Operation and Security Considerations of Federated Learning Platform Based on Compute First Network

Zhao, Lijun; Tang, Xiaoyong; You, Zhengpeng; Pang, Yong; Xue, Hui; Zhu, Lei

doi:10.1109/icccworkshops49972.2020.9209951

Cited by 9 publications

(3 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The purpose of computing power network is to provide flexible computing resources for AI services, such as face recognition, smart surveillance. Meanwhile, AI algorithms, such as deep Q-learning network (DQN) [109], deep deterministic policy gradient (DDPG) [110], and federal learning (FL) [111], are optimizing computing power scheduling for computing power network. For the future of computing power network, it will be more intelligent and flexible with the progress of AI.…”

Section: Artificial Intelligencementioning

confidence: 99%

Computing Power Network: A Survey

Sun¹,

Liu²,

Huang³

et al. 2022

Preprint

View full text Add to dashboard Cite

With the rapid development of cloud computing, edge computing, and smart devices, computing power resources indicate a trend of ubiquitous deployment. The traditional network architecture cannot efficiently leverage these distributed computing power resources due to computing power island effect. To overcome these problems and improve network efficiency, a new network computing paradigm is proposed, i.e., Computing Power Network (CPN). Computing power network can connect ubiquitous and heterogenous computing power resources through networking to realize computing power scheduling flexibly. In this survey, we make an exhaustive review on the state-of-the-art research efforts on computing power network. We first give an overview of computing power network, including definition, architecture, and advantages. Next, a comprehensive elaboration of issues on computing power modeling, information awareness and announcement, resource allocation, network forwarding, computing power transaction platform and resource orchestration platform is presented. The computing power network testbed is built and evaluated. The applications and use cases in computing power network are discussed. Then, the key enabling technologies for computing power network are introduced. Finally, open challenges and future research directions are presented as well.

show abstract

Section: Artificial Intelligencementioning

confidence: 99%

Computing Power Network: A Survey

Sun¹,

Liu²,

Huang³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“… The current MEC structure cannot support hospital dedicated deployment: In principle, the great potential advantage of smart health systems can only be achieved if individuals are confident about the privacy of their health‐related information and providers are optimistic about the security of gathered data [8]. Although many solutions had been proposed to ensure medical data security [11, 17], hospitals also intend to fully supervise the 5G‐based smart healthcare information infrastructure for medical sensitive data protection at the current stage. Considering many network elements (e.g., UPF, AF, NEF, MEC service, etc.)…”

Section: Introductionmentioning

confidence: 99%

“…a few discussions on the MEC's structure designing and the integration methodology between 5G and MEC under the smart healthcare scenario, including but not only AF, NEF, DP, UPF, MEP, MEPM, and MEAO [6,7]. After discussing with more than 100 hospitals and deployment in more than ten hospitals [16], several key issues have been identified on the currently specified MEC structure by ETSI. Those issues are presented in the following (a)-(d), which will directly lead to the 5G-enabled smart healthcare use cases are almost only for demonstration at the current stage, and difficult to integrate with the existing medical systems of the hospital.…”

Section: Introductionmentioning

confidence: 99%

5G‐based smart healthcare system designing and field trial in hospitals

et al. 2021

Self Cite

View full text Add to dashboard Cite

With the 5G worldwide deployment, the scale of vertical applications is innovated benefit from 5G technologies including MEC (Multi-access Edge Computing), network slicing, etc. Especially for healthcare, 5G had been used for COVID-19 protection and intelligent medical processing. However, limited by the hospital's traditional information infrastructures, those 5G-based healthcare applications are hard to be deployed and most only for demonstration, also isolated from the existing medical systems. So what is the next generation of smart healthcare information infrastructures is the key issue for the longterm development of 5G healthcare applications. Even though the standardized 5G MEC framework has been widely used in many vertical scenarios, it is also hard to satisfy hospital-specific requirements such as hospital-dedicated deployment, medical data security, and various network connections, etc. This paper proposes a 5G-based architecture for smart healthcare information infrastructure, a new network element iGW (industry gateway) is defined, and the smart healthcare dedicated cloud platform iMEP (industry multi-access edge platform) is also introduced here, making it possible to satisfy both the hospital-specific requirements and the long-term evolution. Meanwhile, the implementation methodology and the corresponding field test results are presented, which show the significant network performance gain achieved by the proposed new system structure.

show abstract