P2P is increasingly gaining its popularity for streaming multimedia contents. The architecture of streaming has shifted from traditional client server architecture to P2P architecture. Although it is scalable and robust it faces its own challenges and problems such as churn. In tree topology frequent joining and leaving of users in search for better quality and reliable streaming makes the P2P network instable. This thesis provides an effective approach to achieve a resilient network for streaming. Relying on a single tree to receive data from single parent may leave the user deprived of getting the data if any of its ancestors leaves the network. Therefore we present an ideal solution to this problem by introducing a backup tree for the existing base tree. The backup tree is constructed based on parameter such as bandwidth and delay. In case of failure of a node, its children along the tree receive the data from the nodes of backup tree. We present an efficient algorithm for the construction of base tree as well as the backup tree which are based on normalization of two entities of nodes: bandwidth and delay. Through mathematical formulation and experimental setups we show that introducing a backup tree for an existing base tree can help provide resilience to the network. / UOIT

Over the last few years, peer-to-peer (P2P) applications have relentlessly grown to represent a formidable component of Internet traffic. In contract to P2P networks witch used well-defined port number, current P2P applications have use of arbitrary ports. As P2P applications continue to evolve, robust and effective methods are methods are needed for P2P traffic identification. Many P2P applications are bandwidth-intensive. Understanding the Internet traffic profile is important for several reasons, including traffic engineering, network service pricing. In this Thesis, we integrated port-based method into original Classifier which is using content-based method only. Therefore, we can improve the recognition rate for Classifier and identify more applications. We also verified our Classifier recognition rate by using the results of Service Control Engine.

Peer-to-Peer (P2P) Networks have been presenting many fascinating capabilities to the internet since their inception, which has made and is still gathering so much interest. As a result, it is being used in many domains, particularly in transferring a large amount of data, which is essential for modern computing needs. A P2P network contains many independent nodes to form a highly distributed system. These nodes are used to exchange all kinds of files without using a single server as in a Client-Server architecture. Such types of files make the network highly vulnerable to malicious attackers. Nevertheless, P2P systems have become susceptible to different malicious attacks due to their widespread usage, including the threat of sharing malware and other dangerous programs, which can be significantly damaging and harmful. A significant obstacle with the current P2P network traffic monitoring and analysis involves many newly emerging P2P architectures possessing more intricate communication structures and traffic patterns than the traditional client-server architectures. The traffic volume generated by these networks, such as uTorrent, Gnutella, Ares, etc., was once well over half of the total internet traffic. The dynamic use of port numbers, multiple sessions, and other smart features of these applications complicate the characterization of current P2P traffic. Transport-level traffic identification is a preliminary but required step towards traffic characterization, which this thesis addresses. Therefore, a novel detection mechanism that relies on transport-level traffic characterization has been presented for P2P network investigation The importance of the investigation necessitates the formalization of frameworks to leverage the integration of forensics standards and accuracy to provide integrity to P2P networks. We employed the standard Analysis, Design, Development, Implementation, and Evaluation (ADDIE) model to aid a credible digital investigation. We considered the ADDIE model for validation as a standard digital forensic model for P2P network investigations using the United States’ Daubert Standard, the United Kingdom's Forensic Science Regulator Guidance – 218 (FSR-G-218), and Forensic Science Regulator Guidance – 201 (FSR-G-201) methodologies. The solution was evaluated using a realistic P2P investigation and showed accurate load distribution and reliable digital evidence. / Petroleum Technology Development Fund (PTDF) Nigeria

yes / Peer-to-Peer (P2P) network is increasingly becoming the most important means of trading content throughout the last years due to the constant evolvement of the cyber world. This popularity made the P2P network susceptible to the spread of malware. The detection of the cause of malware propagation is now critical to the survival of P2P networks. This paper offers a review of the current relevant mathematical propagation models that have been proposed to date to predict the propagation behavior of a malware in a P2P network. We analyzed the models proposed by researchers and experts in the field by evaluating their limitations and a possible alternative for improving the analysis of the expected behavior of a malware spread.

