Orateur
Maya Stavrianakou
(NOA/NESTOR)
Description
The KM3Net experiment data arrive on shore where they are processed in real time at a Data Filter Farm and are subsequently stored and backed up at a central computing centre located on site. From there we propose a system whereby the data are distributed to participating institutes equipped with large computing centres for further processing, duplication and distribution to smaller centres. The data taking site hosts the central data management services, including the database servers, bookkeeping systems and file catalogue services, the data access and file transfer systems, data quality monitoring systems and transaction monitoring daemons and is equipped with fast network connection to all large computing sites. Data and service challenges in the course of the preparatory phase must be anticipated in order to test the hardware and software components in terms of robustness and performance, scalability as well as modularity and replaceability, given the rapid evolution of the market both in terms of CPU performance and storage capacity. The role of the GRID would also have to be evaluated and the appropriate implementation selected in time for an eventual test in the context of a data challenge before the start of data taking.
Auteur principal
Maya Stavrianakou
(NOA/NESTOR)
Co-auteur
Tasos Belias
(NOA/NESTOR)
