Browsing by Author "Shahriar, Nashid"

Now showing 1 - 14 of 14

Open Access
A robust intrusion detection system utilizing uncertain reasoning techniques in artificial intelligence
(Faculty of Graduate Studies and Research, University of Regina, 2024-05) Singh, Harpreet; Louafi, Habib; Yao, Yiyu; Shahriar, Nashid
Network Intrusion Detection Systems (NIDS) are essential components in cybersecurity, but they face several challenges, including uncertainty and a significant computational overhead. Network attacks and unauthorized access to remote computers can be detected by NIDS. Artificial Intelligence (AI) techniques have been used to automate the intrusion detection process and reduce human intervention, thereby enhancing intrusion detection systems (IDS) performance. AI techniques, such as fuzzy logic, neural networks, and evolutionary computing can also be used in IDS. One proposed application of AI is to utilize evidential reasoning to handle uncertainty in NIDS. This approach leads to more efficient abnormality detection in user behaviour, making it a powerful tool for NIDS. This research primarily focuses on NIDS based on uncertain reasoning AI. The latter is more explainable than machine learning and deep learning approaches because it relies on well-understood principles, such as probabilities. In contrast, machine learning approaches are often considered black boxes, which makes them challenging to explain. We primarily work on real-time network traffic or packet-captured files, with the main objective of looking for attack signs of various types, using Bayesian belief networks (probabilistic graphical models). Throughout this thesis, we describe the IDS and the analysis of network traffic using a BN and Markov network (MN). BN is used to formulate the problem domain, whereas the MN is used for the inference and calculation of marginal distribution. To do so, different propagation algorithms are explored, such as Variable Elimination (VE), Lauritzen-Spiegelhalter Propagation (LS), Shafer-Shenoy Propagation (SS), and Lazy Propagation(LP). The data used in the experiments originated from the CAIDA Lab. CAIDA dataset contains network traffic packets from Distributed Denial-of-Service (DDoS) attacks. Therefore, the main objective of this thesis is to develop an uncertain reasoning-based NIDS system capable of predicting DoS/DDoS attacks with higher accuracy while reducing the computation overhead. Extensive experiments are conducted using the above-listed inference algorithms, and thus three models are trained for each protocol on the CAIDA dataset. The experiments show appealing results, measured using well-known metrics, such as Precision, Recall, and F1-Score. Overall, the SS and LP are efficient, but with minor differences.
Open Access
Active eavesdroppers detection system in multi-hop wireless sensor networks
(Faculty of Graduate Studies and Research, University of Regina, 2022-08) Abedini, Masih; Al-Anbagi, Irfan; Laforge, Paul; Shahriar, Nashid
Wireless Sensor Networks (WSNs) are vulnerable to eavesdropping attacks that endanger their privacy, confidentiality, and authenticity. As the broadcast nature of the wireless channel makes it susceptible to eavesdropping by adversaries, the detection of eavesdroppers in wireless networks can lessen the chances of more damaging attacks. Historically, researchers have attempted to reduce the risk of covert eavesdropping through the use of cryptographic protocols, information-theoretic solutions, and transmission range control. These methods are not suitable for WSNs with resource constraints. It is noteworthy that active eavesdroppers are legitimate nodes that are compromised by adversaries to eavesdrop on traffic while performing their normal responsibilities in ad-hoc networks. Detecting such malicious nodes slows the ongoing destructive attacks. In this thesis, we present a novel Active Eavesdroppers Detection (AED) system for homogeneous multi-hop WSNs. The AED system consists of two major modules: a Monitoring module and a Detection Engine module. The Monitoring module plays a vital role in the AED system to provide accurate measurements for the Detection Engine module. The Detection Engine module is provided with a lightweight detection engine module that employs the Z-test method and runs on edge devices. Regarding measurements, we first use intra-node delay measurements as the input feature of the AED system. To measure intra-node delays of nodes, the Monitoring module employs an out-of-band monitoring system using static nodes, Unmanned Aerial Vehicles (UAVs), or both of them. According to simulation results in the Cooja and MATLAB environments, the AED system can detect active eavesdroppers who relay packets to their neighbors. However, it fails to detect active eavesdroppers who do not forward packets for any reason, like placement at the network’s border. To solve this problem, we propose to use Round Trip Time (RTT) as a measurement for the AED system. The monitoring module requests nodes for responses, and the AED’s detection engine can detect active eavesdroppers in WSNs based on the response delay. We focus on three potential monitoring systems for this measurement: static monitoring nodes, UAV-based monitoring, and neighborhood monitoring. To find the optimal places for static monitoring nodes, we utilize a Genetic Algorithm(GA), and to find the path of flight for UAVs for measuring RTT, we use Hamiltonian path planning. The simulation results indicate that the RTT-based AED system can detect active eavesdroppers regardless of their locations, with a high detection rate (≥ 90%) and a low false-positive rate (≤ 5%) and outstanding performance (AUC ≈ 0.97). In addition, we analyze and discuss the network overhead, advantages, and disadvantages of the in-band neighborhood monitoring system.
