International Journal of Network Management最新文献

Advancement and flourishment in mobile edge computing (MEC) have motivated the farmers to deploy an efficient ecosystem in their farms. For further real-time monitoring and surveillance of the environment along with the deployment of intelligent farming, digital twin is considered as one of the emerging and most promising technologies. For proper optimization and utilization of physical systems, the physical components of the ecosystems are connected with the digital space. Further, the smart technologies and devices have convinced to address the expected level of requirements for accessing the rapid growth in farming associated with digital twins. However, with a large number of smart devices, huge amount of generated information from heterogeneous devices may increase the privacy and security concern by challenging the interrupting operations and management of services in smart farming. In addition, the growing risks associated with MEC by modifying the sensor readings and quality of service further affect the overall growth of intelligent farming. In order to resolve these challenges, this paper has proposed a secure surveillance architecture to detect deviations by incorporating digital twins in the ecosystem. Further, for real-time monitoring and preprocessing of information, we have integrated a four-dimensional trust mechanism along with blockchain. The four-dimensional trusted method recognizes the behavior of each communicating device during the transmission of information in the network. Further, blockchain strengthens the surveillance process of each device behavior by continuously monitoring their activities. The proposed mechanism is tested and verified against various abnormalities received from sensors by simulating false use cases in the ecosystem and compared against various security metrics over existing approaches. Furthermore, the proposed mechanism is validated against several security threats such as control command threat, coordinated cyber threats, accuracy, and decision-making and prediction of records over existing methods.

Deep neural network (DNN) and machine learning (ML) models/ inferences produce highly accurate results demanding enormous computational resources. The limited capacity of end-user smart gadgets drives companies to exploit computational resources in an edge-to-cloud continuum and host applications at user-facing locations with users requiring fast responses. Kubernetes hosted inferences with poor resource request estimation results in service level agreement (SLA) violation in terms of latency and below par performance with higher end-to-end (E2E) delays. Lifetime static resource provisioning either hurts user experience for under-resource provisioning or incurs cost with over-provisioning. Dynamic scaling offers to remedy delay by upscaling leading to additional cost whereas a simple migration to another location offering latency in SLA bounds can reduce delay and minimize cost. To address this cost and delay challenges for ML inferences in the inherent heterogeneous, resource-constrained, and distributed edge environment, we propose ProKube, which is a proactive container scaling and migration orchestrator to dynamically adjust the resources and container locations with a fair balance between cost and delay. ProKube is developed in conjunction with Google Kubernetes Engine (GKE) enabling cross-cluster migration and/ or dynamic scaling. It further supports the regular addition of freshly collected logs into scheduling decisions to handle unpredictable network behavior. Experiments conducted in heterogeneous edge settings show the efficacy of ProKube to its counterparts cost greedy (CG), latency greedy (LG), and GeKube (GK). ProKube offers 68%, 7%, and 64% SLA violation reduction to CG, LG, and GK, respectively, and it improves cost by 4.77 cores to LG and offers more cost of 3.94 to CG and GK.

The Internet of Things has emerged as a significant and influential technology in modern times. IoT presents solutions to reduce the need for human intervention and emphasizes task automation. According to a Cisco report, there were over 14.7 billion IoT devices in 2023. However, as the number of devices and users utilizing this technology grows, so does the potential for security breaches and intrusions. For instance, insecure IoT devices, such as smart home appliances or industrial sensors, can be vulnerable to hacking attempts. Hackers might exploit these vulnerabilities to gain unauthorized access to sensitive data or even control the devices remotely. To address and prevent this issue, this work proposes integrating intrusion detection systems (IDSs) with an artificial neural network (ANN) and a salp swarm algorithm (SSA) to enhance intrusion detection in an IoT environment. The SSA functions as an optimization algorithm that selects optimal networks for the multilayer perceptron (MLP). The proposed approach has been evaluated using three novel benchmarks: Edge-IIoTset, WUSTL-IIOT-2021, and IoTID20. Additionally, various experiments have been conducted to assess the effectiveness of the proposed approach. Additionally, a comparison is made between the proposed approach and several approaches from the literature, particularly SVM combined with various metaheuristic algorithms. Then, identify the most crucial features for each dataset to improve detection performance. The SSA-MLP outperforms the other algorithms with 88.241%, 93.610%, and 97.698% for Edge-IIoTset, IoTID20, and WUSTL, respectively.

