Priorização de alertas médicos em redes sem fio definidas por software: uma perspectiva experimental

Abstract

Technological advances in Wireless Body Area Networks (WBAN) have enabled their use in different scenarios. Among their many possible areas of application, health monitoring systems are one of the most notable ones, as these networks make it possible to monitor patients in real time and diagnose many fatal diseases. The collected sensitive data can be fed to systems capable of early detection of critical events related to the patient’s health status. Once a critical event is identified, medical alert packets can be sent with some level of prioritization to medical servers, enabling rapid relief provisioning and reducing the risk of death or serious sequelae for the patient. In a hospital environment, medical alert packets contend with multiple traffic flows, with varying characteristics, for the wireless local area network (WLAN) resources. Thus, one of the main challenges in this context is to ensure lower latencies and loss rates for medical data packets. Furthermore, it is desirable that the impact on the overall performance of the network be kept to a minimum, which is why it is important to adopt an on-demand approach, which acts only when medical data flows are identified, and only on the access points (APs) that participate in the communication. In this sense, this work proposes a platform based on Software Defined Wireless Networks (SDWN) for on-demand access prioritization to the wireless medium for medical alert packets, based on the programmatic configuration of medium access control sublayer parameters in IEEE 802.11 wireless networks. These parameters allow the definition, for each of the four WLAN traffic categories, of the amount of time that the wireless channel must be idle before a data transmission is initiated. By varying these parameters, it’s possible to give traffic categories different wireless channel access priorities. The platform was evaluated in a real wireless network environment, with different quantities of wireless stations connected to the AP. In the highest density scenario among the ones built for the tests, with 20 wireless clients, there was a 34.65% reduction in the average delay of prioritized medical alert packets compared to non-prioritized ones, in addition to a reduction of 58.62% in the amount of packets lost.