Efficient Solution for Large-Scale IoT Applications with Proactive Edge-Cloud Publish/Subscribe Brokers Clustering

Large-scale IoT applications with dozens of thousands of geo-distributed IoT devices creating enormous volumes of data pose a big challenge for designing communication systems that provide data delivery with low latency and high scalability. In this paper, we investigate a hierarchical Edge-Cloud pu...

Full description

Saved in:
Bibliographic Details
Published in:Sensors (Basel, Switzerland) Switzerland), 2021-12, Vol.21 (24), p.8232
Main Authors: Pham, Van-Nam, Lee, Ga-Won, Nguyen, VanDung, Huh, Eui-Nam
Format: Article
Language:English
Subjects:
Brokers
Cloud computing
Clustering
Collaboration
Communication
Communications systems
Delivery services
distributed publish/subscribe systems
implicit collaborative filtering
Internet of Things
Peers
Sensors
Servers
Smart cities
Social networks
Subscriptions
topic similarity
topic subscription prediction
topic-based publish/subscribe
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Large-scale IoT applications with dozens of thousands of geo-distributed IoT devices creating enormous volumes of data pose a big challenge for designing communication systems that provide data delivery with low latency and high scalability. In this paper, we investigate a hierarchical Edge-Cloud publish/subscribe brokers model using an efficient two-tier routing scheme to alleviate these issues when transmitting event notifications in wide-scale IoT systems. In this model, IoT devices take advantage of proximate edge brokers strategically deployed in edge networks for data delivery services in order to reduce latency. To deliver data more efficiently, we propose a proactive mechanism that applies collaborative filtering techniques to efficiently cluster edge brokers with geographic proximity that publish and/or subscribe to similar topics. This allows brokers in the same cluster to exchange data directly with each other to further reduce data delivery latency. In addition, we devise a coordinative scheme to help brokers discover and bridge similar topic channels in the whole system, informing other brokers for data delivery in an efficient manner. Extensive simulation results prove that our model can adeptly support event notifications in terms of low latency, small amounts of relay traffic, and high scalability for large-scale, delay-sensitive IoT applications. Specifically, in comparison with other similar Edge-Cloud approaches, our proposal achieves the best in terms of relay traffic among brokers, about 7.77% on average. In addition, our model's average delivery latency is approximately 66% of PubSubCoord-alike's one.
ISSN:1424-8220
1424-8220
DOI:10.3390/s21248232