Storage

by Abhishek Lekshmanan (CERN)

Wednesday 17 Jul 2024, 15:00 → 16:30 Europe/Zurich

31/3-004 - IT Amphitheatre (CERN)

Description

The Storage & Data Management Group of the CERN-IT Department is responsible for the laboratory data storage activities related to LHC and non-LHC experiments and end-user community, in addition to the long term data archival, preservation and distribution of data across the Worldwide LHC Computing Grid (WLCG). The team develops and operates over a couple of exabytes of disk and tape Storage. All of this is done via software developed in house: EOS (EOS Open Storage), the disk storage system and CTA, the CERN Tape Archive. In addition the group also operates Ceph for providing block, object and file storage.

In this lecture we will look into the fundamental principles & architecture of distributed storage systems, including consistency models, fault tolerance, scalability. We will examine the design principles and tradeoffs with the real world examples of the storage systems we have at CERN. We will also look into the future trends in the horizon in order to tackle the ever exploding data volumes.

Bio

Abhishek is a Computing Engineer at CERN IT's Storage Department. During the past decade, he has worked on Software defined storage systems for the Cloud both at CERN and in the industry. Since 2021, Abhishek works on the Physics Data Services on EOS, as well as on Ceph distributed storage systems on core system features and operations.