
- This event has passed.
The Shift from Legacy Storage Solutions to Modern Data Platforms
Until relatively recently, HPC workloads have been characterized by the needs of the engineering and scientific community largely around modelling and simulation. Applications such as CFD, crash analysis, seismic exploration, reservoir modelling and climate simulation drove similar requirements for the systems they run on. The advent of big data and AI/ML/DL, however, has seen new requirements being placed on all elements of the HPC infrastructure, compute and storage alike. A transition is underway from legacy storage solutions to modern data platforms, not only to satisfy the traditional modeling/simulation workloads but also heterogeneously support the newer AI/ML/DL workloads.
| Speaker | Representing | Presentation | Video | PDF Download |
|---|---|---|---|---|
| All | HPC User Forum Special Event | Play Video | ||
| Mark Nossokoff | Hyperion Research | Modern Data Platforms | Download PDF | |
| Mark Nossokoff | Hyperion Research | HPC Market Update and Observations on Modern Data Platforms | Play Video | Download PDF |
| Shimon Ben-David | WekaIO | Weka Data Platforms | Play Video | Download PDF |
| Kevin Tubbs | Penguin Computing | Breaking Data Silos to Support Mixed Workloads | Play Video | Download PDF |
| James Lowey | TGen | Scale Requirements for Modern Workloads | Play Video | Download PDF |
| Panelists |
|
Perspectives on Modern Workloads | Play Video |