devices, instant responses are produced in the services. In the healthcare industry, this is one of the lifesaving features.
Bandwidth Utilized Efficiently: as the data is processed in the edge devices, the process of data transmission to the cloud is eliminated, thus improving efficiency of the system and reducing network traffic.
Reliability: The data processing that happens nearer to the end user improves the reliability of edge computing, which is invulnerable to network outage and security threats.
Energy Efficient System: A limited amount of power supply is consumed by end devices, thus those devices are energy constrained. Edge computing brings the data computation nearer to the user end device. Thus usage of edge computing generates energy-efficient systems.
Figure 4.3 Advantages of edge computing.
Scalability: Scalability in edge computing can be achieved in an uncomplicated way, which helps in increasing the count of monitoring patients with a healthcare monitoring system. The capacity of the computing can be expanded by combining the Internet of Things devices with data centers in edge to achieve scalability. If there is a new addition in the end user, no demand of substantial bandwidth is imposed on the network core since the devices of edge computing are capable for processing. The scalability feature makes the edge computing more versatile for applications in the healthcare industry.
4.3.5 Challenges
There are some challenges in edge computing [30] when the services are offered. These include: Deployment of Edge Computing: Physical environment and space are shortage in edge devices. So, the design is made more suitable for the environment of the application. In the edge, computing capability can be brought until the space is available.
Power Supply: Only a limited amount of power supply is consumed by end devices, thus those devices are energy constrained. The edge devices must have the capability to process at any instant of time without outages. Thus it is mandatory for edge devices to utilize the power proficiently, which improves the efficiency of the system operations.
Data Backup: All the data are acquired from the sources and the processing of data is done in edge computing. It is crucial to give protection to all the collected data by the service provider. Data access and data storage reliability are critical for adding application security.
Maintenance: As the architecture is distributive in nature, proper maintenance is necessary for a system to work better. Device failures must be sorted to prevent any disturbance in the services offered by the system.
4.4 Smart Healthcare System
4.4.1 Methodology
In the IoT framework, the network resources are prioritized and intelligently used over a trustworthy and secure transmission channel that is used for applications of the health-care sector. The inputs are acquired from the sensors of the leaf devices in the IoT framework. The acquired inputs are preprocessed efficiently. If there is any high computation process involved, then the leaf device takes assistance from cloud servers at the backend. A large quantity of processes can be performed by backend cloud server and end devices can be advised by the backend cloud server on the preprocessing steps to give priority to the incoming data from the sensor. Data mining and machine language concepts are used at the backend for extracting the signatures from the sensor data. The healthcare interpretations are provided respectively to the captured sensor data. These steps are used by frontend devices for providing healthcare assessment of the patient.
This system can further be extended where a physician can be included at the back end to analyze the patient’s healthcare data for prominent fluctuations. Thus remote diagnosis can be improved and rural medication can also be provided as well, even with the healthcare provider far from the patient. The methodology of the smart healthcare system is shown in Figure 4.4.
Figure 4.4 Methodology of Smart Healthcare System.
In healthcare, low network latency, real-time responses are required. So cloud computing is not suitable in such situations because of its high network latency. Thus edge computing is proposed as a new distributive computing architecture that can perform most of the computations within the IoT edge devices instead of the cloud. In this chapter, the major focus is on combining the concepts of IoT and edge computing and improving the techniques of edge computing in the field of healthcare.
4.4.2 Data Acquisition and IoT End Device
IoT sensors and devices are used to collect the patient’s healthcare data. Sensors such as temperature sensor, ECG sensor, glucometer, sphygmomanometer, SpO2 sensor, EEG sensor, EMG sensor, and body position sensor are used in the system. Glucometer is used to measure the glucose content in the patient’s blood. Sphygmomanometer is mainly used for checking the blood pressure of the patient. SpO2 sensor is used to measure oxygen amounts present in the blood. EEG sensor is used to record the brain activities. EMG sensor is used to deduct the movements in the body. All these multiple sensors are connected to the leaf device or end point device, which is connected to the server at the backend via wireless network. IoT leaf device communicates with the backend server over a wireless communication medium by using IoT protocols such as Message Queuing Telemetry Transport (MQTT), Constrained Application Protocol (CoAP).
4.4.3 IoT End Device and Backend Server
Further, control messages and acquired healthcare data are exchanged between the leaf device and edge data server. The end devices send the patient’s condition data collected from various sensors to the edge data server and gets instructions to perform from the backend server. The computation tasks are carried out by using various techniques like machine learning. The severs at the back end can compute any number of heavy data and uses intensive algorithms computation for processing all the data acquired from the end devices; the instructions are sent as a notification to the end device.
Data analytics are performed in real time and for achieving optimized solution with low latency edge computing is used. The computation can be done by the end IoT devices that are based on the instruction and guidelines from the edge device which is provided by the cloud server. The MQTT client runs on IoT end device while the MQTT server runs on the edge server that can also further request various services from the cloud. MQTT can also be replaced by CoAP IoT protocol as an alternative. Tensor Flow can be used for machine learning [1, 6, 32]. It is an open-source free library that consists of tools with a flexible ecosystem, community resources and libraries for building and deploying machine learning applications.
4.5 Conclusion and Future Directions
The smart healthcare system consists of IoT wearables and sensors to collect patients’ healthcare data. Edge computing technology is used to extract the relevant information from the huge set of data. It is mainly used to minimize the response time, to eliminate network latency, saving bandwidth, and to provide an energy-efficient system. As the storage and computation is done in the edge nodes, data can be protected efficiently. The edge computing is coupled
