NIMBIX CEO Explores 'Software for Hybrid Cloud HPC Clusters'

NIMBIX delivers a webinar regarding visionary methods to incorporate HPC scheduling, applications and workflows in clusters with Kubernetes.

An illustration showing NIMBIX applications and workflows.

Steve Hebert, CEO & co-founder of NIMBIX, discussed the motives of NIMBIX and how the company focuses on high performance cloud computing while incorporating different applications. These applications include Artificial Intelligence (AI), Machine Learning (ML) and High-Performance Computing (HPC). During COVID, NIMBIX strives to provide remote access to HPC clusters for customers to help them access and unlock computing capacity.

“Our [NIMBIX] point of entry for cloud computing is via a “processing” API (Application Programming Interface), which defines the application and workflow and enables the platform to configure the infrastructure for you,” Hebert said.

The early years of cloud computing infrastructure differ in comparison to how it appears today. He explained the numerous challenges that occurred in traditional HPC throughout the years.

“One of the challenges was...the resources in the cloud were inadequate and difficult to set up,” Hebert said.

Scaling beyond a single cluster and increasing complexity within multiple hardware architectures are a few of the new challenges. Occasionally, replicating applications and workflows beyond a single cluster can present difficulties to domain experts who are less focused on IT and infrastructure

Over the years, the demand for computation and HPC have increased. Fast Interconnects became more accessible for companies and customers in the cloud. He indicated how next generation platforms can optimize cost and reduce friction in consuming different hybrid cloud and on premises infrastructures.

The demand is increasing for the cloud. Hebert encouraged companies to explore next generation resources, find ways to accelerate computing and drive businesses forward even though it might not be easy.

“Every type of limit in technology is typically challenged,” Hebert said.