Search

Federated learning (FL) is an emerging machine learning technique where machine learning models are trained in a decentralized manner. The main advantage of this approach is the data privacy it provides because the data are not processed in a centralized device. Moreover, the local client models are aggregated on a server, resulting in a global model that has accumulated knowledge from all the different clients. This approach, however, is vulnerable to attacks because clients can be malicious or malicious actors may interfere within the network. In the first case, these types of attacks may refer to data or model poisoning attacks where the data or model parameters, respectively, may be altered. In this paper, we investigate the data poisoning attacks and, more specifically, the label-flipping case within a federated learning system. For an image classification task, we introduce two variants of data poisoning attacks, namely model degradation and targeted label attacks. These attacks are based on synthetic images generated by a generative adversarial network (GAN). This network is trained jointly by the malicious clients using a concatenated malicious dataset. Due to dataset sample limitations, the architecture and learning procedure of the GAN are adjusted accordingly. Through the experiments, we demonstrate that these types of attacks are effective in achieving their task and managing to fool common federated defenses (stealth). We also propose a mechanism to mitigate these attacks based on clean label training on the server side. In more detail, we see that the model degradation attack causes an accuracy degradation of up to 25%, while common defenses can only alleviate this for a percentage of ∼5%. Similarly, the targeted label attack results in a misclassification of 56% compared to 2.5% when no attack takes place. Moreover, our proposed defense mechanism is able to mitigate these attacks.

Cybersecurity in the Energy sector and relevant information sharing is at the foremost of European strategy towards the digital decade targeted by the EC for the years to come. The proposed Pan-European Incidents Information Sharing Platform (I2SP) offers a cyber-shield armour to European Electrical Power and Energy Systems (EPES) enabling cooperative detection of large scale, cyber-human security and privacy incidents and attacks. Via Incidents Information Sharing Platform, early detection and appropriate mitigation, guarantees the continuity of operations and minimization of cascading effects in the infrastructure itself, the environment, the citizens and the end-users.The Incidents’ Information Sharing Platform (I2SP) constitutes the software package which enables secure Cyber-Threat Intelligence (CTI) information sharing among EPES participants, as well as with trusted nominated entities, such as Information Sharing and Analysis Centers (ISACs), Computer Security Incident Response Teams (CSIRTs) and Security Operations Centers (SOC). I2SP facilitates technical information sharing in view of a warning system and incident reporting across the EU, aligning with the pillars of the new Network Code on Cybersecurity.

Food security is currently considered a huge societal challenge which technology providers, technology adopters, policy makers and consumers altogether are facing. A first step towards more secure and safe food has been attempted through the deployment mainly of Internet of Things based solutions and applications that gather information and provide it to the users. However, these systems are susceptible to attacks e.g. data leakage and information modification, keeping our societies away from the target of secure food. This paper explores the requirements of the farm to fork supply chain with respect to security and proposes a platform that aims at alleviating and mitigating a wide set of attacks.

Network intrusion detection is a key pillar towards the sustainability and normal operation of information systems. Complex threat patterns and malicious actors are able to cause severe damages to cyber-systems. In this work, we propose novel Deep Learning formulations for detecting threats and alerts on network logs that were acquired by pfSense, an open-source software that acts as firewall on FreeBSD operating system. pfSense integrates several powerful security services such as firewall, URL filtering, and virtual private networking among others. The main goal of this study is to analyse the logs that were acquired by a local installation of pfSense software, in order to provide a powerful and efficient solution that controls traffic flow based on patterns that are automatically learnt via the proposed, challenging DL architectures. For this purpose, we exploit the Convolutional Neural Networks (CNNs), and the Long Short Term Memory Networks (LSTMs) in order to construct robust multi-class classifiers, able to assign each new network log instance that reaches our system into its corresponding category. The performance of our scheme is evaluated by conducting several quantitative experiments, and by comparing to state-of-the-art formulations.

