Oral Abstract

Birds of a Feather Discussion (B8) Giuliano Taffoni (INAF - OATs)

Science platforms: towards data science.

co-authors: G. Lemson, Z. Meyer, M.Molinaro, D. Morris, A. Schaaff

A Science Platform is a web based environment providing advanced functionality to analyze and process large and complex datasets close to the data. Science platforms allow users to bring their own customized processing to the data and facilitate sharing of results and workflows with other researchers or projects.
Building such a platform poses substantial technological challenges in the identification and staging of existing data collections for analysis;
in the identification and application of software tools and packages for data processing and bringing custom workflows to the platform; in taking advantage of the underlying HPC and HTC computing infrastructure to execute those workflows; and finally in the interoperability of different platforms, allowing tools to be shared between them transparently.
Are we ready to develop and use such astronomical science platforms? Which are the projects or experiments that my benefit from them? What platforms are already available for astronomers?
This BoF session will examine the state of the art and future perspectives involving data centers, computing centers or astrophysical projects. Sign on url:

Poster Abstract

P6.8 Giuliano Taffoni (INAF - OATs)

Theme: Local and global cloud infrastructure for processing and storage

CHIPP: INAF pilot project for HTC, HPC and HPDA

CHIPP ("Computing HTC in INAF – Pilot Project”) is an Italian project funded by the Italian Institute for Astrophysics (INAF).
The main purpose of the CHIPP project is to coordinate the use of, and access to, already existing high throughput computing and high-performance computing and data processing resources (for small/medium size programs) for the INAF community. Today, Tier-2/Tier-3 systems (1,200 CPU/core) are provided at the INAF institutes at Trieste and Catania, but in the future, the project will evolve including also other computing infrastructures.
During the last two years more than 30 programs have been approved for a total request of 30 Million CPU-h.
Most of the programs are HPC, data reduction and analysis, machine learning.
In this poster, we describe in details the CHIPP infrastructures and the results of the first two years of activity.