Serverless Edge Computing is an extension of the serverless computing paradigm that enables the deployment and execution of modular software functions on resource-constrained edge devices. However, it poses several challenges due to the edge network's dynamic nature and serverless applications' latency constraints. In this work, we introduce EdgeFn, a lightweight distributed data store for the serverless edge computing system. While serverless comput- ing platforms simplify the development and automated management of software functions, running serverless applications reliably on resource-constrained edge devices poses multiple challenges. These challenges include a lack of flexibility, minimum control over management policies, high data shipping, and cold start latencies. EdgeFn addresses these challenges by providing distributed data storage for serverless applications and allows users to define custom policies that affect the life cycle of serverless functions and their objects. First, we study the challenges of existing serverless systems to adapt to the edge environment. Sec- ond, we propose a distributed data store on top of a Distributed Hash Table (DHT) based Peer-to-Peer (P2P) Overlay, which achieves data locality by co-locating the function and its data. Third, we implement programmable callbacks for storage operations which users can leverage to define custom policies for their applications. We also define some use cases that can be built using the callbacks. Finally, we evaluate EdgeFn scalability and performance using industry-generated trace workload and real-world edge applications. / Master of Science / Serverless Edge Computing is an extension of the serverless computing paradigm that enables the deployment and execution of modular software functions on resource-constrained edge devices. However, it poses several challenges due to the edge network's dynamic nature and serverless applications' latency constraints. In this work, we introduce EdgeFn, a lightweight distributed data store for the serverless edge computing system. While serverless comput- ing platforms simplify the development and automated management of software functions, running serverless applications reliably on resource-constrained edge devices poses multiple challenges. These challenges include a lack of flexibility, minimum control over management policies, high data shipping, and cold start latencies. EdgeFn addresses these challenges by providing distributed data storage for serverless applications and allows users to define custom policies that affect the life cycle of serverless functions and their objects. First, we study the challenges of existing serverless systems to adapt to the edge environment. Sec- ond, we propose a distributed data store on top of a Distributed Hash Table (DHT) based Peer-to-Peer (P2P) Overlay, which achieves data locality by co-locating the function and its data. Third, we implement programmable callbacks for storage operations which users can leverage to define custom policies for their applications. We also define some use cases that can be built using the callbacks. Finally, we evaluate EdgeFn scalability and performance using industry-generated trace workload and real-world edge applications.

This thesis presents an innovative strategy for distributed Graph Neural Network (GNN) training, leveraging a peer-to-peer network of heterogeneous edge devices interconnected through a Distributed Hash Table (DHT). As GNNs become increasingly vital in analyzing graph-structured data across various domains, they pose unique challenges in computational demands and privacy preservation, particularly when deployed for training on edge devices like smartphones. To address these challenges, our study introduces the Adaptive Load- Balanced Partitioning (ALBP) technique in the GraphDHT system. This approach optimizes the division of graph datasets among edge devices, tailoring partitions to the computational capabilities of each device. By doing so, ALBP ensures efficient resource utilization across the network, significantly improving upon traditional participant selection strategies that often overlook the potential of lower-performance devices. Our methodology's core is weighted graph partitioning and model aggregation in GNNs, based on partition ratios, improving training efficiency and resource use. ALBP promotes inclusive device participation in training, overcoming computational limits and privacy concerns in large-scale graph data processing. Utilizing a DHT-based system enhances privacy in the peer-to-peer setup. The GraphDHT system, tested across various datasets and GNN architectures, shows ALBP's effectiveness in distributed GNN training and its broad applicability in different domains and structures. This contributes to applied machine learning, especially in optimizing distributed learning on edge devices. / Master of Science / Graph Neural Networks (GNNs) are a type of machine learning model that focuses on analyzing data structured like a network, such as social media connections or biological systems. These models can help identify patterns and make predictions in various tasks, but training them on large-scale datasets can require significant computing power and careful handling of sensitive data. This research proposes a new method for training GNNs on small devices, like smartphones, by dividing the data into smaller pieces and using a peer-to-peer (p2p) network for communication between devices. This approach allows the devices to work together and learn from the data while keeping sensitive information private. The main contributions of this research are threefold: (1) examining existing ways to divide network data and how they can be used for training GNNs on small devices, (2) improving the training process by creating a localized, decentralized network of devices that can communicate and learn together, and (3) testing the method on different types of datasets and GNN models, showing that it works well across a variety of situations. To sum up, this research offers a novel way to train GNNs on small devices, allowing for more efficient learning and better protection of sensitive information.

In recent years, Internet-of-Things (IoT) devices generate a large amount of personal data. However, due to the privacy concern, collecting the private data in cloud centers for training Machine Learning (ML) models becomes unrealistic. To address this problem, Federated Learning (FL) is proposed. Yet, central bottleneck has become a severe concern since the central node in traditional FL is responsible for the communication and aggregation of mil- lions of edge devices. In this paper, we propose Scalable Federated Learning via Distributed Hash Table Based Overlays for network (Scaled) to conduct multiple concurrently running FL-based applications over edge networks. Specifically, Scaled adopts a fully decentral- ized multiple-master and multiple-slave architecture by exploiting Distributed Hash Table (DHT) based overlay networks. Moreover, Scaled improves the scalability and adaptability by involving all edge nodes in training, aggregating, and forwarding. Overall, we make the following contributions in the paper. First, we investigate the existing FL frameworks and discuss their drawbacks. Second, we improve the existing FL frameworks from centralized master-slave architecture by using DHT-based Peer-to-Peer (P2P) overlay networks. Third, we implement the subscription-based application-level hierarchical forest for FL training. Finally, we demonstrate Scaled's scalability and adaptability over large scale experiments. / Master of Science / In recent years, Internet-of-Things (IoT) devices generate a large amount of personal data. However, due to privacy concerns, collecting the private data in central servers for training Machine Learning (ML) models becomes unrealistic. To address this problem, Federated Learning (FL) is proposed. In traditional ML, data from edge devices (i.e. phones) should be collected to the central server to start model training. In FL, training results, instead of the data, are collected to perform training. The benefit of FL is that private data can never be leaked during the training. However, there is a major problem in traditional FL: a single point of failure. When power to a central server goes down or the central server is disconnected from the system, it will lose all the data. To address this problem, Scaled: Scalable Federated Learning via Distributed Hash Table Based Overlays is proposed. Instead of having one powerful main server, Scaled launches many different servers to distribute the workload. Moreover, since Scaled is able to build and manage multiple trees at the same time, it allows multi-model training.

