关键词：可扩展性；灵活性；多代理集群调度
摘 要：This dissertation presents a taxonomy and evaluation of three cluster scheduling architectures for scalability and exibility using a common high level taxonomy of cluster scheduling, a Monte Carlo simulator, and a real system implementation. We begin with the popular Monolithic State Scheduling (MSS), then consider two new architectures: Dynamically Partitioned State Scheduling (DPS) and Replicated State Scheduling (RSS). We describe and evaluate DPS, which uses pessimistic concurrency control for cluster resource sharing. We then present the design, implementation, and evaluation of Mesos, a real-world DPS cluster scheduler that allows diverse cluster computing frameworks to eciently share resources.