Open Access
Deep transfer learning-based DDoS attack detection in 5G and beyond networks
(Faculty of Graduate Studies and Research, University of Regina, 2024-09) Farzaneh, Behnam; Shahriar, Nashid; Louafi, Habib; Yao, Yiyu
Network slicing is a crucial technology for enabling 5G and beyond mobile networks which support a wide range of new services such as Enhanced Mobile Broadband (eMBB), Ultra-Reliable and Low Latency Communications (URLLC), and Massive Machine-Type Communications (mMTC) on the same physical infrastructure. However, this technology also makes networks more vulnerable to cyber threats, especially Distributed Denial-of-Service (DDoS) attacks. These kinds of attacks can degrade service quality by overwhelming essential network functions necessary for the seamless operation of network slices. To address this issue, an Intrusion Detection System (IDS) is needed to protect against various DDoS attacks. A promising solution is the use of Deep Learning (DL) models to detect potential DDoS attacks, a method already proving effective in the field. However, DL models require large amounts of labeled data for effective training, which are often scarce in operational networks. To address this, Transfer Learning (TL) techniques can be used by transferring knowledge from previously trained models to a target domain with limited labeled data. In this thesis, Bidirectional Long Short-Term Memory (BiLSTM), Convolutional Neural Network (CNN), Residual Network (ResNet), and Inception are used as base models for Deep Transfer Learning (DTL) methods that look into how they can improve DDoS attack detection in 5G networks. A comprehensive dataset generated in our 5G network slicing testbed, which contains both benign and various DDoS attack traffic, serves as the source dataset for DTL. After learning features, patterns, and representations from the source dataset, the base models are fine-tuned using different TL processes on a target DDoS attack dataset. The 5G-NIDD (5G Network Intrusion Detection Dataset), which has limited annotated traffic from several DDoS attacks generated in a real 5G network, is chosen as the target dataset. The results indicate that the proposed DTL models improve the detection of various DDoS attacks in the 5G-NIDD dataset compared to models without TL. Specifically, the BiLSTM and Inception models are identified as the top performers. BiLSTM shows an improvement of 13.90%, 21.48%, and 12.22% in terms of accuracy, recall, and F1-score, respectively, while Inception demonstrates a 10.09% increase in precision compared to models not using TL.