Les services Internet populaires modèlent et remodèlent fondamentalement les moyens traditionnels de communication des personnes, ayant ainsi un impact majeur sur leur vie sociale. Deux des services Internet très populaires avec cette caractéristique sont les Réseaux sociaux en ligne (OSN) et les systèmes Peer-to-Peer (P2P). Les ONS fournissent un environnement virtuel où les gens peuvent partager leurs informations et leurs intérêts tout en étant en contact avec d'autres personnes. D'autre part, les systèmes P2P, qui sont toujours l'un des services populaires avec une grande proportion de l'ensemble du trafic Internet, offrent une occasion en or pour leurs clients de partager un type de contenu différent, y compris le contenu protégé. En dehors de l'énorme popularité des ONS et des systèmes de P2P parmi les utilisateurs réguliers, ils sont intensivement utilisés par les professionnels (grandes entreprises, politiciens, athlètes, célébrités en cas d'ONS et éditeurs de contenu professionnels en cas de P2P) afin d'interagir avec les gens à des fins différentes (campagnes marketing, les commentaires des clients, amélioration de la réputation publique, etc.) Dans cette thèse, nous caractérisons le comportement des utilisateurs réguliers et professionnels dans les deux services mentionnés populaires (ONS et P2P) en termes de stratégies de publication, de consommation de contenu et d'analyse comportementale. À cette fin, cinq de nos études menées sont présentées dans ce manuscrit comme suit: - "L'évolution des contenus multimédias", qui présente une analyse approfondie sur l'évolution du contenu multimédia disponible en BitTorrent en se concentrant sur quatre mesures pertinentes à travers différentes catégories de contenu : la disponibilité du contenu, la popularité du contenu, la taille de contenu et les commentaires de l'utilisateur - "La réaction des utilisateurs professionnels face aux actions de lutte contre le piratage", en examinant l'impact de deux grandes actions de lutte contre le piratage - la fermeture de Megaupload et la mise en œuvre de la loi anti-piratage française (HADOPI) - sur le comportement des publicateurs professionnels dans le plus grand portail de BitTorrent qui sont les principaux fournisseurs de contenu en ligne protégé. - "La quantité d'informations divulguées sur Facebook", en enquêtant sur l'exposition publique des profils utilisateurs, une grande base de données comprenant un demi-million d'utilisateurs réguliers. - "Les utilisateurs professionnels Cross Posting Activity», en analysant le modèle de publication des utilisateurs professionnels de mêmes informations sur trois grands ONS à savoir Facebook, Google+ et Twitter. - "Les stratégies des utilisateurs professionnels dans les ONS", où nous étudions la stratégie globale d'utilisateurs professionnels par secteur (par exemple, les entreprises de voitures, l'habillement, politiques, etc.) sur Facebook, Google+ et Twitter. Les résultats de cette thèse fournissent une vision d'ensemble pour comprendre certains aspects comportementaux importants de différents types d'utilisateurs des services Internet populaires et ces contributions peuvent être utilisées dans divers domaines (par exemple analyse de campagne marketing et publicité, etc.) et les différentes parties peuvent bénéficier des résultats et des méthodologies mises en œuvre telles que les FAI et les propriétaires des services pour leur planification ou l'expansion des services actuels à venir, ainsi que les professionnels pour accroître leur succès sur les médias sociaux / Popular Internet services are fundamentally shaping and reshaping traditional ways of people communication, thus having a major impact on their social life. Two of the very popular Internet services with this characteristic are Online Social Networks (OSNs) and Peer-to-Peer (P2P) systems. OSNs provide a virtual environment where people can share their information and interests as well as being in contact with other people. On the other hand, P2P systems, which are still one of the popular services with a large proportion of the whole Internet traffic, provide a golden opportunity for their customers to share different type of content including copyrighted content. Apart from the huge popularity of OSNs and P2P systems among regular users, they are being intensively used by professional players (big companies, politician, athletes, celebrities in case of OSNs and professional content publishers in case of P2P) in order to interact with people for different purposes (marketing campaigns, customer feedback, public reputation improvement, etc.). In this thesis, we characterize the behavior of regular and professional users in the two mentioned popular services (OSNs and P2P systems) in terms of publishing strategies, content consumption and behavioral analysis. To this end, five of our conducted studies are presented in this manuscript as follows: - “The evolution of multimedia contents", which presents a thorough analysis on the evolution of multimedia content available in BitTorrent by focusing on four relevant metrics across different content categories: content availability, content popularity, content size and user's feedback. - “The reaction of professional users to antipiracy actions", by examining the impact of two major antipiracy actions, the closure of Megaupload and the implementation of the French antipiracy law (HADOPI), on professional publishers behavior in the largest BitTorrent portal who are major providers of online copyrighted content. - “The amount of disclosed information on Facebook", by investigating the public exposure of Facebook users' profile attributes in a large dataset including half million regular users. - “Professional users Cross Posting Activity", by analyzing the publishing pattern of professional users which includes same information over three major OSNs namely Facebook, Google+ and Twitter. - “Professional Users' Strategies in OSNs", where we investigate the global strategy of professional users by sector (e.g., Cars companies, Clothing companies, Politician, etc.) over Facebook, Google+ and Twitter. The outcomes of this thesis provide an overall vision to understand some important behavioral aspects of different types of users on popular Internet services and these contributions can be used in various domains (e.g. marketing analysis and advertising campaign, etc.) and different parties can benefit from the results and the implemented methodologies such as ISPs and owners of the Services for their future planning or expansion of the current services as well as professional players to increase their success on social media

