Um arcabouço para o desenvolvimento e simulação de aplicações de internet das coisas em um cenário de computação na borda.

Abstract

The increase in the number of devices connected to the Internet has encouraged the development of several applications and services in different sectors of industry engineering. However, the possibility of using these devices in different applications is limited to the capacity of the network and its components. For instance, for applications that require a short response time, a network scenario where most of the data is processed in the cloud, may not be feasible. In order to have an environment that allows the development of these applications and services, the Edge Computing paradigm emerges, which enables data processing close to its generating source. This new paradigm enables to combine the best of both worlds, the fast data processing by a distributed network of equipment at Edge Computing, and the high storage and processing capacity of data centers in the cloud. Considering this context, one of the challenges inherent to this new paradigm is how to develop applications and validate them in an development and simulation environment. With this challenge in mind, this work proposes a framework that allows IoT applications to be developed using an IoT application middleware, and to validated in a Edge Computing network simulation environmet. For that, a proposal of integration framework is presented that aims to unite two distinct tools, Middleware Node-RED and FogNetSim ++, which is an extension of OMNeT ++, through a manager that controls the data exchange between them. Details of how the choice of tools was made, and the adaptations made that allow integration between the them, are presented in this work. In addition, experimental results of the operation of integration are presented in a validation use case. These tests demonstrate how the system reacts when modifying parameters such as interval of sending and receiving messages, type of mobility used by the devices, number of devices on the network, among other changes in the test scenario. Thus, depending on the application requirements of each developer, it is possible to tailor the simulation environment to their needs, and evaluate the behavior of the application before its launch.