Recently, with the advent of various Internet of Things (IoT) applications, a massive amount of network traffic is being generated. A network operator must provide different quality of service, according to the service provided by each application. Toward this end, many studies have investigated how to classify various types of application network traffic accurately. Especially, since many applications use temporary or dynamic IP or Port numbers in the IoT environment, only payload-based network traffic classification technology is more suitable than the classification using the packet header information as well as payload. Furthermore, to automatically respond to various applications, it is necessary to classify traffic using deep learning without the network operator intervention. In this study, we propose a traffic classification scheme using a deep learning model in software defined networks. We generate flow-based payload datasets through our own network traffic pre-processing, and train two deep learning models: 1) the multi-layer long short-term memory (LSTM) model and 2) the combination of convolutional neural network and single-layer LSTM models, to perform network traffic classification. We also execute a model tuning procedure to find the optimal hyper-parameters of the two deep learning models. Lastly, we analyze the network traffic classification performance on the basis of the F1-score for the two deep learning models, and show the superiority of the multi-layer LSTM model for network packet classification.

Edge computing is an emerging computing paradigm that distributes the computational capability to the edge of networks for enabling the computation near to the environment where the sensors and actuators are deployed. Therefore, from the network edge, heterogeneous solutions can be provided to the Internet based on sufficient computing ability. Nevertheless, computing and networking resources are constrained for devices in the network edge. Providing secure services from edge computing is a challenge based on constrained resources. In this paper, we propose a secure edge computing to provide management of device, data, user and additional services based on deploying independent microservices providers with a security gateway on an edge gateway. The edge gateway is the hub of a local network where multiple IoT devices are deployed to interact with the physical environment for sensing and actuating. The gateway provides the management functionalities through microservices based on multiple independent server modules. Each gateway-centric local network has a standalone management service based on the gateway. For providing secure edge computing services through the edge gateway, a security gateway is deployed on the proposed edge gateway to provide Representational State Transfer Application Programming Interfaces to expose the security services to the Internet instead of microservices from management modules. Moreover, a client support gateway is deployed in the edge gateway to provide services of User Interface and access forwarding based on web sessions to support user authentication and authorization with the security gateway. Based on the proposed edge gateway including client support and security gateway, IoT clients and IoT devices are enabled to communicate for providing secure edge services of access and visualization to users.

The development of the Internet of Things (IoT) has increased the ubiquity of the Internet by integrating all objects for interaction via embedded systems, leading to a highly distributed network of devices communicating with human beings as well as other devices. In recent years, cloud computing has attracted a lot of attention from specialists and experts around the world. With the increasing number of distributed sensor nodes in wireless sensor networks, new models for interacting with wireless sensors using the cloud are intended to overcome restricted resources and efficiency. In this paper, we propose a novel sensor-cloud based platform which is able to virtualize physical sensors as virtual sensors in the CoT (Cloud of Things) environment. Virtual sensors, which are the essentials of this sensor-cloud architecture, simplify the process of generating a multiuser environment over resource-constrained physical wireless sensors and can help in implementing applications across different domains. Virtual sensors are dynamically provided in a group which advantages capability of the management the designed platform. An auto-detection approach on the basis of virtual sensors is additionally proposed to identify the accessible physical sensors nodes even if the status of these sensors are offline. In order to assess the usability of the designed platform, a smart-space-based IoT case study was implemented, and a series of experiments were carried out to evaluate the proposed system performance. Furthermore, a comparison analysis was made and the results indicate that the proposed platform outperforms the existing platforms in numerous respects.

Intralogistics is a technology that optimizes, integrates, automates, and manages the logistics flow of goods within a logistics transportation and sortation center. As the demand for parcel transportation increases, many sortation systems have been developed. In general, the goal of sortation systems is to route (or sort) parcels correctly and quickly. We design an n-grid sortation system that can be flexibly deployed and used at intralogistics warehouse and develop a collaborative multi-agent reinforcement learning (RL) algorithm to control the behavior of emitters or sorters in the system. We present two types of RL agents, emission agents and routing agents, and they are trained to achieve the given sortation goals together. For the verification of the proposed system and algorithm, we implement them in a full-fledged cyber-physical system simulator and describe the RL agents’ learning performance. From the learning results, we present that the well-trained collaborative RL agents can optimize their performance effectively. In particular, the routing agents finally learn to route the parcels through their optimal paths, while the emission agents finally learn to balance the inflow and outflow of parcels.

In recent years the ever-expanding internet of things (IoT) is becoming more empowered to revolutionize our world with the advent of cutting-edge features and intelligence in an IoT ecosystem. Thanks to the development of the IoT, researchers have devoted themselves to technologies that convert a conventional home into an intelligent occupants-aware place to manage electric resources with autonomous devices to deal with excess energy consumption and providing a comfortable living environment. There are studies to supplement the innate shortcomings of the IoT and improve intelligence by using cloud computing and machine learning. However, the machine learning-based autonomous control devices lack flexibility, and cloud computing is challenging with latency and security. In this paper, we propose a rule-based optimization mechanism on an embedded edge platform to provide dynamic home appliance control and advanced intelligence in a smart home. To provide actional control ability, we design and developed a rule-based objective function in the EdgeX edge computing platform to control the temperature states of the smart home. Compared to cloud computing, edge computing can provide faster response and higher quality of services. The edge computing paradigm provides better analysis, processing, and storage abilities to the data generated from the IoT sensors to enhance the capability of IoT devices concerning computing, storage, and network resources. In order to satisfy the paradigm of distributed edge computing, all the services are implemented as microservices. The microservices are connected to each other through REST APIs based on the constrained IoT devices to provide all the functionalities that accomplish a trade-off between energy consumption and occupant-desired environment setting for the smart home appliances. We simulated our proposed system to control the temperature of a smart home; through experimental findings, we investigated the application against the delay time and overall memory consumption by the embedded edge system of EdgeX. The result of this research work suggests that the implemented services operated efficiently in the raspberry pi 3 hardware of IoT devices.

To seamlessly deliver artificial intelligence (AI) services using object detection, both inference latency from a system perspective as well as inference accuracy should be considered important. Although edge computing can be applied to efficiently operate these AI services by significantly reducing inference latency, deriving an optimized computational offloading policy for edge computing is a challenging problem. In this paper, we propose inference latency prediction approaches for determining the optimal offloading policy in edge computing. Since there is no correlation between the image size and inference latency during object detection, approaches to predict inference latency are required for finding the optimal offloading policy. The proposed approaches predict the inference latency between devices and object detection algorithms by using their statistical information on the inference latency. By exploiting the predicted inference latency, a client may efficiently determine whether to execute an object detection task locally or remotely. Through various experiments, the performances of predicted inference latency according to the object detection algorithms are compared and analyzed by considering two communication protocols in terms of the root mean square error. The simulation results show that the predicted inference latency matches the actual inference latency well.