Proteus

Harlap, Aaron; Tumanov, Alexey; Chung, Andrew; Ganger, Gregory R.; Gibbons, Phillip B.

doi:10.1145/3064176.3064182

Cited by 68 publications

(5 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several strategies have been put out to deal with these limitations and improve cloud computing's capabilities. For instance, Proteus, an elastic PS framework created to scale up training on public clouds, was introduced by Harlap et al [89]. The framework dynamically assigns PSs and personnel using three transitional stages, maximizing cost reductions, particularly when temporary revocable resources become available.…”

Section: Parameter Servermentioning

confidence: 99%

Deep Learning Neural Networks in the Cloud

Awan

2023

IJAEMS

View full text Add to dashboard Cite

Deep Neural Networks (DNNs) are currently used in a wide range of critical real-world applications as machine learning technology. Due to the high number of parameters that make up DNNs, learning and prediction tasks require millions of floating-point operations (FLOPs). Implementing DNNs into a cloud computing system with centralized servers and data storage sub-systems equipped with high-speed and high-performance computing capabilities is a more effective strategy. This research presents an updated analysis of the most recent DNNs used in cloud computing. It highlights the necessity of cloud computing while presenting and debating numerous DNN complexity issues related to various architectures. Additionally, it goes into their intricacies and offers a thorough analysis of several cloud computing platforms for DNN deployment. Additionally, it examines the DNN applications that are already running on cloud computing platforms to highlight the advantages of using cloud computing for DNNs. The study highlights the difficulties associated with implementing DNNs in cloud computing systems and provides suggestions for improving both current and future deployments.

show abstract

Section: Parameter Servermentioning

confidence: 99%

Deep Learning Neural Networks in the Cloud

Awan

2023

IJAEMS

View full text Add to dashboard Cite

show abstract

“…With the need for training DNNs beyond convolutional neural networks [72], [73], DNN training remains a challenging cloud engineering problem -even with the emergence of serverless training services [74], [75]. Additionally, as the training scenario shifts from dedicated clusters (one training job per cluster) to shared clusters, problems such as resource provisioning (with cheap transient resources) [76], [77] and GPU scheduling [78]- [80] still remain unresolved to effectively trade-off cluster utilization and training accuracy and throughput. Edge resources and micro-data centers close to data sources can be utilized for distributed DL training [81], which requires solving challenges of data distribution and resource heterogeneity.…”

Section: Iot and Ai Are Becoming The Main Applicationsmentioning

confidence: 99%

On the Future of Cloud Engineering

Bermbach

Chandra

Krintz

et al. 2021

2021 IEEE International Conference on Cloud Engineering (IC2E)

View full text Add to dashboard Cite

Ever since the commercial offerings of the Cloud started appearing in 2006, the landscape of cloud computing has been undergoing remarkable changes with the emergence of many different types of service offerings, developer productivity enhancement tools, and new application classes as well as the manifestation of cloud functionality closer to the user at the edge. The notion of utility computing, however, has remained constant throughout its evolution, which means that cloud users always seek to save costs of leasing cloud resources while maximizing their use. On the other hand, cloud providers try to maximize their profits while assuring service-level objectives of the cloud-hosted applications and keeping operational costs low. All these outcomes require systematic and sound cloud engineering principles. The aim of this paper is to highlight the importance of cloud engineering, survey the landscape of best practices in cloud engineering and its evolution, discuss many of the existing cloud engineering advances, and identify both the inherent technical challenges and research opportunities for the future of cloud computing in general and cloud engineering in particular.

show abstract

“…Narrow Applicability: Some existing approaches often suffer from limited applicability, restricting their effectiveness in specific scenarios. Proteus [11] only considers distributed learning environments with a Parameter Server (PS) architecture, and as the number of Spot VMs increases, the proportion of On-Demand VM usage also increases, reducing cost-efficiency. Spotnik [12] is limited to situations where only part of the cluster faces preemption and is incapable of avoiding complete checkpoint approaches when the entire cluster is revoked.…”

Section: Existing Approaches and Their Limitationsmentioning

confidence: 99%

“…Limitations of Existing Approaches: There is a very small effort to construct cost-effective clusters that leverage both Spot and On-Demand VMs. Some studies [11,12,13,14] are limited by their narrow applicability, as they are effective only in certain usecases and/or are tailored for specific architectures. Other approaches [15,16,17,18] fail adequately propose methods for cluster configuration, lack comprehensive analysis or experimentation in complex scenarios, and fail to simultaneously consider both price and performance.…”

mentioning

confidence: 99%