Currently, there are over 14 billion IoT devices [7], and with many devices come many protocols, the main ones being MQTT and CoAP. We are interested in connecting the many diverse IoT devices to the cloud. To do so, we use the middleware architecture proposed by article [8] in which a device, called the middleware, acts as the middleman between the various IoT networks and the cloud. Since IoT devices typically operate in real-time, performance is of great concern. Therefore, we conducted a simulation to measure the data latency of using middleware and the overall fairness between different IoT networks. Our simulation had an MQTT and a CoAP network interacting with the middleware. The simulation results showed that CoAP always had a lower travel time than MQTT, mainly because CoAP is a more lightweight protocol. However, we also found that MQTT had slightly more throughput, which was unexpected since we initially thought that CoAP would have had higher throughput. We have shown that analyzing data via a middleware device is possible and that there are potential directions to explore, such as evaluating different Quality of Service Algorithms in the context of having a middleware device.

翻译：目前，物联网设备数量已超过140亿台[7]，众多设备带来了多种协议，其中MQTT和CoAP是主要协议。我们致力于将各类异构物联网设备连接到云端。为此，采用文献[8]提出的中间件架构：一个称为中间件的设备充当物联网网络与云端之间的中介。由于物联网设备通常需要实时运行，性能至关重要。因此，我们通过仿真实验评估使用中间件时的数据延迟以及不同物联网网络间的整体公平性。仿真中，MQTT与CoAP网络分别与中间件交互。结果表明，CoAP的传输时间始终低于MQTT，主要原因是CoAP属于更轻量级的协议。然而，我们还发现MQTT的吞吐量略高于CoAP，这一结果出乎意料，因为初始假设认为CoAP会具备更高吞吐量。本研究证明，通过中间件设备分析数据是可行的，并指明了潜在的研究方向，例如在中间件环境下评估不同服务质量算法。