FASTER is an H2020 RIA project that develops a set of tools for enhancing the operational capacity of first responders while increasing their safety in the field. It will introduce augmented reality technologies for improved situational awareness and early risk identification and mobile and wearable technologies for better mission management and information delivery to first responders. Body- and gesture-based user interfaces will be employed to enable new capabilities while reducing equipment clutter, offering unprecedented ergonomics. Moreover, FASTER will provide a platform of autonomous vehicles aiming to collect valuable information from the disaster scene prior to operations, extend situational awareness and offer physical response capabilities to first responders. Furthermore, first responders will improve their situational awareness receiving information gathered and analysed by a Portable Common Operational Picture (PCOP). PCOP will gather multimodal data from the field, utilising an IoT network, and social media content to extract meaningful information and to orchestrate an intelligent response to the disaster. The whole system will be facilitated by tools for Resilient Communications Support featuring opportunistic relay services, emergency communication devices and 5G-enabled communication capabilities.

Modern critical infrastructures are increasingly turning into distributed, complex Cyber-Physical systems that need proactive protection and fast restoration to mitigate physical or cyber incidents or attacks. Addressing the need for early stage threat detection against physical intrusion, the paper presents two physical security sensors developed within the DEFENDER project for detecting the intrusion of drones and humans using video analytics. The continuous stream of media data obtained from the region of vulnerability and proximity is processed using Region based Fully Connected Neural Network deep-learning model. The novelty of the pro-posed system relies in the processing of multi-threaded media input streams for achieving real-time threat identification. The video analytics solution has been validated using NVIDIA GeForce GTX 1080 for drone detection and NVIDIA GeForce RTX 2070 Max-Q Design for detecting human intruders. The experimental test bed for the validation of the proposed system has been constructed to include environments and situations that are commonly faced by critical infrastructure operators such as the area of protection, tradeoff between angle of coverage against distance of coverage.

5G is considered to be the technology that will accommodate the development and management of innovative services with stringent and diverse requirements from end users, calling for new business models from the industry. In this context, the development and efficient management of Network Services (NS) serving specific vertical industries and spanning across multiple administrative domains and heterogeneous infrastructures is challenging. The main challenges regard the efficient provision of NSs considering the Quality of Service (QoS) requirements per vertical industry along with the optimal usage of the allocated resources. Towards addressing these challenges, this paper details an innovative approach that we have developed for managing and orchestrating such NSs, called SONATA, and compare it with OSM and Cloudify, which are two of the most known open-source Management and Orchestration (MANO) frameworks. In addition to examining the supported orchestration mechanisms per MANO framework, an evaluation of main operational and functional KPIs is provided based on experimentation using a real testbed. The final aim is the identification of their strong and weak points, and the assessment of their suitability for serving diverse vertical industry needs, including of course the Internet of Things (IoT) service ecosystem.

Recent technological innovations along with the vast amount of available data worldwide have led to the rise of cyberattacks against network systems. Intrusion Detection Systems (IDS) play a crucial role as a defense mechanism in networks against adversarial attackers. Machine Learning methods provide various cybersecurity tools. However, these methods require plenty of data to be trained efficiently, which may be hard to collect or to use due to privacy reasons. One of the most notable Machine Learning tools is the Generative Adversarial Network (GAN), and it has great potential for tabular data synthesis. In this work, we start by briefly presenting the most popular GAN architectures, VanillaGAN, WGAN, and WGAN-GP. Focusing on tabular data generation, CTGAN, CopulaGAN, and TableGAN models are used for the creation of synthetic IDS data. Specifically, the models are trained and evaluated on an NSL-KDD dataset, considering the limitations and requirements that this procedure needs. Finally, based on certain quantitative and qualitative methods, we argue and evaluate the most prominent GANs for tabular network data synthesis.

