Abstract

Presently, big companies such as Amazon, Google, Facebook, Microsoft, and Yahoo! own huge datacenters with thousands of nodes. These clusters are used simultaneously by many clients. The users submit jobs containing one or more tasks. The task flow is usually a mix of short, long, interactive, and batch tasks with different priorities. The cluster scheduler decides on which server the task should be run as a process, container, or virtual machine. Scheduler optimizations are important as they provide higher server utilization, lower latency, improved load balancing, and fault tolerance. Optimal task placement is a complex problem that has multiple dimensions and requires algorithmically complex optimizations. This increases placement latency and limits cluster scalability. In this paper, we consider different cluster scheduler architectures and optimization problems.

中文翻译：

---

计算集群中的有效调度

摘要

目前，亚马逊，谷歌，Facebook，微软和Yahoo!等大公司 拥有拥有数千个节点的巨大数据中心。这些群集被许多客户端同时使用。用户提交包含一个或多个任务的作业。任务流通常是具有不同优先级的短期，长期，交互式和批处理任务的混合。群集调度程序决定应在哪个服务器上将任务作为进程，容器或虚拟机运行。调度程序优化很重要，因为它们可以提高服务器利用率，降低延迟，改善负载平衡和容错能力。最佳任务放置是一个复杂的问题，具有多个维度，并且需要算法上复杂的优化。这会增加放置延迟并限制群集的可伸缩性。在本文中，