Behavioral Models and Scenario Selection for Testing IoT Trickle-Based Lossy Multicast Networks

Abstract

The Trickle algorithm is widely used to manage information dissemination in Wireless Sensor Networks (WSN) and it continues to be studied, in order to evaluate its performance or to derive variants or use in higher-level protocols. In this paper, we propose behavioural test models for the Trickle-based Multicast Protocol for Low Power and Lossy Networks (MPL). MPL proposes to use Trickle not only to schedule on-demand retransmissions but also to discover new information that has not yet been received, eventually due to non-deterministic packets loss. Testing such a protocol is challenging as it is used in particular to update sensors functionalities by disseminating prioritized large-scale firmware upgrades in the network. As part of our contribution, we have implemented reachability-based techniques which allow the selection of distributed test objectives in the form of timed scenarios with high-coverage of critical multicast communications achieving an up-to-date state of the sensors network.