Open Access
Detection and monitoring of ransomware attacks using machine learning and Deep Learning
(Faculty of Graduate Studies and Research, University of Regina, 2024-01) Al Ahasan, Md Abdullah; Shahriar, Nashid; Sadaoui, Samira; Louafi, Habib
This thesis presents a comprehensive investigation into the threat of ransomware and explores recent advancements in detection techniques. With the rise in the popularity of ransomware, a unique ecosystem of cybercriminals has emerged, leveraging encryption technology, anonymous cybersecurity, and easily accessible ransomware code. To address this growing concern, this thesis emphasizes the need for a machine learning (ML) and Deep Learning (DL) solution to detect ransomware attacks. Additionally, the study introduces the utilization of Software Defined Networking (SDN) combined with ML and DL for enhanced ransomware detection and mitigation. In our pursuit of demonstrating ransomware detection capabilities, we introduce an architectural design aimed at providing a highly efficient solution for proactively countering ransomware attacks. Experimental results demonstrate the efficacy of the developed mechanism in promptly detecting and preventing the spread of ransomware. Moreover, considering the significant damage caused by ransomware attacks, the thesis explores the training and testing of various ML and DL models for ransomware detection. A novel and flexible ransomware detection model is proposed, achieving good accuracy and F1-scores on different domains of the dataset. The proposed method is applicable to any domain of network traffic analysis data. In the context of the dynamic malware landscape, this thesis explores the detection of ransomware attacks by monitoring network traffic between infected computers and command and control servers. By extracting high-level flow features and utilizing a random forest classifier, a flow-based detection method is developed to identify and classify ransomware without deep packet inspection. The proposed solution demonstrates a high detection rate and low false negative rate, proving its feasibility and accuracy. The proposed approach significantly improves detection accuracy, making it effective for detecting both ransomware and specific types of malware. The method achieves feature reduction and quick convergence means that our method is attributed to its adept feature reduction capabilities, showcasing its efficiency and efficacy.
Open Access
Detection of DoS and DDoS attacks on 5G network slices using deep learning approach
(Faculty of Graduate Studies and Research, University of Regina, 2023-09) Khan, Md. Sajid; Shahriar, Nashid; Yao, Yiyu; Louafi, Habib; Morgan, Yasser
A new degree of connectedness and interaction has been introduced by the development of 5G networks. By dividing a physical network into several logical networks, 5G network slicing is a special feature that gives network operators the ability to allocate specific resources and services to various applications and customers. However, 5G network slicing is susceptible to cyberattacks, particularly Denial-of-Service (DoS) or Distributed Denial-of-Service (DDoS) attacks, just like any other network. Such attacks can have a significant negative effect on network performance, degrading services and reducing the availability of slices. The primary objective of this thesis is to examine the impact of DoS/DDoS attacks on 5G network slicing and their potential to disrupt the performance of legitimate users and slice availability. Additionally, a novel dataset specifically tailored to DoS/DDoS attacks in 5G network slicing is generated, as there is no available dataset based on a 5G network slice. Through extensive research, key features relevant to DoS/DDoS attacks are identified and prioritized. To categorize and detect different types of DoS/DDoS attacks, two deep learning techniques, namely the convolutional neural network (CNN) and the Bidirectional Long Short-Term Memory (BLSTM) models, are employed. These models not only utilize the newly created dataset but also enable comparison with existing datasets to assess their effectiveness. This thesis emphasizes how crucial it is to create strong security measures to guard against DoS/DDoS attacks on 5G network slicing. A step in the right direction toward reaching this goal is the construction of a deep learning model for the classification, detection, and production of a new dataset specifically for 5G network slicing. To keep enhancing the security and stability of 5G network slicing, more study in this area will be required. The results indicate that the proposed models have a high accuracy rate of 99.96% in distinguishing different types of DoS/DDoS attacks within the networking slice environment. This achievement is noteworthy as it pertains to a novel context. Additionally, the newly developed models exhibit comparable performance in terms of other confusion metrics. To verify the research outcome, some well-known data sets are used to show the results.
