Optimizing Job Scheduling in National Grid Computing System: Theory and Practice
| Authors | |
|---|---|
| Year of publication | 2015 |
| Type | Article in Proceedings |
| Conference | MEMICS |
| MU Faculty or unit | |
| Citation | |
| Field | Informatics |
| Keywords | Grid; Performance; Simulation; Job scheduling; Workload |
| Description | The work presented explains that the problem of optimizing the performance of an actual production scheduler is far more complex than just choosing a proper scheduling algorithm. Using our experience from the Czech National Grid Infrastructure MetaCentrum, we show that production systems need to balance around a set of policies that are established to satisfy both resource providers and users, e.g., a fairsharing policy guaranteeing fair resource allocation, as well as proper queue setup, guaranteeing reasonable distribution of system resources among various job classes. Most importantly, our work provides both the theoretical analysis covering impact of various system policies as well as the practical evaluation using data gathered from the real production system. |
| Related projects: |