Abstract

The present disclosure relates to a system and a method of locking and ranking the resource for utilization in a distributed scheduling system. The method comprises prioritizing the jobs from the job pool based on the job properties and allocating the jobs to the scheduler. Before allocating the jobs to the scheduler, if the jobs are consumed, then a lock is added to the jobs based on state management. The jobs may be assigned a lock state based on the time of their access by the schedulers. The scheduler is configured to pick up the jobs which are not in lock state. Thereafter, the set of jobs from the list selected by the job scheduler is ranked using a rank function. The ranking may be re-evaluated, if one or more jobs are added within the time and reflected in the new job pool. Ranking of jobs ensures that there is unbiased throttling of the jobs. Further, the job scheduler allocates the job on the resource based on the resource availability. The advantages of the present disclosure are that is helps in prioritizing the jobs based on their properties, and scheduling the jobs on a job scheduler by avoiding scheduler starvation and race-condition among the schedulers.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS