Job Centric Monitoring on the Grid 7 years of experience with L&B and JP services
| Authors | |
|---|---|
| Year of publication | 2008 |
| Type | Article in Proceedings |
| Conference | CESNET CONFERENCE 2008-SECURITY, MIDDLEWARE, AND VIRTUALIZATION-GLUE OF FUTURE NETWORKS |
| MU Faculty or unit | |
| Citation | |
| Field | Informatics |
| Keywords | grid; job; provenance |
| Description | Users are approaching Grids with expectations of transparency and easy to use environment. They look for simple interface to submit a job and to retrieve results of its run, eventually following changes in job state. Logging and Bookkeeping service (L&B) introduced in the EU DataGrid project 7 years ago, replaces tedious inspection of log files at individual Grid components and provides job state information through a single interface. Based on a push model, L&B is an infrastructure that collects events from all Grid components job is passing through and process them in a fault tolerant way. Users can also add annotations, to further classify their jobs. As L&B is focused on running jobs, it is complemented by the Job Provenance service (JP) that provides a long term storage and access to the job information. While L&B is for several years in production use on the largest Grid infrastructure operated by the EGEE projects and has been already deployed in other environments, JP capabilities to support scientific workflows and enhance the ways scientists can deal with a Grid had been demonstrated recently at several occasions. |
| Related projects: |