▷ the Bright blog

We are excited to announce that Bright Computing has been invited to participate in the next Dell Technologies HPC Community Event:

The HPC industry is marching towards the delivery of the first exascale systems, composed of tens-of-thousands or hundreds-of-thousands of server nodes and exotic software stacks to manage it all. While the first systems will have a limited audience, exascale systems will increasingly move down-market to reach large numbers of enterprises. As exascale systems make their way into commercial enterprises, the need for more standardized, out-of-the-box management tools will follow.   

HPC is undergoing massive changes that present both challenges and opportunities for organizations across industries. Historically, HPC has been defined by a consistent and well-known set of applications for modeling and simulation running on bare metal servers in an organization’s data center—but that is changing. We now have machine learning and data analytics applications vying for access to trusted HPC systems, ultimately throttling important work in one place or another.

ICYMI, Bright recently announced that Bright Cluster Manager was selected to support the new artificial intelligence supercomputer to be installed at National Supercomputer Centre (NSC) at Linköping University.

Earlier this week, Bright CEO Bill Wagner wrote an article featured by InsideHPC where he discusses how changes in HPC are demanding that businesses unify their HPC infrastructure from edge to core to cloud. He goes on to explain that, at the edge, while sensors and new apps that require low-latency compute certainly necessitate systems that live in close proximity (outside of public clouds and traditional data centers), the reality is, new compute-intensive/data-intensive apps in machine learning and analytics are driving the need for organizations to have a growing number of high-performance clusters in an expanding number of physical locations, as well as in the public cloud. With compute and systems destined to become increasingly decentralized in this way, the notion of the edge being an outlier to some dominant central computing infrastructure overlord will quickly become dated.

Did you see that Bright Cluster Manager 9.1 introduces support for Redfish BIOS management? This feature enables administrators to view and update BIOS parameters on any/all nodes of the cluster that support the Redfish specification from the command line or Bright View. This reduces complexity and the effort that admins face when updating BIOS settings across their cluster.

Bright Cluster Manager 9.1 includes an integration with Jupyter Notebook that accomplishes several important things that make Jupyter a more effective and powerful tool for users:

Ever since I joined Bright back in early 2014, I have been delighted with the level of collaboration we have with NVIDIA; in fact, it was not long after joining that we announced Day Zero support for NVIDIA’s latest K80 GPU. Fast forward a small number of years and many hours of Bright engineering investment on NVIDIA technologies, and I am once again pleased that Bright is now providing the cluster management component for NVIDIA and NetApp’s latest turnkey AI solution.

Bright Cluster Manager includes a feature called Auto Scaler that dynamically scales HPC workload management (WLM) and Kubernetes clusters based on workload demand, subject to configured policies. Auto Scaler is the engine that drives the hybrid cloud. It can repurpose and/or re-image on-prem and edge nodes and it can automatically extend the cluster to the cloud. 

I’m pleased to say that now Bright provides cluster management for OpenShift, organizations can manage their OpenShift infrastructure through Bright, unlocking many Bright benefits such as gaining the increased capabilities, flexibility, and extensibility.