RDIG monitoring

The Russian RDIG consortium (Russian Grid for Intensive Operations with the Data – Russian Data Intensive Grid, RDIG) was organized in 2003. Its basic purpose was the creation of a grid-infrastructure for intensive operations with scientific data. This infrastructure was necessary for support of participating Russian scientists in experiments in high-energy physics, biology, chemical physics, earth sciences, etc. By the end of the EGEE project (May, 2010), the consortium included 15 resource centers (the so-called grid-sites) from Russia, and one Belarus and one Ukrainian centers.

Alongside with an essential geographical distribution, RDIG possessed considerable resources: more than 4000 computing slots and about 2 Petabytes of data storage. The consortium infrastructure supported work of more than forty virtual organizations. Tools of monitoring and the accounting of use of resources have been developed at JINR for tracing the state of this scale system. A specialized package for grid-monitoring has been developed and installed in all the resource centers within RDIG. For the distributed data gathering of monitoring, the package MonALISA (http://monalisa.cacr.caltech.edu) was used as a basis. All information was accessible on the site of RDIG monitoring (http://rocmon.jinr.ru). It displayed both a current state of the system and the history of change of the most important components.

As the most significant parameters of functioning the grid-system and state of jobs of the virtual organizations, the following indicators have been chosen:

  • volume of the resources accessible to the virtual organizations in the resource centers;
  • number of computing slots and their performance;
  • disk storage volume;
  • current use of an infrastructure;
  • quantity of the jobs submitted by the virtual organizations (the number both running and  waiting);
  • volumes of the data, which are kept by the virtual organizations in each of the resource centers and the grid-infrastructure as a whole;
  • loading of the Computing Elements which are carrying out distribution users’ jobs on Working Nodes;
  • for the experiment — the use of storage resources by various groups of users in the virtual organization.

Since the 2011 year, i.e. after completion of the EGEE project, only the archive (2004-2010 years) of the statistics of using the RDIG computing resources by the virtual organizations is accessible on the monitoring site (http://rocmon.jinr.ru:8080/accounting/). From the moment of the organization of the RDIG consortium to the end of 2010 it processed already over 26 million jobs, and more than 30 million hours of pure processor time was spent for it. The most part of the resources was consumed by the virtual organizations connected with the Large Hadron Collider.

At present full information on the use of resources both individual RDIG sites and the entire consortium can be obtained at the accounting portal of the EGI project (http://accounting.egi.eu/country.php).

Comments are closed.