Currently, information and communication technologies (ICT) deeply penetrate into many areas of modern life. For example, the concept of integrating ICT and the Internet of Things for managing Smart City infrastructure allows city authorities to monitor changes and the situation in the city using sensors. Thus, specialized systems collect data automatically without human intervention. An important parameter in determining the performance of wireless networks of machine-to-machine (M2M) interaction - data transfer rates, blocking probabilities, becomes the distance between the device (sensor) and the radio transmitting equipment (base station, BS). Therefore, describing such a network in the form of a queuing system with streaming (guaranteed data transfer rate) or elastic traffic (nonguaranteed speed), it is necessary to consider the incoming stream of requests for data transmission of M2M devices in such a way as to take into account the distance between devices and BS. In this paper, there is built a cell model of a wireless network with stationary M2M devices that are in a passive or active state, shown by points that appear randomly on a plane. The devices generate streaming traffic which depends on the distance from the BS, the device transmit power, and the noise level. The state of the system is described by the vector of variable length, the components of which are the distance of each active device to the BS. Two different disciplines of radio resource separation are considered - "round robin" and "full power," which differ in the distribution of the time interval for servicing an M2M device and the data transfer rate provided. There is built a random process with states enlarged by the number of users and a formula for calculating the probability of blocking a data transfer request is proposed. © 2019 Federal Research Center "Computer Science and Control" of Russian Academy of Sciences. All rights reserved.