NAPLES, Fla., Sept. 18, 2020 — Adaptive Computing, a trusted leader in high-performance computing technology, announced today the release of the latest version of their cloud resource management software, NODUS Cloud OS 5.1. The enterprise-grade solution is used to automatically deploy and build clusters on any major cloud provider, automatically run workloads on those clusters in the cloud, and then destroy the clusters, assuring you only pay for what you use.
Using NODUS Cloud OS 5.1, companies can achieve a true “hybrid cloud” experience and expand their on-premise HPC resources by “bursting” their workload backlog to the cloud, eliminating long wait times in job queues and providing a better end-user experience. Studies have shown a 65% reduction in workload time-to-complete using NODUS Cloud OS 5.1. It is also ideal for compute demanding use cases requiring specialized hardware such as GPUs and larger instance sizes than are available on-premise. With NODUS Cloud OS 5.1, users can match application compute requirements with cloud compute hardware.
If you have a desktop HPC Workstation or do not have High-Performance Computing resources in your existing data center, NODUS Cloud OS 5.1 will enable you to run simulation or other HPC compute or data intensive applications in the cloud that would otherwise require an up-front investment in a complex HPC hardware installation. Provisioning the compute that is necessary can be automated in NODUS Cloud OS 5.1 without any upfront investment.
Batch jobs can run automatically with NODUS Cloud OS 5.1 by deploying the compute nodes necessary to do the work that is requested, using that compute for exactly the amount of time it is needed, and then shutting down all compute when the cloud instances are no longer in use. The cost-savings becomes significant because cloud instances are never left running.
New features in NODUS Cloud OS 5.1 enhance functionality and cloud integration, and include:
•Automation for “on-demand” cloud workloads: release a cloud assigned job, NODUS Cloud OS 5.1 will then deploy the cluster in the cloud automatically and destroy the cluster when the workload completes.
•Deploying a cluster: with NODUS Cloud OS 5.1, you can now deploy a cluster to run using either TORQUE or SLURM.
•No job scheduler needed: NODUS Cloud OS 5.1 can run workloads submitted from Moab, any other HPC workload scheduler such as PBS Pro, LSF, SLURM, etc., or without a job scheduler in place. Workloads can be routed directly to the NODUS Cloud OS 5.1 job queue and sent to instances on any major cloud provider such as AWS, Azure, Google Cloud, Oracle Cloud, or Open Telekom Cloud.
•Connect via VPN: this allows a more secure connection of on-premise clusters to the NODUS Cloud OS 5.1 clusters created in AWS, Azure, Google Cloud, Oracle Cloud, Open Telecom Cloud, or other cloud service provider.
•Advanced options: to allow the creation of clusters within AWS GovCloud and using public or private networks within all the major cloud providers.
“With NODUS Cloud OS 5.1, all enterprises and departments can achieve the nirvana of a fully automated “hybrid cloud” experience”, says Art Allen, President and CEO of Adaptive Computing Enterprises.
Source: Adaptive Computing