Security is essential to communication between entities in the internet. Delay tolerant and disconnected Mobile Ad Hoc Networks (MANET) are a class of networks characterized by high end-to-end path latency and frequent end-to-end disconnections and are often termed as challenged networks. In these networks nodes are sparsely populated and without the existence of a central server, acquiring global information is difficult and impractical if not impossible and therefore traditional security schemes proposed for MANETs cannot be applied. This thesis reports trust management schemes for peer to peer (P2P) application in delay tolerant disconnected MANETs. Properties of a profile based file sharing application are analyzed and a framework for structured P2P overlay over delay tolerant disconnected MANETs is proposed. The framework is implemented and tested on J2ME based smart phones using Bluetooth communication protocol. A light weight Content Driven Data Propagation Protocol (CDDPP) for content based data delivery in MANETs is presented. The CDDPP implements a user profile based content driven P2P file sharing application in disconnected MANETs. The CDDPP protocol is further enhanced by proposing an adaptive opportunistic multihop content based routing protocol (ORP). ORP protocol considers the store-carry-forward paradigm for multi-hop packet delivery in delay tolerant MANETs and allows multi-casting to selected number of nodes. Performance of ORP is compared with a similar autonomous gossiping (A/G) protocol using simulations. This work also presents a framework for trust management based on dynamicity aware graph re-labelling system (DA-GRS) for trust management in mobile P2P applications. The DA-GRS uses a distributed algorithm to identify trustworthy nodes and generate trustable groups while isolating misleading or untrustworthy nodes. Several simulations in various environment settings show the effectiveness of the proposed framework in creating trust based communities. This work also extends the FIRE distributed trust model for MANET applications by incorporating witness based interactions for acquiring trust ratings. A witness graph building mechanism in FIRE+ is provided with several trust building policies to identify malicious nodes and detect collusive behaviour in nodes. This technique not only allows trust computation based on witness trust ratings but also provides protection against a collusion attack. Finally, M-trust, a light weight trust management scheme based on FIRE+ trust model is presented.

Sharing economy services have grown significantly in the last decade. Thereby, P2P accommodation sharing services represent one of the largest sectors and play a crucial role in the tourism industry. The purpose of this paper is to explain the relationship between motives as well as concerns and the customer satisfaction of users in accommodation sharing services. Furthermore, it is investigated if the generation is affecting this relationship. In order to answer the research questions a quantitative method was used. A survey was carried out, which delivered 157 valid responses from existing users of accommodation sharing services. The findings of this research support the positive impact of the motives and the negative impact of the concerns on customer satisfaction. However, no effect of generation on the relationship could be found, whereas there is an indication that nationality might be an influencing factor. Since only the motives and concerns of existing users are studied, the sample is limited to a small share of the whole population, which has already used accommodation sharing services. The paper tests empirically the concept of generation and its impact on the relationship between the motives as well as concerns and the customer satisfaction in the setting of accommodation sharing services. This study shows that it is important for P2P accommodation sharing platform providers and hosts of P2P accommodations to deal with the concerns of the customers as well as to address aspects of the motives in their marketing campaigns to increase the customer satisfaction.