作者: Subrata Mitra , Shanka Subhra Mondal , Nikhil Sheoran , Neeraj Dhake , Ravinder Nehra
关键词:
摘要: Large multi-tenant production clusters often have to handle a variety of jobs and applications with complex resource usage characteristics. It is non-trivial non-optimal manually create placement rules for scheduling that would decide which should co-locate. In this paper, we present DeepPlace, scheduler learns exploits various temporal patterns using Deep Reinforcement Learning (Deep RL) reduce competition across running in the same machine while at time optimizing overall cluster utilization.