Online Performance Monitoring of the Third ALICE Data Challenge (ADC III)

The ALICE data acquisition system has been designed for a maximum bandwidth of 2.5 GB/s for event building and of 1.25 GB/s for mass storage. In order to attain a gradual integration of the overall computing infrastructure, the present hardware components and software prototypes are tested during regular ALICE data challenges. The third one (ADC III) took place from January to March 2001 as a joint effort between the ALICE online/offine team and the CERN IT division. The main goal of this data challenge was to achieve a stable 300 MB/s throughput in the event building network and a 100 MB/s throughput to CASTOR over periods of a few days. Performance monitoring was another goal of this exercise, where a prototype (dateStat ) was developed to collect and display statistics. In this paper we will introduce this online monitoring system and report on some of the obtained results. It is structured in three parts: (1) An overview will be given on the testbed hardware, the software running on it, and the data flow. (2) The architecture of the monitoring system will be described, which consists of a set of C programs, Perl/gnuplot/CGI scripts ,and a MySQL database. It allows to measure individual/aggregate data rates,collected data volumes, and CPU loads. All these values can be visualized on web pages both on a run-by-run and global basis. (3) Various plots will be shown to illustrate the usefulness of this online monitoring system and to document the outcome of the ADC III. Finally, some ideas will be pointed out how to advance dateStat.