Multi-Scale Workflow Scheduling and Energy Data for Data Centers
收藏资源简介:
This dataset provides traces for data center workload and energy management, including task-level workflows, server specifications, real-time electricity prices, and photovoltaic (PV) generation. Workloads cover three scales of job arrivals (10, 100, 1000 jobs) with tasks\u2019 resource requirements, dependencies, processing times, and deadlines. Server data describe five heterogeneous types with resource capacities and energy consumption parameters. Electricity and PV data are from CAISO, while workloads are derived from the Huawei 6GANA Network AI Challenge. This dataset supports research on topology-aware workflow scheduling and energy-aware resource allocation.
本数据集提供数据中心工作负载与能源管理相关的追踪数据,涵盖任务级工作流、服务器规格、实时电价以及光伏发电(Photovoltaic, PV)数据。工作负载涵盖三类作业到达规模,分别为10、100、1000个作业,包含任务的资源需求、依赖关系、处理时长与截止时限。服务器数据包含五类异构服务器的相关信息,涵盖其资源容量与能耗参数。电力与光伏发电数据源自CAISO(加州独立系统运营商),而工作负载数据集则衍生自华为6GANA网络AI挑战赛。本数据集可支撑拓扑感知工作流调度与能源感知资源分配方向的相关研究。