Open Access
Development of Failure and Consequence Analysis Frameworks for Natural Gas Transmission Pipeline
(Faculty of Graduate Studies and Research, University of Regina, 2020-12) Ahmed, Sk Kafi; Kabir, Golam; Aroonwilas, Adisorn; Almehdawe, Eman; Shahriar, Nashid
In our daily activities, natural gas is an essential part for cooking, heating, electricity production as well as manufacturing activities. In last two decades, the natural gas consumption rate has been increasing exponentially in every corner of the world. Subsequently, any natural gas pipeline failure can lead human fatalities, financial losses, and interruption of manufacturing activities as well as environmental impact. To overcome this situation or reduce the losses, a risk analysis method can play a very significant role the potential threats. The main objective of this research is to identify the natural gas pipeline failure causes, influences and consequences analysis to improve the public life safety, security and to reduce the potential environmental hazards. In first part of this study, a Rough Analytic Hierarchy Process (AHP), Rough Decision-making Trial and Evaluation Laboratory (DEMATEL), Interpretive Structural Modeling (ISM) and Bayesian Belief Network (BBN) are integrated to analyze the failure causes and influences of natural gas transmission pipeline. After that, to identify the consequence of the natural gas transmission pipeline failure, the potential threat zone simulation is performed using ALOHA software and a BBN-based consequence model is developed. The outcome of this thesis will help the energy provider and government agency to take necessary safety precaution plan as well as forecast the crisis management budget.
Open Access
Discretization of nature-inspired techniques for combinatorial problems
(Faculty of Graduate Studies and Research, University of Regina, 2023-11) Sadeghilalimi, Mehdi; Mouhoub, Malek; Louafi, Habib; Shahriar, Nashid; Volodin, Andrei; Bagheri, Ebrahim
Scientists across various domains like scheduling, computational biology, and machine learning face constraint-solving and optimization problems. Classical systematic and mathematical methods often fall short of providing suitable solutions for such complex problems, leading to the introduction of metaheuristic algorithms. These algorithms exhibit diverse characteristics and can effectively address specific optimization problems. The primary motivation is to develop robust metaheuristics that can efficiently handle scaling problems. However, one challenge with metaheuristics is their immature convergence. In the context of Constraint Satisfaction Problems (CSPs), a framework applicable to numerous real-world problems, metaheuristics play a significant role. To address these objectives and challenges, this thesis investigates the applicability of metaheuristics, including the Whale Optimization Algorithm (WOA), Genetic Algorithm (GA), Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), and Water Cycle Algorithm (WCA). More precisely, we propose a discrete version of nature-inspired techniques for solving the Electricity Technician Dispatch Problems (ETDP), the Nurse Scheduling Problem (NSP), and the Task Scheduling Problem in Mobile Cloud Computing. We also propose a discrete version of WOA for Type 2 Diabetes Diagnosis. Experimentation showcases the efficiency of the proposed techniques in finding a good trade-off between running time and the quality of the solution returned.
Open Access
DTL-IDS: Deep Transfer Learning-based Intrusion Detection System in 5G Networks
(2023-11-02) Farzaneh, Behnam; Shahriar, Nashid; Al Muktadir, Abu Hena; Towhid, Md. Shamim
Open Access
A Dynamic Predictive Analysis on Gas Pipeline Failure Consequence Using Bayesian Network
(Faculty of Graduate Studies and Research, University of Regina, 2021-09) Aalirezaei, Armin; Kabir, Golam; Henni, Amr; Khondoker, Mohammad; Shahriar, Nashid
Nowadays, due to the economical, environmental, and social concerns which are the three pillars of sustainable development objectives, natural gas is considered as one of the most important and critical energies. Natural gas pipeline failure and related losses is a kind of devastating disaster because the threats of natural gas failure consequences may provide a huge extension that can simply lead to cascading disasters. Therefore, maintenance and repair of the deteriorating infrastructure like buried gas pipelines have been considered significant considerations amongst stakeholders and researchers in recent years. In this thesis, a dynamic Bayesian network (BN) was used to investigate natural gas pipeline network failure consequences in a probabilistic way. Seven parent nodes including, age, diameter, length, depth, population, time of occurrence, and land-use have been considered for the developed model. Twelve consequences factors also were recognized based on the literature review and expert’s opinion. Production loss, asset loss, environmental loss, and social and reputational loss were also measured as the overall losses due to the failure consequences. The proposed model can manage both static and dynamic systems employing quantitative and/or qualitative data. Methods of the extreme-condition test, scenario analysis, sensitivity analysis, and partial validation test were utilized to validate the model and to detect the key and leading parent nodes and their effect on the overall loss index. The developed static and dynamic BN models will help the decision-makers to manage and prioritize their asset effectively. To demonstrate the applicability and effectiveness of the developed model, the gas pipeline network of the City of Regina, SK, in Canada is studied. The results show that age and diameters are the most important and sensitive parameters amongst others which need to take into consideration for disaster reaction decision-making and loss prevention.
