The Internet of Things (IoT) is becoming more and more integrated into our daily lives with the high number of connected devices around the world. However, IoT devices are constrained in terms of computational and networking resources, including bandwidth. Thus, IoT applications are distinguished by their criticality regarding response time, which we call latency, from a simple smart weather system to a remote surgery application that requires 1ms of latency. To face this challenge, researchers are always trying to optimize the network infrastructure and architecture of their applications, as well as the implemented communication protocols. In this context, we propose a new peer-to-peer architecture for topic-based pub/sub systems deployed at the edge that allows balancing the traffic load of IoT nodes to meet their bandwidth constraints and minimize the average data exchange latency. Our solution is compatible with the MQTT communication protocol and reduces average latency up to 5% compared to a local central server and up to 60% compared to a central cloud server in a local deployment, thus reducing average latency up to 44% compared to a central cloud server in a national deployment and reducing average latency up to 18% compared to a central cloud server in a worldwide deployment.
| Date | 15 Sept 2021 |
|---|
| Original language | French |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Julien Gascon-Samson (Supervisor) |
|---|
Bouallegue, C. (Author),
Gascon-Samson (Supervisor),
15 Sept 2021Student thesis: Master's thesis › Master in Engineering: Information Technology Engineering