Mixed Workload Scheduling (混部) in Kubernetes: Challenges, Core Technologies, and Koordinator Enhancements

The article analyzes low CPU utilization in pure online Kubernetes clusters, introduces mixed‑workload (online + offline) scheduling to improve resource efficiency, explains core techniques, kernel QoS features, and details Koordinator‑based implementations such as node resource reservation and scheduling adjustments.

Cloud NativeKoordinatorKubernetes

0 likes · 13 min read

Mixed Workload Scheduling (混部) in Kubernetes: Challenges, Core Technologies, and Koordinator Enhancements

iQIYI Technical Product Team

Nov 17, 2023 · Big Data

Mixed Workload Co-location of Big Data and Online Services at iQIYI: Design, Implementation, and Results

iQIYI’s mixed‑workload system colocates Spark/Hive big‑data jobs with online video services by running YARN NodeManagers inside Kubernetes, using an Elastic YARN Operator, Koordinator‑driven CPU oversubscription, and remote shuffle, boosting online CPU utilization from ~9 % to over 40 % and saving tens of millions of RMB annually.

Big DataCloud NativeKubernetes

0 likes · 19 min read

Mixed Workload Co-location of Big Data and Online Services at iQIYI: Design, Implementation, and Results

Alibaba Cloud Native

Nov 4, 2022 · Cloud Native

How Koordinator Revolutionizes Cloud‑Native Mixed‑Workload Scheduling

Koordinator, an open‑source cloud‑native scheduler launched in April 2022, unifies heterogeneous workloads on Kubernetes through zero‑intrusion plugins, fine‑grained resource oversubscription, QoS‑aware scheduling, and a flexible descheduler framework, dramatically improving resource utilization and latency‑sensitive service performance.

Cloud NativeKubernetesMixed Workload

0 likes · 22 min read

How Koordinator Revolutionizes Cloud‑Native Mixed‑Workload Scheduling

HomeTech

Jan 13, 2022 · Cloud Native

AutoKH: A Mixed‑Workload Resource Management Solution on Kubernetes and Hadoop

AutoKH is a cloud‑native mixed‑workload framework that integrates Kubernetes and Hadoop to dynamically schedule online and offline tasks, improve CPU and memory utilization, enforce priority classes, and ensure service stability through operators, CronHPA, and resource‑control components.

CPU ManagerHadoopKubernetes

0 likes · 19 min read

AutoKH: A Mixed‑Workload Resource Management Solution on Kubernetes and Hadoop

Baidu Geek Talk

Jan 5, 2022 · Cloud Native

Baidu Cloud‑Native Mixed Workload (Offline Co‑location) Technology Overview

Baidu’s mixed‑workload approach co‑locates offline batch jobs with latency‑sensitive online services on shared nodes, using a dynamic resource view, priority‑based scheduling, cpuset/NUMA isolation, eBPF policies, and predictive profiling, boosting CPU utilization above 40 % and saving billions of RMB in total cost of ownership.

KubernetesMixed Workloadcloud-native

0 likes · 17 min read

Baidu Cloud‑Native Mixed Workload (Offline Co‑location) Technology Overview