Open Access
An efficient Internet of Things (IoT) device fingerprinting approach using machine learning
(Faculty of Graduate Studies and Research, University of Regina, 2022-10) Osei, Richmond; Louafi, Habib; Mouhoub, Malek; Shahriar, Nashid; Al-Anbagi, Irfan
Internet of Things (IoT) usage is steadily becoming a way of life. IoT devices have applications in smart homes, factories, and farming. Thousands of IoT devices are hooked up to the cyberspace each day. However, the increased usage of IoT devices comes with many security concerns due to their small and constrained build-up; consequently, perpetrators could target an IoT device as an access point to attack the whole network. Further, the number of IoT device manufacturers keeps expanding by the day, although a high number of these manufacturers have less or little knowledge about IoT security. Therefore, a comprised IoT device in a network presents a vulnerability for an attacker to exploit the network. Notwithstanding, relying on the most typical approaches to securing IoT devices is becoming increasingly complex and less prone to attackers. To determine the identities (fingerprints) and nature of these devices, Mobile Network Operators (MNOs) often analyze the traffic generated by these devices (legitimate, faulty, or malicious). When a new IoT device is attached to the network and is compromised, the fingerprinting process takes quick action to determine the real identity of the device. This thesis presents an effective way to fingerprint IoT devices using machine learning. Many security benefits come with fingerprinting IoT devices on a network, including automatic vulnerability patching, the introduction of behaviour-based anomaly detection, and dynamic attack mitigation. A better way of securing IoT devices is by using a machine-learning algorithm to fingerprint IoT devices in the network accurately. This process should be achieved using minimal features to reduce the amount of data to process, which is critical in real-time prediction applications. Focusing on the optimal (minimal) features of IoT devices requires accurately fingerprinting the device. In this study, new features were generated from the original features of an IoT dataset that contains IoT network traffic. These new features, extracted using reduction methods, were then trained on selected machine learning algorithms, and the prediction quality of each model was analyzed. The obtained results are appealing. Indeed, with only five (5) features instead of the 17 features from the original dataset, the proposed solution is capable of accurately fingerprinting IoT devices with an accuracy of 97%, as measured using the Precision, Recall, and Harmonic mean metrics.
Open Access
Feistel network-based prefix-preserving network trace anonymization
(Faculty of Graduate Studies and Research, University of Regina, 2022-05) Dandyan, Shaveta; Louafi, Habib; Sadaoui, Samira; Shahriar, Nashid
Network traces represent a critical piece of data for the network security analysts to ensure the security of the data and detecting/correcting network issues. Due to lack of expertise, companies are forced to outsource their network traces to third parties to perform analytics on the traces and provide security feedbacks and recommendations. In fact, outsourcing network traces to third party analysts for monitoring and analytics is a common service requested by companies. However, these companies are reluctant to share their network traces, as they comprise sensitive information (e.g., IP addresses), which may be exploited for attacks. Therefore, such sensitive information needs to be hidden before outsourcing the network traces. Network trace anonymization is a solution that provides the privacy of the data and preserving its utility. The latter is important for the analytics, that is, the data needs to be anonymized (some information are changed), in such a way the essence of the data remains valid. Otherwise, the analytics provided by third party analysts cannot reflect the actual state of the network. Existing solutions, such as CryptoPAN, preserves the data utility (by preserving the IP prefixes), but are vulnerable to semantic attacks. In this thesis, we propose an anonymization solution, which is based on the Feistel network and preserves the data privacy and utility at the same time. Besides, the proposed solution requires less computation and resources, since it is based on the Feistel network that guarantees the anonymization and de-anonymization with the same architecture. The Feistel network is widely used in cryptography because of its flexible structure. Thus, in this thesis, we adapt it to perform both the anonymization and de-anonymization. We validate our solution using Kddcup99 dataset, from which distinct IP addresses have been filtered to better measure the data leakage (dual of privacy) provided by our solution. The obtained results show that the proposed solution provides consistent results throughout the different traces under the same experimental parameters. We evaluate the security of our solution using the avalanche property, which is widely used to measure the security of encryption systems. Moreover, the efficacy of our solution is evaluated against Injection attacks. Overall, the obtained results, avalanche property and resistance to Injection attacks, are appealing.
