DSPT#93:A Multi-Cloud Infrastructure with Kubernetes,Prometheus and CNCF Project

Our #93 Webinar Meetup will be hosted by Ricardo Rocha and dedicated to the topic of "A Multi-Cloud Infrastructure with Kubernetes, Prometheus and other CNCF Projects".

The schedule for the webinar is the following:
• 18:30 - 18:45: Opening the meeting
• 18:45 - 19:25: A Multi-Cloud Infrastructure with Kubernetes, Prometheus and other CNCF Projects by Ricardo Rocha
• 19:25 - 19:50: Q&A
• 19:50: Networking (meet.google.com/pdd-uczr-fwj)

Title:
A Multi-Cloud Infrastructure with Kubernetes, Prometheus and other CNCF Projects

Abstract:
CERN is the European Organization for Particle Physics, the largest particle physics laboratory in the world. The Large Hadron Collider (LHC) and its experiments generate over 70PB of data every year, which needs to be stored and analyse using large amounts of resources.
This talk will focus on how external resources, particularly the public
clouds can be beneficial for workloads that are spiky, short lived or dependent on resources that are scarce or not available on-premises - including GPUs, TPUs or IPUs.
It will present and describe an infrastructure managing service and application deployment across multiple cloud providers. Relying on Kubernetes, Prometheus and other CNCF projects it allows our on-premise capacity to be extended with external, on-demand resources with minimal or no changes to the applications or the infrastructure itself. It will give examples of workloads covering GPU benchmarking and scaling out ML model training. Finally, it will cover the costs of using public clouds, and what can be done to optimize them.

Short-bio:
Ricardo is a Computing Engineer in the CERN cloud team focusing on containerized deployments, networking and more recently machine learning platforms. He has led for several years the internal effort to transition services and workloads to use cloud-native technologies, as well as dissemination and training efforts. Ricardo got CERN to join the CNCF and is a leader of the CNCF Research User Group, and a member of the CNCF Technical Oversight Committee. Prior to this work Ricardo helped develop the grid computing infrastructure serving the Large Hadron Collider (LHC).