The recent advancements in the fields of artificial intelligence (AI) and machine learning (ML) have affected several research fields, leading to improvements that could not have been possible with conventional optimization techniques. Among the sectors where AI/ML enables a plethora of opportunities, industrial manufacturing can expect significant gains from the increased process automation. At the same time, the introduction of the Industrial Internet of Things (IIoT), providing improved wireless connectivity for real-time manufacturing data collection and processing, has resulted in the culmination of the fourth industrial revolution, also known as Industry 4.0. In this survey, we focus on the vital processes of fault detection, prediction and prevention in Industry 4.0 and present recent developments in ML-based solutions. We start by examining various proposed cloud/fog/edge architectures, highlighting their importance for acquiring manufacturing data in order to train the ML algorithms. In addition, as faults might also occur from sources beyond machine degradation, the potential of ML in safeguarding cyber-security is thoroughly discussed. Moreover, a major concern in the Industry 4.0 ecosystem is the role of human operators and workers. Towards this end, a detailed overview of ML-based human–machine interaction techniques is provided, allowing humans to be in-the-loop of the manufacturing processes in a symbiotic manner with minimal errors. Finally, open issues in these relevant fields are given, stimulating further research.

Live streaming is becoming an essential for the entertainment and media industry. Quality of Service (QoS) and reliability is a major factor in live streaming, whereas availability and interaction between users are considered very important. Furthermore, the amount of data those services produce is huge, while the data production rate has dramatically increased the last years. Thus, the use of 5G networking technologies, such as Network Function Virtualization (NFV), appears to be a promising solution for addressing the above-mentioned challenges. To this end, in this paper, we describe how an existing infrastructure of live streaming and social media services can take advantage of the NFV-enabled 5GTANGO Service Platform (SP). Also, in this paper we describe the capabilities of the SP and how it can support such services. To be more specific, an efficient monitoring framework, that supports the monitoring of Service Level Agreements (SLAs) is introduced. It aims to guarantee the QoS that is agreed in a high-level objective in the SLA, while this scenario is applied on an immersive media NS. Apart from the SP architecture and specific components description, the paper also provides explicit implementation requirements, needed for deployment of the presented media service and discusses promising results in terms of service management, monitoring flexibility and SLA assurance.

Energy consumption reserves a large portion of the budget for school buildings. At the same time, the students that use such facilities are the adults of the years to come and thus, should they embrace energy-aware behaviors, then sustainable results with respect to energy efficiency are anticipated. GAIA is a research project targeting this user domain, proposing a set of applications that a) aims at raising awareness, prompting action and fostering engagement in energy efficiency enhancement, and b) is adaptable to the needs of each facility/community. This application set relies on an IoT sensing infrastructure, as well as on the use of humans as sensors to create situational awareness.

Billions of embedded processors are being attached to everyday objects and houseware equipment to enhance daily activities and enable smart living. These embedded processors have enough processing capabilities to process sensor data to produce smart insights, and are designed to operate for months without the need of physical interventions. Despite the compelling features of Internet of Things (IoT), applied at several home-oriented use cases (e.g., lighting, security, heating, comfort), due to the lack of a physical flow of information (e.g., absence of switches and cable-based gateways), the security of such networks is impeding their rapid deployment. In this work we look into IPv6 based IoT deployments, since it is the leading standard for interconnecting the wireless devices with the Internet and we propose a data-driven anomaly detection system that operates at the transport-layer of 6LoWPAN deployments. We present a comprehensive experimental evaluation carried out using both simulated and real-world experimentation facilities that demonstrates the accuracy of our system against well-known network attacks against 6LoWPAN networks.

