The significant increase in mobile phone usage in the past few decades has resulted in various health hazards in individuals due to continuous exposure to radiation for a longer duration of time. To address the health hazards caused by mobile phone radiation, in this paper, we propose a mobile application (app)-based recommendation system to analyze smartphone usage patterns and Generative Artificial Intelligence (GAI)-based alert generation and recommendations to the user. We collect data on user’s mobile phone usage, including screen time, number of near-the-ear long-duration calls, and exposure to external radiation. These collected data offer a deeper understating of radiation exposure through mobile phone usage patterns of the user using the k-means clustering algorithm. As the excessive use of mobile phones leads to various health risks, including adverse physical and mental health conditions besides reduced productivity and social isolation, the designed application tracks the duration of user’s different application usage and generates an alert on excessive mobile usage, including longer near-the-ear calls and screen time as well as exposure to high external radiation. The app-based recommendation system also uses the GAI algorithm to inform users about their mobile usage pattern on a daily basis and the possibility of health hazards to help users take corrective action to reduce mobile usage. From the obtained result, we observe that the designed application is capable of precisely identifying and alerting the users based on their mobile usage.

The advent of Internet of Things (IoT) has resulted in a massive influx of data from remote networks, with big data originators located at the edge of the Internet. Software-Defined Networking (SDN) has recently emerged as an effective tool for centralized network control, including access nodes, to manage and optimize the flow of data generated by a vast number of IoT devices. However, remote and relay-positioned access nodes are often vulnerable to security attacks. In this paper, we propose DL-IPS, a novel intrusion detection mechanism for Software-Defined IoT (SD-IoT) based on Deep Learning (DL) techniques. Our proposed approach employs a Deep Neural Network (DNN) algorithm to monitor the traffic behaviours generated from IoT devices and predict potential intruders in the network. The proposed mechanism identifies malicious nodes and packets by monitoring incoming traffic behaviours at the local access nodes, thereby providing protection to the switch and flow tables from various attacks. Furthermore, our approach efficiently places flow rules to devices by removing malicious flows, reducing space consumption and overhead, and detecting anomalies effectively.

In this work, we propose Soft-Safe, a Software Defined Safety-as-a-Service (Safe-aaS) model for provisioning safety-related decisions to the registered end-users. In Safe-aaS, the end-users register to the infrastructure, provide their initial and destination location, select certain decision parameters, and make payment through a Web portal. As the safety-related decisions are time-critical in nature, therefore timely delivery of these decisions is essential. Considering these facts and road transportation as the application scenario of Safe-aaS, we address the problem of efficient decision delivery to the end-users in two stages. In the first stage, we propose a Software Defined Safe-aaS platform to address the problems of heterogeneity among the SDN switches present in the edge layer. Further, based on the utility of each of the SDN switches present within the vicinity of the end-users, we optimally select a suitable SDN switch among the available ones, for delivering them decisions in the second stage. To obtain the maximum utility for delivering decisions to the end-users, we map the interactions between the SDN controller and SDN switches as a Non-cooperative Single Leader Multiple Follower game. Then, we estimate the optimal delay incurred by an SDN switch applying the Lagrangian function and Karush-Kuhn-Tucker (KKT) conditions. Exhaustive simulation results illustrate that the energy consumed and delay incurred using our proposed scheme, Soft-Safe, is reduced compared to the existing schemes, Traditional Safe-aaS and MoRule.

With the growing number of green Internet of Things (IoT) applications, the underlining network and decision services need more Machine Learning (ML) models at the same time while reducing latency and utilized energy. In this paper, we propose HI-SDN, a Heterogeneous fog Intelligence enabled architecture for complex Software-Defined Network (SDN)-based IoT network running applications such as a smart city and healthcare. HI-SDN proposes a fog node and ML algorithm selection mechanism with dynamic traffic forwarding for green IoT to reduce delay and energy consumption. It adopts a reconfigurable approach to the ML method to dynamically change the required prediction policy and flow rules to be executed in a selected edge/fog node for critical computation. The results from the performance analysis show that HI-SDN has a substantial reduction in delay by 32% and consumed energy by 25%, in comparison to the existing state-of-the-art, besides having an increase in packet delivery ratio.

