Análise de desempenho de protocolos para computação em névoa

Abstract

With advancement in data processing and miniaturization in modern processors, mobile devices are becoming more popular. These devices use embedded systems with scarse resources such as battery, processing power and memory, which are important to provide a good user experience. Mobile devices, which are located at the edge of the network, could act as data producers and it is generally not feasible to centralize the data. In order to reduce problems related to data scalability at the edge of the network, the fog computing layer is used as a means of providing storage and communication services between cloud computing and mobile devices, providing real-time decentralized data in a scalable way. In fog computing there can be an entity called message broker, which is responsible for managing and distributing information through one or more communication channels in order to demand less computational resource. When using a message broker, it is also necessary to define which communication protocol will be used in the message transmission. This work aims to analyze some communication protocols for fog computing, such as MQTT (Message Queuing Telemetry Transport), AMQP (Advanced Message Queuing Protocol), and STOMP (SimpleText Oriented Messaging Protocol) working with RabbitMQ as message broker. Based on the results, it is possible to observe that MQTT presented the best performance regarding power consumption and processing, AMQP consumed less memory and STOMP presented a shorter time to complete the transmission of each message.