Currently, the coexistence of multiple users and devices challenges the network's ability to reliably connect them. This article proposes a novel communication architecture that satisfies the requirements of fifth-generation (5G) mobile network applications. In particular, this architecture extends and combines ultra-dense networking (UDN), multi-access edge computing (MEC), and virtual infrastructure manager (VIM) concepts to provide a flexible network of moving radio access (RA) nodes, flying or moving to areas where users and devices struggle for connectivity and data rate. Furthermore, advances in radio communications and non-orthogonal multiple access (NOMA), virtualization technologies and energy-awareness mechanisms are integrated towards a mobile UDN that not only allows RA nodes to follow the user but also enables the virtualized network functions (VNFs) to adapt to user mobility by migrating from one node to another. Performance evaluation shows that the underlying network improves connectivity of users and devices through the flexible deployment of moving RA nodes and the use of NOMA.

Autonomous fault detection plays a major role in the Critical Energy Infrastructure (CEI) domain, since sensor faults cause irreparable damage and lead to incorrect results on the condition monitoring of Cyber-Physical (CP) systems. This paper focuses on the challenging application of wind turbine (WT) monitoring. Specifically, we propose the two challenging architectures based on learning deep features, namely&mdash, Long Short Term Memory-Stacked Autoencoders (LSTM-SAE), and Convolutional Neural Network (CNN-SAE), for semi-supervised fault detection in wind CPs. The internal learnt features will facilitate the classification task by assigning each upcoming measurement into its corresponding faulty/normal operation status. To illustrate the quality of our schemes, their performance is evaluated against real-world&rsquo, s wind turbine data. From the experimental section we are able to validate that both LSTM-SAE and CNN-SAE schemes provide high classification scores, indicating the high detection rate of the fault level of the wind turbines. Additionally, slight modification on our architectures are able to be applied on different fault/anomaly detection categories on variant Cyber-Physical systems.

Intrusion detection plays a critical role in cyber-security domain since malicious attacks cause irreparable damages to cyber-systems. In this work, we propose theI2SPprototype, which is a novel Information Sharing Platform, able to gather, pre-process, model, and distribute network-traffic information. Within the I2SP prototype we build several challenging deep feature learning models for network-traffic intrusion detection. The learnt representations will be utilized for classifying each new network measurement into its corresponding threat level. We evaluate our prototype’s performance by conducting case studies using cyber-security data extracted from the Malware Information Sharing Platform (MISP)-API. To the best of our knowledge, we are the first that combine the MISP-API in order to construct an information sharing mechanism that supports multiple novel deep feature learning architectures for intrusion detection. Experimental results justify that the proposed deep feature learning techniques are able to predict accurately MISP threat-levels.

With the proliferation of energy-hungry small- and medium-sized data centers in urban contexts and granted the ever-increasing demand for digitalization of increasingly more services, an urgent need for managing the energy consumption of urban data centers has arisen. Such energy consumption management should be subject to limitations of variable elasticity such as the balancing needs of the smart grid, the service level agreements (SLAs) signed between the cloud services providers and their customers, the actual computational capacity of the data centers, and the needs of the data center owners for maximizing profit through service provisioning. The EU research project DOLFIN has proposed an approach towards optimizing the energy consumption of federated data centers by means of continuous resources monitoring and by exploiting flexible SLA renegotiation, coupled with the operation of predictive, multilayered optimization components.

Current partially or even no automated in-hospital processes often allow unacceptable errors to occur, the effects of which may span from simple injury to mortality. Recent advances in the Internet of Things, smart tags and cloud technologies may decisively alter this fact and minimize such "never events". In this paper, MATISSE is presented as a smart hospital ecosystem, aiming to decisively decrease "never events" in the hospital value chain. Real-time drugs/pills lifetime/aptness verification and medication administration, patients' identification and association with drugs and medical exams, as well as real-time quality assessment are supported in the proposed ecosystem. The ecosystem leverages the exploitation of information lying in (dynamic) smart tags attached to physical and virtual entities of the ecosystem, as well as the integration of a smart medication cart. The paper presents the supported use cases and the architectural design of the proposed solution, while it provides insights in both the hardware and software assembly of the smart medication cart.

