Dribble: A learn-based timer scheme selector for mobility management in IoT

Abstract

In this work, we present Dribble a learn-based timer scheme selector to manage topology changes caused by mobility in the Internet of Things (IoT) context. IoT has turned smart devices part of our everyday lives. They are in everywhere with many shapes, sizes, and capabilities. For IoT to become even more ubiquitous, it is necessary to overcome the challenges posed by mobility. One of them is the management of topology changes, especially at the network layer. Currently, routing protocols check the topology through an advertisement timer scheme. Such schemes face a basic trade-off between being fast to find topology problems and concurrently be energy and overhead control saver. Although there are timer schemes designed to mobile context, all devices are governed by the same one, which is a hard assumption since IoT is heterogeneous and naturally, devices have different behaviors. Thus, Dribble learns the devices' mobility pattern and then it assigns a custom-made timer scheme conveniently for each device. Our results show that personalized timer schemes present better performance than single traditional timer schemes such as Trickle Timer (TT) and Reverse Trickle Timer (RevTT).