Dynamic Monitoring of Very Large Wireless Systems

Recently we observe a wide-spread penetration of mobile devices which has been enlarged with the adoption of {em Internet-of-Things} (IoT) technologies. Consequently monitoring, configuring and sending queries to the large groups of wireless devices become a major challenge since current schemes require each device to be contacted individually. Addressing this shortage we propose DyMo for Dynamic Monitoring of very large wireless systems with thousands or millions of devices. AMUSE leverages the multicast capabilities of wireless networks such as {em LTE-eMBMS} (evolved~Multimedia~Broadcast/Multicast~Service) for distributing {em instructions} and {em queries} to all the relevant devices and analyzes their responses. This approach simplifies the monitoring and control of very large wireless systems and significantly reduces the communication overhead. Such solution is attractive to various applications of wireless network monitoring and IoT management. As an example use case we demonstrate the attractiveness of DyMo for monitoring LTE-eMBMS video multicasting in crowded areas, such as sports arenas. Monitoring the quality of such service is very challenging due to the lack of real-time feedback from the {em user equipment} (UEs). As a result, each eMBMS deployment requires an extensive and expansive field trail for tuning the service parameters, such as the eMBMS modulation and coding scheme (MCS) for ensuring the user satisfaction. After describing the practical realization of AMUSE for eMBMS monitoring, our performance analysis and large-scale evaluation show that Dymo infers the optimal eMBMS MCS, while meeting strict quality of service (QoS) requirement and with extremely low overhead. For instance, by having 20-30 reports per seconds we can ensures standard deviation of 0.1% of our estimation regardless of the population size.