"Data center workload dynamic scheduling dataset"
收藏资源简介:
"This dataset provides a comprehensive resource for research in cloud computing, workload scheduling, and energy-aware resource management. It includes:\u200b\u200bWorkload traces\u200b\u200b (two scales: large and small), derived from the Alibaba Cluster Trace 2020, capturing task arrival times, resource demands (CPU, memory), and deadlines after rigorous cleaning and filtering.\u200b\u200bServer configurations\u200b\u200b (two scales: large and small), specifying hardware performance parameters (e.g., CPU cores, memory, energy profiles) to model heterogeneous infrastructures.\u200b\u200bTime-varying electricity prices\u200b\u200b at 15-minute granularity, sourced from CAISO, to enable cost and energy efficiency analysis.The dataset supports studies on dynamic workload scheduling, server provisioning, and demand-response optimization in data centers. The inclusion of multi-scale workload and server data ensures flexibility for simulating diverse scenarios, from small-scale experiments to large-scale cluster analyses."
本数据集为云计算(Cloud Computing)、工作负载调度(Workload Scheduling)以及能源感知资源管理(Energy-aware Resource Management)领域的研究提供了全面的支撑资源。其包含以下内容: 工作负载跟踪数据(Workload traces):分为大、小两种规模,源自阿里巴巴集群跟踪数据集2020版(Alibaba Cluster Trace 2020),经过严格的清洗与筛选流程后,涵盖任务到达时间、资源需求(CPU、内存)以及任务截止时限等核心信息。 服务器配置数据(Server configurations):同样包含大、小两种规模,明确了异构基础设施建模所需的硬件性能参数(如CPU核心数、内存容量、能源特性曲线)。 时变电价(Time-varying electricity prices):粒度为15分钟,数据源自加州独立系统运营商(CAISO),可用于开展成本与能源效率分析。 本数据集可支撑数据中心内动态工作负载调度、服务器资源配置以及需求响应优化等相关研究。多尺度的工作负载与服务器数据,能够灵活模拟从小型实验到大型集群分析的各类应用场景。



