Home – Home – Jornals – Avtometriya 2018 number 2

In a modern experiment in the field of high-energy physics, special attention is paid to the global integration of information and computing resources into a single system for efficient storage and processing of experimental data. The ATLAS experiment conducted at the Large Hadron Collider European Organization for Nuclear Research annually produces tens of petabytes of data from the recording electronics, as well about one petabyte of data from the simulation system. For processing and storage of such super-large volumes of data, the computer model of the ATLAS experiment is based on the technology geographically distributed parallel computing, which includes the global grid infrastructure of the WLCG project (Worldwide LHC Computing Grid) and is able to meet the requirements of the experiment on processing huge data sets and provide a high degree of their accessibility (hundreds of petabytes). The paper considers the AGIS (ATLAS Grid Information System) central information system used by the ATLAS collaboration to describe the topology and resources of the computer infrastructure of the experiment, configure and connect the high-level software systems of computer centers, describe and store all possible parameters, control, configuration, and other supporting information required for the effective operation of the services of the global distributed processing system. The role of the AGIS system in the development of the concept of a general description of the resources of computing centers of grid nodes, supercomputer centers, and cloud computing resources in a single information model for the ATLAS experiment. This approach allowed the collaboration to expand the computing capabilities of the WLCG project and integrate supercomputers and cloud computing platforms into the software components of the production and distributed analysis workload management system (PanDA, ATLAS).