The goal of this study was to investigate the performance of a realistic wireless sensor nodes deployment in order to support modern building management systems (BMSs). A three-floor building orientation is taken into account, where each node is equipped with a multi-antenna system while a central base station (BS) collects and processes all received information. The BS is also equipped with multiple antennas; hence, a multiple input–multiple output (MIMO) system is formulated. Due to the multiple reflections during transmission in the inner of the building, a wideband code division multiple access (WCDMA) physical layer protocol has been considered, which has already been adopted for third-generation (3G) mobile networks. Results are presented for various MIMO orientations, where the mean transmission power per node is considered as an output metric for a specific signal-to-noise ratio (SNR) requirement and number of resolvable multipath components. In the first set of presented results, the effects of multiple access interference on overall transmission power are highlighted. As the number of mobile nodes per floor or the requested transmission rate increases, MIMO systems of a higher order should be deployed in order to maintain transmission power at adequate levels. In the second set of results, a comparison is performed among transmission in diversity combining and spatial multiplexing mode, which clearly indicate that the first case is the most appropriate solution for indoor communications.

The energy sector represents undoubtedly one of the most significant “test cases” for 5G enabling technologies, due to the need of addressing a huge range of very diverse requirements to deal with across a variety of applications(stringent capacity for smart metering/AMI versus latency for supervisory control and fault localization). However, to effectively support energy utilities along their transition towards more decentralized renewable-oriented systems, several open issues still remain as to 5G networks management automation, security, resilience, scalability and portability. To face these issues, we outline a novel 5G PPP-compliant software framework specifically tailored to the energy domain, which combines i)trusted, scalable and lock-in free plug ‘n’ play support for a variety of constrained devices; ii) 5G devices’ abstractions to demonstrate mMTC (massive Machine Type Communications), uMTC (critical MTC) and xMBB (Extended Massive BroadBand) communications coupled with partially distributed, trusted, end-to-end security and MCM to enable secure, scalable and energy efficient communications; iii) extended Mobile Edge Computing (xMEC) micro-clouds to reduce backhaul load, increase the overall network capacity and reduce delays, while facilitating the deployment of generic MTC related NFVs (Network Function Virtualisation) and utility-centric VNF(Virtual Network Functions).

The massive deployment of IoT devices, broadband andmission criticalservices are paving the way for 5G communication networks, which will enable massive capacity, zerodelay, elasticity and optimal deployment, enhanced security, privacy by design and connectivity to billions of devices with less predictable traffic patterns. This paper targets a very important and demanding application the Preventive Maintenance as a Service in Critical Infrastructures and more precisely in the energy (electricity and gas) transmission and distribution network that combines the 5G technology with secure IoT and drones flight control. In more details, it addresses the 5G advances at the edge network and proposes a number of VNFs to support surveillance using swarms of drones.ENG

Current Internet architecture is being challenged by the overwhelming multiplication of available content and the increasing requirements of the content delivery applications. Content Delivery Networks have been playing a catalytic role in content delivery solutions, replicating (caching) content at the edges of the network. In this paper, we have investigated the potential benefits offered by adaptability in overlay content replication, considering changing conditions such as the popularity and the size of the content objects and a system of collaborating content caches. Pursuing the understanding of the influence of these parameters we have designed and implemented a simulation framework, with enhanced visualization and parameterization capabilities, and performed a series of experiments, under varying conditions.

During the last decades Intelligent Transportation Systems (ITS) have been attracting the interest of an increasing number of researchers, engineers and entrepreneurs, as well as citizens and civil authorities, since they can contribute towards improving road transport safety and efficiency and ameliorate environmental conditions and life quality. Emerging technologies yield miniaturized sensing, processing and communication devices that enable a high degree of integration and open the way for a large number of smart applications that can exploit automated fusion of information and enable efficient decisions by collecting, processing and communicating a large number of data in real-time. The cornerstone of these applications is the realization of an opportunistic wireless communication system between vehicles as well as between vehicles and infrastructure over which the right piece of information reaches the right location on time. In this paper, the authors present the design and implementation of representative safety and traffic management applications. Specifically the authors discuss the hardware and software requirements presenting a use case based on the NEC Linkbird-MX platform, which supports IEEE 802.11p based communications. The authors show how the functionality of IEEE 802.11p can be exploited to build efficient road safety and traffic management applications over mobile opportunistic systems and discuss practical implementation issues.

