Federated learning empowered mobility-aware proactive content offloading framework for fog radio access networks

Sanaullah Manzoor*, Adnan Noor Mian, Ahmed Zoha, Muhammad Ali Imran

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

11 Citations (Scopus)

Abstract

Proactive content caching has emerged as a promising solution to cope with exponentially increasing mobile data traffic. The popular user contents can be cached near the network edge for faster retrieval and processing. Current state-of-the-art approaches adopt a centralized model training mechanism that requires high communication and data exchange overheads to predict content popularity. Moreover, these approaches fail to deal with the dynamicity of the environment since they do not take into account the users’ mobility information and are unable to incorporate content offload timings. In this paper, we address these limitations by proposing a novel federated learning-based Mobility and Demand-aware Proactive Content Offloading (MDPCO) framework. MDPCO exploits distributed learning strategies and capitalizes on users’ mobility and demand information for proactive content offloading. Extensive simulations are carried out to validate the efficacy of MDCPO against local and cloud-based models. Our proposed model yields an average performance improvement of 6.7% in comparison to the cloud-based model. Furthermore, with the increase in the number of fog servers, the MDPCO achieves a 9.8% higher data offloading ratio and 1.18% increase in the downlink rates while being more energy-efficient than cloud-based approaches.

Original languageEnglish
Pages (from-to)307-319
Number of pages13
JournalFuture Generation Computer Systems
Volume133
Early online date25 Mar 2022
DOIs
Publication statusPublished - Aug 2022
Externally publishedYes

Keywords

  • Content offloading
  • F-RANs
  • Federated machine learning
  • Mobility management
  • Proactive caching

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Federated learning empowered mobility-aware proactive content offloading framework for fog radio access networks'. Together they form a unique fingerprint.

Cite this