In this article, we propose a software-defined fog architecture, named as Soft-Health, to serve various Internet-of-Things (IoT)-based healthcare applications. The health conditions of the patients fluctuate over time. Further, specialized medical care may not always be available in all healthcare facilities. The use of wireless body area network (WBAN) for continuous patient monitoring addresses the issue to a certain extent. However, as the physiological parameters of a patient are time-critical in nature, any delay, packet loss, and network overhead, may result in deterioration of the patient’s health conditions. Considering this, we design a Software-defined fog-enabled IoT platform for various healthcare applications. We consider that the fog layer comprises SDN switches that allocate the packet to the appropriate fog/cloud depending upon the criticality index (CI) of the data packets originating from patients. We mathematically formulate the CI, based on the physiological parameters sensed and transmitted to the switches. Further, we design an optimization function to obtain the maximum utility of a fog node, for an optimal number of processes executed by that node. We apply the Lagrangian method to simplify the optimization function and solve it using Karush–Kuhn–Tucker (KKT) conditions. We apply the auto-regression model to predict the total delay incurred and the total energy consumed by the proposed scheme. Exhaustive analysis of our proposed scheme, Soft-Health, demonstrates that the delay incurred decreases by 24.57% and 40.1% approximately, compared to the existing schemes, Mobi-Flow, and CARE, respectively.

The resource constrained nature of IoT devices set about task offloading over the Internet for robust processing. However, this increases the Turnaround Time (TAT) of the IoT services. High TATs may cause catastrophe in time-sensitive environments such as chemical and steel industries, vehicular networks, healthcare, and others. Moreover, the unreliable Internet in rural parts of underdeveloped and developing countries is unsuitable for time-critical IoT systems. In this work, we propose a framework for continuous delivery of IoT services to address the issue of high latency/TAT with poor/no-internet coverage. The proposed framework guarantees service delivery in such areas. To demonstrate the proposed framework, we implemented an IoT-based mobile patient monitoring system. It predicts the patient's criticality using actual sensor data. When the sensed parameters exceed the pre-set threshold in the rule-base, it initiates data transfer to the fog or cloud server. If fog or the cloud is unreachable, it performs onboard predictions. Thus, the framework ensures essential service delivery to the user at all times. Our test-bed-based evaluation demonstrates edge CPU and RAM load reduction of 16% and 26%, respectively, in the ML model's test phase. Also, the results confirm continuous service delivery, reduced latency, power and computing resource consumption.

In this paper, we propose a microservices and deep learning-based scheme, termed as Micro-Safe, for provisioning Safety-as-a-Service (Safe-aaS) in a 6G environment. A Safe-aaS infrastructure provides customized safety-related decisions dynamically to the registered end-users. As the decisions are time-sensitive in nature, the generation of these decisions should incur minimum latency and high accuracy. Further, scalability and extension of the coverage of the entire Safe-aaS platform are also necessary. Considering road transportation as the application scenario, we propose Safe-aaS, which is a microservices- and deep learning-based platform for provisioning ultra-low latency safety services to the end-users in a 6G scenario. We design the proposed solution in two stages. In the first stage, we develop the microservices-enabled application layer to improve the scalability and adaptability of the traditional Safe-aaS platform. Moreover, we apply the state space model to represent the decision parameters requested and the decision delivered to the end-users. During the second stage, we use deep learning models to improve the accuracy in the decisions delivered to the end-users. Additionally, we apply an assortment of activation functions to analyze and compare the accuracy of the decisions generated in the proposed scheme. Extensive simulation of our proposed scheme, Micro-Safe, demonstrates that latency is improved by 26.1 – 31.2%, energy consumption is reduced by 22.1 – 29.9%, throughput is increased by 26.1 – 31.7%, compared to the existing schemes.

Due to the lack of an integrated communication and computation architecture for Software-Defined Healthcare IoT (SD-HI), provisioning critical services is challenging. In this paper, we propose SD-Health, an edge-based decision making and task allocation (EDT) scheme for SD-HI. The proposed SD-HI network uses Machine Learning (ML)-based approach to predict the criticality of flows and location of mobile devices. Based on the predicted values, the controller delegates the required EDT module to the respective edge node. The controller identifies the future healthcare-related decisions for an edge node and prepares the module accordingly. The ML-based trajectory prediction allows to find the future location of mobile devices in the network. Once the location of the mobile device is predicted, a set of computation tasks is dynamically allocated to the edge node. The results of performance analysis show that SD-Health has a significant improvement in latency by 43.3% and energy consumption by 30%, compared to the existing state-of-the-art, along with a fair improvement in packet delivery ratio.

In this paper, we propose Health-Flow, a criticality-aware traffic forwarding scheme for mobile devices to maximize the efficiency of a software-defined healthcare network. The proposed scheme uses a machine learning-based approach to find the criticality of flows and the location of the mobile device. Concerning the criticality levels in traffic, the proposed protocol dynamically places or removes flow-rules at the edge access points. Consequently, it helps to take adequate actions for the incoming requests adaptively with improved network reconfiguration overhead, latency, and energy consumption. We mathematically formulate Integer Linear Programming for optimally selecting access points. We mathematically formulate the resource reallocation problem in terms of optimization by minimizing the network overhead subject to the packet flows' criticality requirements. The proposed scheme has the potential to reduce latency by 52%, overhead by 19%, and energy consumption by 12% as compared to the existing schemes.