Open Access
Prediction of Waste Generation and Disposal Using System Dynamics Modeling
(Faculty of Graduate Studies and Research, University of Regina, 2022-01) Eslami, Sanaalsadat; Kabir, Golam; Ng, Tsun Wai Kelvin; Khondoker, Mohammad; Shahriar, Nashid
Solid waste management plays an important role in protecting the environment, the health system and minimizing biological threats, so serious actions need to be taken to control and manage these threats on the environment. Different situations such as seasonal changes and cities policies have had effects on solid waste generation and disposal behaviors, therefore the proper prediction of the amount of waste disposal is needed to improve the waste management system. In this thesis, the system dynamics (SD) model is used to predict the amount of waste generation and disposal in Canada. Different types of wastes are considered in the SD model. Gross domestic product (GDP) and population are the socio-economic variables that have a significant effect on the amount of waste generation and disposal in the SD model. The developed model can show the relationship between variables to predict the amount of waste generation and disposal, so by using the SD model with a better attitude, governments and organizations can make decisions to design an appropriate model for waste management and apply it in the system. In this thesis, the most significant contribution is using effective variables such as GDP, population, and education to predict the amount of waste generation and disposal seasonally and yearly. The results show that the increasing and decreasing trends in the amount of waste disposal from the prediction value depend on the variables in the model. Population and GDP are the main effective variables in the SD model, in which waste disposal will increase in the future as the population increases. Therefore, by the effect of various policies and conditions, there will be different trends in waste management.
Open Access
A study of factors influencing cyberattacks
(Faculty of Graduate Studies and Research, University of Regina, 2022-12) Ikuomenisan, Gbenga Tola; Morgan, Yasser; Yow, Kin-Choong; Shirif, Ezeddin; Shahriar, Nashid
There has been a substantial global cyberattack evolution and Internet Use revolution during the current Coronavirus pandemic, consequently forcing countries across the globe to begin reviewing their national cybersecurity strategies. However, to adequately minimize the risks of future attacks and associated losses, new and/or contemporary cyberattack studies are required, particularly at the country level. This research examines the chances of attack initiation, during the present Coronavirus pandemic, indicated in this study by Country-Level Cyberattacks (CLCA) and how it is affected by Country-Level Internet Use (CLIU), a technological Country-Level Influential Factor (CLIF). It presents a quantitative, nonexperimental, correlational approach that examines CLIU and CLCA indexes. Three omnibus research questions are raised to quantify and evaluate how International Internet Users (IIU), International Internet Subscriptions (IIS), and International Internet Bandwidth (IIB), as predictor variables, individually correlates with International Source Addresses (ISA). Corresponding datasets from the Economist Intelligence Unit (EIU), International Telecommunication Union (ITU), and the Hornet Honeypot Project (HHP) are analyzed using correlational tests and regression models, and interpretation and inferences are drawn, using a measure of association and the frequentist methods. In line with existing studies, data analysis results show a strong positive correlation and predictive power between CLIU and CLCA indexes (R2(%) > 50) suggesting that countries having higher CLIU are more likely to have higher chances of attack initiation or attacker infestation. Contrary to existing studies, the IIB (R2(%) > 70) is found to very strongly influence a CLCA more than the IIU and IIS. Hence, relevant digital-capacity indexes are stronger than others, indicating a more decisive influence of particular indexes on the relationship between CLIU and CLCA. Additionally, the BRIC economic countries are observed to be true outliers with China having significant eccentricities. China’s IIU, IIB, and IIS-related chances of attack