Invited Paper: Improving Data Center Efficiency Through Holistic Scheduling In Kubernetes

Townend, Paul; Clement, Stephen; Burdett, Dan; Yang, Renyu; Shaw, Joseph A.; Slater, Brad; Xu, Jie

doi:10.1109/sose.2019.00030

Cited by 36 publications

(16 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Finally, another trend worth-mentioning is the need to reduce energy consumption in data centers, with multiple vendors such as Amazon AWS [4], Google [45], Microsoft [78] committing eforts in this direction. As a result, diferent works such as [50,52,89,106] have been proposed to address the energy eiciency and carbon footprint concerns.…”

Section: The State Of the Artmentioning

confidence: 99%

Custom Scheduling in Kubernetes: A Survey on Common Problems and Solution Approaches

Rejiba

Chamanara

2022

ACM Comput. Surv.

View full text Add to dashboard Cite

Since its release in 2014, Kubernetes has become a popular choice for orchestrating containerized workloads at scale. In order to determine the most appropriate node to host a given workload, Kubernetes makes use of a scheduler that takes into account a set of hard and soft constraints defined by workload owners and cluster administrators. Despite being highly configurable, the default Kubernetes scheduler cannot fully meet the requirements of emerging applications, such as machine/deep learning workloads and edge computing applications. This has led to different proposals of custom Kubernetes schedulers that focus on addressing the requirements of the aforementioned applications. Since the related literature is growing in this area, we aimed, in this survey, to provide a classification of the related literature based on multiple criteria, including scheduling objectives as well as the types of considered workloads and environments. Additionally, we provide an overview of the main approaches that have been adopted to achieve each objective. Finally, we highlight a set of gaps that could be leveraged by the academia or industry to drive further research and development activities in the area of custom scheduling in Kubernetes.

show abstract

Section: The State Of the Artmentioning

confidence: 99%

Custom Scheduling in Kubernetes: A Survey on Common Problems and Solution Approaches

Rejiba

Chamanara

2022

ACM Comput. Surv.

View full text Add to dashboard Cite

show abstract

“…Expanding beyond the hardware level, Townend et al [14] presented a scheduler that takes multiple levels of infrastructure into account when distributing containers on a Kubernetes cluster while Baek et al [2] use RL to balance workloads over heterogeneous servers while optimizing throughput and energy usage.…”

Section: Related Workmentioning

confidence: 99%

Towards a Holistic Controller

Heimerson

Brännvall

Sjölund

et al. 2021

Proceedings of the Twelfth ACM International Conference on Future Energy Systems

View full text Add to dashboard Cite

The increased use of cloud and other large scale datacenter IT services and the associated power usage has put the spotlight on more energy-efficient datacenter management. In this paper, a simple model was developed to represent the heat rejection system and energy usage in a small DC setup. The model was then controlled by a reinforcement learning agent that handles both the load balancing of the IT workload, as well as cooling system setpoints. The main contribution is the holistic approach to datacenter control where both facility metrics, IT hardware metric and cloud service logs are used as inputs. The application of reinforcement learning in the proposed holistic setup is feasible and achieves results that outperform standard algorithms. The paper presents both the simplified DC model and the reinforcement learning agent in detail and discusses how this work can be extended towards a richer datacenter model. CCS CONCEPTS• Computer systems organization → Sensors and actuators; • Hardware → Enterprise level and data centers power issues; Temperature control; • Computing methodologies → Reinforcement learning; Modeling and simulation.

show abstract

“…Townend et al [29] propose a solution that coordinates software and hardware isolation mechanisms to improve the efficiency of data centers and implement their own scheduler based on Kubernetes. However, its focus is on energy saving, and it does not consider the services SLOs.…”

Section: ) Production Cluster Analysismentioning

confidence: 99%

Zeus: Improving Resource Efficiency via Workload Colocation for Massive Kubernetes Clusters

Zhang

Wang

et al. 2021

IEEE Access

View full text Add to dashboard Cite

With the popularity of container-based microservices and cloud-native architectures, Kubernetes has established itself as the de facto standard for container orchestration. Kubernetes is known for its advantage in easy deployment and operations for applications; however, it suffers from low resource utilization, incurring high server provisioning and operational costs, due to following reasons. First, it is common practice for latency-sensitive services to be over-provisioned for the peak load: Kubernetes might consider such peak-load provisioning as constant, even though the actual resource utilization is low. Second, the isolation between different containers is poor and cannot prevent performance degradation when besteffort jobs and latency-sensitive services run together. Users may have to run these two classes of workloads on separate machines to avoid interference, resulting in higher provisioning costs. This paper presents a highly scalable cluster scheduling system named Zeus, which is designed based on Kubernetes extension mechanisms. Zeus achieves safe colocation of best-effort jobs and latency-sensitive services. Furthermore, Zeus schedules best-effort jobs based on the real server utilization and can adaptively allocate resources between the two classes of workloads. In addition, Zeus enhances container isolation by coordinating software and hardware isolation features. As a result, Zeus can effectively improve the resource utilization of Kubernetes clusters. We discuss the design and implementation of Zeus and evaluate its effectiveness using latency-sensitive services and best-effort jobs in a massive production environment. The results show that by colocating latency-sensitive services with best-effort jobs, Zeus can increase the average CPU utilization from 15% to 60% without SLO violations.

show abstract

Invited Paper: Improving Data Center Efficiency Through Holistic Scheduling In Kubernetes

Cited by 36 publications

References 15 publications

Custom Scheduling in Kubernetes: A Survey on Common Problems and Solution Approaches

Custom Scheduling in Kubernetes: A Survey on Common Problems and Solution Approaches

Towards a Holistic Controller

Zeus: Improving Resource Efficiency via Workload Colocation for Massive Kubernetes Clusters

Contact Info

Product

Resources

About