Datacenter operations today provide a plethora of new queueing and scheduling problems. The notion of a “job” has become more general and multi-dimensional. The ways in which jobs and servers can interact have grown in complexity, involving parallelism, speedup functions, precedence constraints, and task graphs. The workloads are vastly more variable and more heavy-tailed. Even the performance metrics of interest are broader than in the past, with multi-dimensional service-level objectives in terms of tail probabilities. The purpose of this article is to expose queueing theorists to new models, while providing suggestions for many specific open problems of interest, as well as some insights into their potential solution.

当今的数据中心运营提供了许多新的排队和调度问题。“工作”的概念已变得更加普遍和多维。作业和服务器之间进行交互的方式越来越复杂，涉及并行性，加速功能，优先级约束和任务图。工作负载的可变性更大，并且要求更严格。甚至感兴趣的性能指标也比过去更宽泛，在尾部概率方面具有多维服务级别目标。本文的目的是向排队论者介绍新模型，同时为许多感兴趣的特定开放问题提供建议，并对他们的潜在解决方案提供一些见解。