Large-scale cluster management at Google with Borg

Verma, Abhishek; Pedrosa, Luis; Korupolu, Madhukar; Oppenheimer, David; Tune, Eric; Wilkes, John

doi:10.1145/2741948.2741964

Cited by 987 publications

(614 citation statements)

References 54 publications

Supporting

Mentioning

606

Contrasting

Unclassified

Order By: Relevance

“…In actual practice the cloud infrastructure is the enabler of these essential characteristics. Due to the economy of scale, infrastructure installations are large and typically built of commodity hardware so that failures are the norm rather than the exception [Verma et al (2015)]. Finally, cloud applications often rely on third-party services, as part of the application functionality, support (e.g., monitoring) or both.…”

Section: Cna: Definitions and Requirementsmentioning

confidence: 99%

Self-managing cloud-native applications: Design, implementation, and experience

Toffetti

Brunner

Blöchlinger

et al. 2017

Future Generation Computer Systems

View full text Add to dashboard Cite

Running applications in the cloud efficiently requires much more than deploying software in virtual machines. Cloud applications have to be continuously managed : 1) to adjust their resources to the incoming load and 2) to face transient failures replicating and restarting components to provide resiliency on unreliable infrastructure. Continuous management monitors application and infrastructural metrics to provide automated and responsive reactions to failures (health management) and changing environmental conditions (auto-scaling) minimizing human intervention.In the current practice, management functionalities are provided as infrastructural or third party services. In both cases they are external to the application deployment. We claim that this approach has intrinsic limits, namely that separating management functionalities from the application prevents them from naturally scaling with the application and requires additional management code and human intervention. Moreover, using infrastructure provider services for management functionalities results in vendor lock-in effectively preventing cloud applications to adapt and run on the most effective cloud for the job.In this paper we discuss the main characteristics of cloud native applications, propose a novel architecture that enables scalable and resilient selfmanaging applications in the cloud, and relate on our experience in porting a legacy application to the cloud applying cloud-native principles.

show abstract

Section: Cna: Definitions and Requirementsmentioning

confidence: 99%

Self-managing cloud-native applications: Design, implementation, and experience

Toffetti

Brunner

Blöchlinger

et al. 2017

Future Generation Computer Systems

View full text Add to dashboard Cite

show abstract

“…Based on the lightweight architecture of container-based virtualization, containers offer several advantages over virtual machines such as high performance, resource efficiency, and agile environment. As a result of these advantages, containers have been adopted in the IT industry in areas such as cloud data centers, mobile systems, and networks [16][17][18].…”

Section: Background: Container-based Virtualizationmentioning

confidence: 99%

The Impact of Container Virtualization on Network Performance of IoT Devices

Lee

Kim

2018

Mobile Information Systems

View full text Add to dashboard Cite

Container-based virtualization offers advantages such as high performance, resource efficiency, and agile environment. ese advantages make Internet of ings (IoT) device management easy. Although container-based virtualization has already been introduced to IoT devices, the different network modes of containers and their performance issues have not been addressed. Since the network performance is an important factor in IoT, the analysis of the container network performance is essential. In this study, we analyze the network performance of containers on an IoT device, Raspberry Pi 3. e results show that the network performance of containers is lower than that of the native Linux, with an average performance difference of 6% and 18% for TCP and UDP, respectively. In addition, the network performance of containers varies depending on the network mode. When a single container runs, bridge mode achieves higher performance than host mode by 25% while host mode shows better performance than bridge mode by 45% in the multicontainer environment.

show abstract

“…Choosing Mesos with Marathon [13] is an excellent combination if you already have a Mesos cluster. Kubernetes [14] is the third container-management system for automating deployment, scaling, and management of containerized applications influenced by Borg [9]. It aims to provide better ways of managing related, distributed components across varied infrastructure, while still benefiting from the improved utilization containers enable.…”

Section: Related Workmentioning

confidence: 99%

“…The development of containerization technology has greatly accelerated the popularization and application of deep learning. To put deep learning task into container and make the container as the unit of management can enable broader and easier deployment of deep learning algorithms for building cloud-based distributed deep learning platform [9]. In addition, it enables higher levels of utilization and can transition from training to serving smoothly.…”

Section: Introductionmentioning

confidence: 99%

DLTAP: A Network-efficient Scheduling Method for Distributed Deep Learning Workload in Containerized Cluster Environment

Wang

Liu

2017

ITM Web Conf.

View full text Add to dashboard Cite

Abstract:Deep neural networks (DNNs) have recently yielded strong results on a range of applications. Training these DNNs using a cluster of commodity machines is a promising approach since training is time consuming and compute-intensive. Furthermore, putting DNN tasks into containers of clusters would enable broader and easier deployment of DNN-based algorithms. Toward this end, this paper addresses the problem of scheduling DNN tasks in the containerized cluster environment. Efficiently scheduling data-parallel computation jobs like DNN over containerized clusters is critical for job performance, system throughput, and resource utilization. It becomes even more challenging with the complex workloads. We propose a scheduling method called Deep Learning Task Allocation Priority (DLTAP) which performs scheduling decisions in a distributed manner, and each of scheduling decisions takes aggregation degree of parameter sever task and worker task into account, in particularly, to reduce cross-node network transmission traffic and, correspondingly, decrease the DNN training time. We evaluate the DLTAP scheduling method using a state-of-the-art distributed DNN training framework on 3 benchmarks. The results show that the proposed method can averagely reduce 12% cross-node network traffic, and decrease the DNN training time even with the cluster of low-end servers.

show abstract

Large-scale cluster management at Google with Borg

Cited by 987 publications

References 54 publications

Self-managing cloud-native applications: Design, implementation, and experience

Self-managing cloud-native applications: Design, implementation, and experience

The Impact of Container Virtualization on Network Performance of IoT Devices

DLTAP: A Network-efficient Scheduling Method for Distributed Deep Learning Workload in Containerized Cluster Environment

Contact Info

Product

Resources

About