-results for all HPC workloads. Our newly issued patent is recognition for years of innovation and operational experience with containers at scale, on bare metal, and we are excited to continue to push the envelope with ever-evolving technologies in this space.”

Sign up for our insideHPC Newsletter