The exponential growth in a number of the smart devices connected to the Internet of Things (IoT) in recent years has resulted in a massive increase in the volume of user generated content. Edge computing has been proposed to reduce the communication latency and improve the security of the user data by moving the computation and storage functions of these services closer to the users. However, the transmission delay between the devices and the edge server has a critical effect on the performance of the data offloading service since most IoT applications are delay-intolerant. Accordingly, the present study presents the use of measurement-based selection to determine the server with the shortest delay. Overall, the results show that irrespective of the scale of the networks, this kind of method helps to select the server with a reasonable cost.