This work presents the implementation of a midsize Data Center devoted to Physics data analysis, realized in 2023-24 at the University of Naples, as the second upgrade to a previous Data Center built on 2007 during SCoPE project. Data Center had a first upgrade, during RECAS project, in 2014. After PNRR projects (STILES, CN1), the Data Center handles about 100 servers, 10.000 cores, 100 GPUs, 10 Pbytes of storage, and is used for big data analysis in astrophysics, high energy physics, material science and theoretical physics. The purpose of this work is to illustrate and demonstrate how it is possible to effectively and efficiently perform physical and astrophysical analyses in the mid-size Data Center, using Big Data, on unstructured datasets
Big-Data-Oriented Computing on a Mid-Size Data Center: the ADHOC Infrastructure / Tortora, Augusto; Brescia, Massimo; Conte, Stefania; Guida, Paolino; Russo, Guido; Spisso, Bernardino. - (2024), pp. 1-7. (Intervento presentato al convegno FUTURE COMPUTING 2024 : The Sixteenth International Conference on Future Computational Technologies and Applications tenutosi a Venice nel 14-18 aprile 2024).
Big-Data-Oriented Computing on a Mid-Size Data Center: the ADHOC Infrastructure
Augusto Tortora
;Massimo Brescia;Stefania Conte;Guido Russo;Bernardino Spisso
2024
Abstract
This work presents the implementation of a midsize Data Center devoted to Physics data analysis, realized in 2023-24 at the University of Naples, as the second upgrade to a previous Data Center built on 2007 during SCoPE project. Data Center had a first upgrade, during RECAS project, in 2014. After PNRR projects (STILES, CN1), the Data Center handles about 100 servers, 10.000 cores, 100 GPUs, 10 Pbytes of storage, and is used for big data analysis in astrophysics, high energy physics, material science and theoretical physics. The purpose of this work is to illustrate and demonstrate how it is possible to effectively and efficiently perform physical and astrophysical analyses in the mid-size Data Center, using Big Data, on unstructured datasetsI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.