originations or attacker infestation compared to other countries are about 32, 32, and 43 Average Euclidean Distance (AED) units apart respectively, suggesting that China is a peculiar country that demonstrates a rare combination of CLIF hardly found in many countries and therefore, should be given special policy considerations. The findings in this study provide new insight on how CLIU interrelates with CLCA during the present Coronavirus pandemic and therefore have significant implications for research and practice. To cybersecurity researchers, practitioners, policymakers, and equipment manufacturers, this present study can help in identifying where to anticipate a higher density of attack initiations during the current pandemic. It can help them in developing models that can predict or track countries having high chances of attack initiations and attacker infestations, which is useful in many areas, such as in development of intelligent security fences and building of well-informed cybersecurity policies, strategies and bilateral agreements. Cybersecurity researchers can use this study as a related literature and a basis for future works.
Open Access
Trust-aware virtual network embedding algorithms for wireless sensor networks
(Faculty of Graduate Studies and Research, University of Regina, 2024-06) Rezaeimoghaddam, Parinaz; Al-Anbagi, Irfan; Bais, Abdul; Zhang, Lei; Shahriar, Nashid; Yassine, Abdulsalam
Network virtualization (NV) in wireless sensor networks (WSNs) enables the utilization of their shared sensing capabilities. Efficient assignment of WSN resources to maximize the infrastructure provider’s revenue can be achieved by virtual network embedding (VNE) while considering the quality of information (QoI), quality of service (QoS), and wireless interference handling constraints. Improving the acceptance rate of VNE is essential because the more the virtual network requests (VNRs) can be mapped onto the substrate network, the more revenue they will generate for the infrastructure provider. However, the shared and complex nature of VNE exposes WSNs to security risks. In this thesis, we apply security constraints and address the trust-aware VNE problem with different algorithms to maximize the VNR acceptance rate while minimizing the cost for WSNs. This research includes four main objectives. In the first objective, we develop a novel centralized trust-aware virtual wireless sensor network (TA-VWSN) algorithm to improve QoI, QoS, and security and enhance the average network throughput, measurement error efficiency, and processing time when the trust attributes are assigned, making the VNE algorithm more practical. Since centralized algorithms suffer from scalability issues, in the second objective, we design a novel distributed trust-aware virtual wireless sensor networks (DTA-VWSN) algorithm by using multiagent systems (MAS) to scale these algorithms to network size. Our heuristic TA-VWSN and DTA-VWSN algorithms achieve a high-quality sub-optimal solution in a real-time manner, enabling us to investigate the tradeoff between solution quality and search time. However, the heuristic algorithms use manual embedding rules, which are incompatible with actual VNE situations. Therefore in the third objective, we develop a novel reinforcement learning-based (RL-based) trust-aware virtual wireless sensor network (RLT-VWSN) algorithm by employing a policy network and extracting attributes of the substrate nodes to get the mapping probability of each one. This algorithm is superior to heuristic algorithms in terms of VNR acceptance rate and cost. In the fourth objective, our first approach addresses node failures during mapping with a pioneering heuristic— the survivable trust-aware virtual wireless sensor networks (STA-VWSN) algorithm. It employs a failure recovery procedure, prioritizing nodes based on the technique for order of preference by similarity to ideal solution (TOPSIS) method. In the second approach, we enhance WSN substrate resilience against individual node and link failures through the development of the survivable and trust-aware reinforcement learning-based virtual network embedding for WSNs (SRLT-VWSN) algorithm. Utilizing the deep Q-Learning (DQL) method, this algorithm ensures end-to-end failure recovery and improves physical resource utilization intelligently.