The complexity of the power grid, in conjunction with the ever increasing demand for electricity, creates the need for efficient analysis and control of the power system. The evolution of the legacy system towards the new smart grid intensifies this need due to the large number of sensors and actuators that must be monitored and controlled, the new types of distributed energy sources that need to be integrated and the new types of loads that must be supported. At the same time, integration of human-activity awareness into the smart grid is emerging and this will allow the system to monitor, share and manage information and actions on the business, as well as the real world. In this context, modeling and simulation is an invaluable tool for system behavior analysis, energy consumption estimation and future state prediction. In this paper, we review current smart grid simulators and approaches for building and user behavior modeling, and present a federated smart grid simulation framework, in which building, control and user behavior modeling and simulation are decoupled from power or network simulators and implemented as discrete components. This framework enables evaluation of the interactions between the communication infrastructure and the power system taking into account the human activities, which are at the focus of emerging energy-related applications that aim to shape user behavior. Validation of the key functionality of the proposed framework is also presented.

In the last few years Intelligent Transportation Systems (ITS) based on wireless vehicular networks have been attracting interest, since they can contribute towards improving road transport safety and efficiency and ameliorate environmental conditions and life quality. In order to widely spread these technologies, standardization at each layer of the networking protocol stacks has to be done. Therefore, a suite of protocols along with the architecture for the wireless environments with vehicles has been developed and standardized. Both in the US as well as in Europe the selected wireless communication protocol has been the 802.11p protocol developed by the IEEE. In this chapter, we discuss the potential impact of ITS towards achieving the above targets of improving road safety and traffic control. We review the overall architecture and the protocol functionality and present in detail a number of applications that have been developed demonstrating selected use-cases on an 802.11p compliant system prototype. Specifically, we discuss the implementation of selected applications on the NEC's Linkbird-MX platform, which supports IEEE 802.11p based communications, showing how its functionality can be exploited to build efficient road safety and traffic management applications, and evaluate the performance of the system using an experimental testbed.

As cloud services are becoming increasingly popular, the number of operating data centers is accordingly increasing, together with the need of implementing federated data centers and clouds. In this context, we consider a framework for achieving energy efficiency in federated clouds, by means of continuous monitoring and SLA renegotiation, coupled with the operation of prediction and multi-layered optimization components. In this paper, relevant prediction and optimization components, based on Support Vector Regression and Bin-Packing solving heuristics, operating at local data center level are examined and the experimental results of their deployment in a real-life testbed are presented and discussed.

In this work we discuss a mechanism for the monitoring and management of energy consumption in Wireless Sensor Networks. We consider that the Wireless Sensor Network consists of nodes that operate individually and collaborate with each other. After briefly discussing the typical network topologies and associating with the expected communications needs, we describe a conceptual framework for monitoring and managing the energy consumption on per process basis.

Wireless Sensor Networks (WSN) applications continue to expand and already cover almost all our daily activities improving from security and environmental efficiency to gaming experience. The diverse applications running on top of WSNs have led to the design of an immense number of routing protocols. Few years ago, the IETF standardized the IPv6 routing protocol for low-power and lossy networks (RPL) which is based on routing metrics to build communication paths between a source and the destination node. While significant efforts have focused on the design of routing metrics that satisfy the various applications, limited work has been reported on validating their performance using real-life motes. In this paper, we focus on validating the simulation results (obtained through the JSim simulator) using TelosB motes for a set of routing metrics that have been proposed in our previous articles.