Smart-Contract-Based Automation for OF-RAN Processes: A Federated Learning Use-Case

The opportunistic fog radio access network (OF-RAN) expands its offloading computation capacity on-demand by establishing virtual fog access points (v-FAPs), comprising user devices with idle resources recruited opportunistically to execute the offloaded tasks in a distributed manner. OF-RAN is attr...

Full description

Bibliographic Details
Main Authors: Jofina Jijin, Boon-Chong Seet, Peter Han Joo Chong
Format: Article
Language:English
Published: MDPI AG 2022-09-01
Series:Journal of Sensor and Actuator Networks
Subjects:
Online Access:https://www.mdpi.com/2224-2708/11/3/53
Description
Summary:The opportunistic fog radio access network (OF-RAN) expands its offloading computation capacity on-demand by establishing virtual fog access points (v-FAPs), comprising user devices with idle resources recruited opportunistically to execute the offloaded tasks in a distributed manner. OF-RAN is attractive for providing computation offloading services to resource-limited Internet-of-Things (IoT) devices from vertical industrial applications such as smart transportation, tourism, mobile healthcare, and public safety. However, the current OF-RAN design is lacking a trusted and distributed mechanism for automating its processes such as v-FAP formation and service execution. Motivated by the recent emergence of blockchain, with smart contracts as an enabler of trusted and distributed systems, we propose an automated mechanism for OF-RAN processes using smart contracts. To demonstrate how our smart-contract-based automation for OF-RAN could apply in real life, a federated deep learning (DL) use-case where a resource-limited client offloads the resource-intensive training of its DL model to a v-FAP is implemented and evaluated. The results validate the DL and blockchain performances of the proposed smart-contract-enabled OF-RAN. The appropriate setting of process parameters to meet the often competing requirements is also demonstrated.
ISSN:2224-2708