The Internet of Things (IoT) is an evolving paradigm that has dramatically transformed the traditional style of living into a smart lifestyle. IoT devices have recently attained great attention due to their wide range of applications in various sectors, such as healthcare, smart home devices, smart industries, smart cities, and so forth. However, security is still a challenging issue in the IoT environment. Because of the disparate nature of IoT devices, it is hard to detect the different kinds of attacks available in IoT. Various existing works aim to provide a reliable intrusion detection system (IDS) technique. But they failed to work because of several security issues. Thus, the proposed study presents a blockchain-based deep learning model for IDS. Initially, the input data are preprocessed using min-max normalization, converting the raw input data into improved quality. In order to detect the presented attacks in the provided dataset, the proposed work introduced Gaussian mixture–fully convolutional variational autoencoder (GM-FCVAE) model. The implementation is performed in Python, and the performance of the proposed GM-FCVAE model is analyzed by evaluating several metrics. The proposed GM-FCVAE model is tested on three datasets and attained superior accuracy of 99.18%, 98.81%, and 98.4% with UNSW-NB15, CICIDS 2019, and N_BaIoT datasets, respectively. The comparison reveals that the proposed GM-FCVAE model obtained higher results than the other deep learning techniques. The outperformance shows the efficacy of the proposed study in identifying security attacks.

Traditional techniques for detecting threats in mobile edge networks are limited in their ability to adapt to evolving threats. We propose an intelligent reinforcement learning (RL)–based method for real-time threat detection in mobile edge networks. Our approach enables an agent to continuously learn and adapt its threat detection capabilities based on feedback from the environment. Through experiments, we demonstrate that our technique outperforms traditional methods in detecting threats in dynamic edge network environments. The intelligent and adaptive nature of our RL-based approach makes it well suited for securing mission-critical edge applications with stringent latency and reliability requirements. We provide an analysis of threat models in multiaccess edge computing and highlight the role of on-device learning in enabling distributed threat intelligence across heterogeneous edge nodes. Our technique has the potential, significantly enhancing threat visibility and resiliency in next-generation mobile edge networks. Future work includes optimizing sample efficiency of our approach and integrating explainable threat detection models for trustworthy human–AI collaboration.

The healthcare industry has undergone a digital transformation in recent years, with the adoption of electronic health records (EHRs) becoming increasingly prevalent. While this digitization offers various advantages, concerns regarding the security and privacy of sensitive medical data have also intensified. Data breaches and cyber-attacks targeting healthcare organizations have underscored the need for robust solutions to protect patient data. Blockchain technology has emerged as a promising solution due to its decentralized and immutable nature, which ensures secure and transparent data recording. This paper proposes a novel approach that combines blockchain with advanced encryption scheme and privacy protection technique to establish a secure and privacy protected medical data sharing environment. The proposed system consists of three phases such as initialization phase, data processing phase, and authentication phase. The hybrid Feistal-Shannon homomorphic encryption algorithm (HFSHE) is proposed to encrypt the medical data to ensure data confidentiality, integrity, and availability. Ring signature is integrated to the system to provide additional anonymity and protect the identities of the participants involved in data transactions. In addition, the smart contract developed performs authentication checks on users, generates a time seal, and verifies the ring signature. Through this enhancement, the system becomes more resilient to both external and internal threats, enhancing overall security as well as privacy. A comprehensive security analysis is conducted to compare the proposed method's performance against existing techniques. The results demonstrate the effectiveness of the proposed approach in safeguarding sensitive medical information within the blockchain ecosystem.

