Architectural design of IoT-Based healthcare hetero genous traffic: performance analysis and optimization

Abstract

The existing Internet of Things framework requires adjustments to suit healthcare setups by ensuring interoperability through flexible models for IoT devices that accommodate various communication protocols and standards, while managing resource allocation and optimizing the network for increased efficiency. Additionally, IoT healthcare systems encompass diverse traffic types requiring service guarantees for critical parameters like delay and throughput. Failure to meet these parameters could endanger human lives. Numerous scheduling schemes for IoT-based healthcare systems have been proposed like First Come First Served (FCFS), Earliest Deadline First (EDF) Rate-Monotonic, Preemptive Resume Service Priority, Dynamic Transmission Mechanism-L priority (DTM-L). Unfortunately, these schemes have limitations that range from large request starving short requests, process starvation for processes which require a long time to complete if short processes are continually added, performing poorly in overloaded conditions, not being optimal for multiprocessors, low throughput, high packet loss probability, to high priority requests starving lower priority requests. Previous attempts to schedule packets in IoT have used preemptive scheduling to improve the performance of delay sensitive packets. However, under high arrival rate of delay sensitive packets, the delay tolerant packets are starved due to unlimited preemptions by delay sensitive packets. This study designed an IoT-based healthcare architecture to alleviate the challenges of developing countries affected by limited resources in terms of IT infrastructure and poor health infrastructure, performed a performance analysis for heterogeneous multi-server priority queuing system based on the formulated packet transmission scheduling, and optimized the data traffic to meet specific critical parameters. In particular, the study proposed an IoT architecture that enables connection of medical sensors to collect patients' health data and process it to reduce on the burden on the healthcare systems in developing countries. The architecture utilizes Long Range (LoRa) technology. LoRa's low power consumption, long-range communication coverage, and low cost make it an attractive technology for healthcare monitoring. The proposed system architecture consists of IoT wearable ii

devices (sensors) that are LoRa-enabled, LoRa Gateway, Low Earth Orbit (LEO) satellite antenna, internet cloud server and end-user application terminals. The performance of the proposed architecture is evaluated in terms of battery power consumption, and sensor battery life using Framework for grained LoRa (FLoRa) network simulator within OMNeT environment. The results indicate that the LoRa-based network architecture outperforms the 4G-based network architecture in terms of energy efficiency. Specifically, the power consumption and battery life should be longer in the LoRa-based network compared to the 4G-based network. Additionally, we proposed an analytical model of a prioritized scheme (PS) that provides service differentiation in terms of delay sensitive packets receiving service before delay tolerant packets and also in terms of packet size with the short packets being serviced before large packets. The numerical results obtained from the derived models show that the PS offers better performance than FCFS and Shortest Job First (SJF) scheduling schemes for both short and large packets, except the shortest short packets that perform better under SJF than the prioritized scheme in terms of mean slowdown metric. We propose a system model of an Optimal Selective Pre-emption (OSP) scheduling scheme where, under high arrival rate of delay sensitive packets, only a certain fraction of the delay sensitive packets are served to prevent starvation of the delay tolerant packets. The numerical results show that the OSP scheme improves the performance of delay tolerant packets without appreciably degrading the service of delay sensitive packets. In the same breath, the performance of delay sensitive packets which is expected to be worse are within the tolerable limits. Furthermore, this study proposed a system model for the PS scheme. The PS scheme is an improvement of the EDF scheme for IoT-based healthcare applications. The PS scheme uses a heterogeneous multi-server priority queuing system to provide service differentiation by prioritizing short packets over large packets and delay sensitive packets are serviced before delay tolerant packets. In particular, we showed that the PS scheme performs better than the EDF scheme in terms of throughput for all packet sizes and at both low and high load values.