Loading…

User preference-aware content caching strategy for video delivery in cache-enabled IoT networks

•Introduction of a novel learning model founded on VAE, trained with historical user request data, enabling accurate prediction of current users' future preferences.•Utilization of predicted content preferences to identify popular content within each SBS coverage area, subsequently stored withi...

Full description

Saved in:
Bibliographic Details
Published in:Computer networks (Amsterdam, Netherlands : 1999) Netherlands : 1999), 2024-02, Vol.240, p.110142, Article 110142
Main Authors: Firouzjaee, Mostafa Taghizade, Jamshidi, Kamal, Moghim, Neda, Shetty, Sachin
Format: Article
Language:English
Subjects:
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Introduction of a novel learning model founded on VAE, trained with historical user request data, enabling accurate prediction of current users' future preferences.•Utilization of predicted content preferences to identify popular content within each SBS coverage area, subsequently stored within the respective SBS cache.•Introduction of an online cache replacement algorithm, post-placement of popular content in SBS caches, designed to address instantaneous data events and deviations between predicted and actual content popularity. This algorithm considers both predicted content popularity and content request frequency and timing.•Integration of a collaborative caching algorithm to enhance caching efficiency, enabling SBSs to request content from adjacent SBSs when user-requested content is unavailable within their cache.•Validation of the proposed approach through evaluations showcasing superior performance compared to prior methods, based on key performance parameters such as cache hit rate and CRD. The escalating growth of content-dependent services and applications within the Internet of Things (IoT) platform has led to a surge in traffic, necessitating real-time data processing. Content caching has emerged as an effective solution to counteract this traffic upswing. Caching not only improves network efficiency but also enhances user service quality. Critical to the development of an optimal caching algorithm is the accurate prediction of future content popularity. This prediction hinges on the ability to anticipate users' content preferences, which is a pivotal method for assessing content popularity. In this study, we introduce a novel caching strategy termed User Preference-aware content Caching Strategy (UPCS) tailored for an IoT platform, where users access multimedia services offered by remote Content Providers (CPs). The UPCS encompasses three key algorithms: a content popularity prediction algorithm that utilizes Variational Autoencoders (VAE) to forecast users' future content preferences based on their prior requests, an online algorithm for dynamic cached content replacement, and a cooperative caching algorithm to augment caching efficiency. The proposed content caching strategy outperforms alternative methods, exhibiting superior cache hit rates and reduced Content Retrieval Delays (CRD).
ISSN:1389-1286
1872-7069
DOI:10.1016/j.comnet.2023.110142