Rutgers University Capstone Automated Task Scheduler
RUCATS is a web-based resource monitoring and job submission tool which includes prediction models. This tool will allow users to monitor the cluster utilization (CPU, network, I/O, etc.) and customize resource access privileges for various classes of users.
Automated job schedulers have been extensively studied and implemented in large clusters and supercomputers. However, many of these clusters do not provive detailed utilization information which users can take advantage of. For example, users working to meet deadlines can utilize the cluster monitoring to make appropriate decisions regarding application deployment. Using a web-based interface will allow users easy access to the cluster monitoring system and job scheduler. One advantage of using the web, is thatit provides systems administrators to customize the access priviledges for individual users or groups of users. Another consequence of implementing both monitoring and scheduling is the ability to generate prediction models. The models would enable the system to react effectively to jobs currently running and the current cluster utilization.
Project Website: https://sites.google.com/site/rutgerscats/