Besides cryptocurrencies, DLTs may be also exploited in enterprise systems operated by a consortium of organizations. Their interaction takes usually place on a permissioned blockchain network that holds a set of data to be queried frequently. In this scope, the main problem of DLTs is their unsuitability for a fast service of complex queries on those data. In order to solve this issue, many proposals dump the ledger contents onto databases that, because of their own goals and design, are already optimized for the execution of those queries. Unfortunately, many of those proposals assume that the data to be queried consist in only a block or (cryptocurrency-related) transaction history. However, those organization consortiums commonly store other structured business-related information in the DLT, and there is an evident lack of support for querying that other kind of structured data. To remedy those problems, DELTA synchronizes, with minimal overhead, the DLT state into a database, providing (1) a modular architecture with event-based handling of DLT updates that supports different DLTs and databases, (2) a transparent management, since DLT end users do not need to learn or use any new API in order to handle that synchronization (i.e., those users still rely on the original interface provided by their chosen DLT), (3) the efficient execution of complex queries on those structured data. Thus, DELTA reduces query times up to five orders of magnitude, depending on the DLT and the database, compared to queries directed to the ledger nodes.

Large language models (LLMs), with their billions to trillions of parameters, excel in natural language processing, machine translation, dialog systems, and text summarization. These capabilities are increasingly pivotal in the metaverse, where they can enhance virtual interactions and environments. However, their extensive use, particularly in the metaverse's immersive platforms, raises significant privacy concerns. This paper analyzes existing privacy issues in LLMs, vital for both traditional and metaverse applications, and examines protection techniques across the entire life cycle of these models, from training to user deployment. We delve into cryptography, embedding layer encoding, differential privacy and its variants, and adversarial networks, highlighting their relevance in the metaverse context. Specifically, we explore technologies like homomorphic encryption and secure multiparty computation, which are essential for metaverse security. Our discussion on Gaussian differential privacy, Renyi differential privacy, Edgeworth accounting, and the generation of adversarial samples and loss functions emphasizes their importance in the metaverse's dynamic and interactive environments. Lastly, the paper discusses the current research status and future challenges in the security of LLMs within and beyond the metaverse, emphasizing urgent problems and potential areas for exploration.

Achieving high-performance aggregation is essential to scaling data-parallel distributed machine learning (ML) training. Recent research in in-network computing has shown that offloading the aggregation to the network data plane can accelerate the aggregation process compared to traditional server-only approaches, reducing the propagation delay and consequently speeding up distributed training. However, the existing literature on in-network aggregation does not provide ways to deal with slower workers (called stragglers). The presence of stragglers can negatively impact distributed training, increasing the time it takes to complete. In this paper, we present Serene, an in-network aggregation system capable of circumventing the effects of stragglers. Serene coordinates the ML workers to cooperate with a programmable switch using a hybrid synchronization approach where approaches can be changed dynamically. The synchronization can change dynamically through a control plane API that translates high-level code into switch rules. Serene switch employs an efficient data structure for managing synchronization and a hot-swapping mechanism to consistently change from one synchronization strategy to another. We implemented and evaluated a prototype using BMv2 and a Proof-of-Concept in a Tofino ASIC. We ran experiments with realistic ML workloads, including a neural network trained for image classification. Our results show that Serene can speed up training by up to 40% in emulation scenarios by reducing drastically the cumulative waiting time compared to a synchronous baseline.

The applicability of the Internet of Things (IoT) cutting across different domains has resulted in newer “things” acquiring IP connectivity. These things, technically known as IoT devices, are vulnerable to diverse security threats. Consequently, there has been an exponential increase in IoT malware over the past 5 years, and securing IoT devices from such attacks is a pressing concern in the current era. However, the traditional peripheral security measures do not comply with the lightweight security requirements of the IoT ecosystem. Considering this, we propose a lightweight intrusion detection model for IoT networks (LIDM-IoT) that demonstrates similar efficiency in exposing malicious activities compared with the existing computationally expensive methods. The crux of the proposed model is that it provides efficient attack detection with lower computational requirements in IoT networks. LIDM-IoT achieves the feat through a novel unified feature selection strategy that unifies filter-based and embedded feature selection methods. The proposed feature selection strategy reduces the feature space by 94%. Also, we use only the records of a single attack type to build the model using the XGBoost algorithm. We have tested LIDM-IoT with unseen attack types to ensure its generalized behavior. The results indicate that the proposed model exhibits efficient attack detection, with a reduced feature set, in IoT networks compared with the state-of-the-art models.