作者: Ronnie Chaiken , Bikas Saha , James David Ryseff
DOI:
关键词:
摘要: Systems, methods, and computer media for scheduling vertices in a distributed data processing network allocating computing resources on node are provided. Vertices, subparts of job including both code that runs the data, assigned by manager to cluster process nodes processing. The run transmit resource usage information, memory core usage, back manager. uses this information estimate other either still running or waiting be run. Using estimated each can multiple concurrently.