Recent research has indicated that although security systems are developing, illegal intrusion to computers is on the rise. The research conducted here illustrates that improving intrusion detection and prevention methods is fundamental for improving the overall security of systems. This research includes the design of a novel Intrusion Detection System (IDS) which identifies four levels of visibility of attacks. Two major areas of security concern were identified: speed and volume of attacks; and complexity of multistage attacks. Hence, the Multistage Intrusion Detection and Prevention System (MIDaPS) that is designed here is made of two fundamental elements: a multistage attack engine that heavily depends on attack trees and a Denial of Service Engine. MIDaPS were tested and found to improve current intrusion detection and processing performances. After an intensive literature review, over 25 GB of data was collected on honeynets. This was then used to analyse the complexity of attacks in a series of experiments. Statistical and analytic methods were used to design the novel MIDaPS. Key findings indicate that an attack needs to be protected at 4 different levels. Hence, MIDaPS is built with 4 levels of protection. As, recent attack vectors use legitimate actions, MIDaPS uses a novel approach of attack trees to trace the attacker's actions. MIDaPS was tested and results suggest an improvement to current system performance by 84% whilst detecting DDOS attacks within 10 minutes.

Récemment, l’informatique distribuée a connu une grande évolution en raison de l’utilisation du paradigme des agents mobiles, doté d’innovantes capacités, au lieu du système client-serveur où les applications sont liées à des nœuds particuliers dans les réseaux. Ayant capturé l’intérêt des chercheurs et de l’industrie, les agents mobiles sont capables de migrer de manière autonome d’un nœud à un autre à travers le réseau, en transférant de leur code et leurs données, ce qui leur permet d’effectuer efficacement des calculs, de recueillir des informations et d’accomplir des tâches. Cependant, en dépit de ses avantages significatifs, ce paradigme souffre encore de certaines limitations qui font obstacle à son expansion, principalement dans le domaine de la sécurité. Selon les efforts actuellement déployés pour évaluer la sécurité des agents mobiles, deux catégories de menaces sont considérées. La première catégorie concerne les attaques menées sur l’agent mobile lors de son voyage à travers des hôtes ou des entités malveillantes, tandis que la seconde catégorie traite les attaques effectuées par un agent mobile illicite afin d’affecter la plate-forme d’hébergement et de consommer ses ressources. Ainsi, il est substantiellement nécessaire de concevoir une infrastructure de sécurité complète pour les systèmes d’agents mobiles, qui comprend la méthodologie, les techniques et la validation. L’objectif de cette thèse est de proposer des approches qui fournissent cette technologie avec des fonctionnalités de sécurité, qui correspondent à sa structure globale sans compromettre ses capacités de mobilité, l’interopérabilité et l’autonomie. Notre première approche est basée sur la sérialisation XML et des primitives cryptographiques, afin d’assurer une mobilité persistante de l’agent ainsi qu’une communication sécurisée avec les plates-formes d’hébergement. Dans la seconde approche, nous avons conçu une alternative à la première approche en utilisant la sérialisation binaire et la cryptographie à base de l’identité. Notre troisième approche introduit l’aspect d’anonymat à l’agent mobile, et lui fournit un mécanisme de traçage pour détecter les intrusions le long de son voyage. La quatrième approche a été développée dans le but de restreindre l’accès aux ressources de la plate-forme de l’agent, en utilisant une politique de contrôle d’accès bien définie à base la cryptographie à seuil. A ce stade, on s’est intéressé à expérimenter l’utilité des agents mobiles avec des fonctionnalités de sécurité, dans la préservation de la sécurité des autres technologies, telles que le Cloud Computing. Ainsi, nous avons proposé une architecture innovante du Cloud, en utilisant des agents mobiles dotés de traces cryptographiques pour la détection d’intrusion et d’un protocole de révocation à base de seuil de confiance pour la prévention. / Recently, the distributed computing has witnessed a great evolution due to the use of mobile agent paradigm, endowed with innovative capabilities, instead of the client-server system where the applications are bound to particular nodes in networks. Having captured the interest of researchers and industry, the mobile agents areable to autonomously migrate from one node to another across the network, transferring their code and data, which allows them to efficiently perform computations, gather information and accomplish tasks. However, despite its significant benefits, this paradigm still suffering from some limitations that obstruct its expansion, primarily in the area of security. According to the current efforts to investigate the security of mobile agents, two categories of threats are considered. The first one concerns the attacks carried out on the mobile agent during its travel or stay by malicious hosts or entities, while the second one deals the attacks performed by a malicious mobile agent in order to affect the hosting platform and consume its resources. Thus, it is substantially needed to conceive a complete security infrastructure for mobile agent systems, which includes methodology, techniques and validation. The aim of this thesis is to propose approaches which provide this technology with security features, that meet with its overall structure without compromising its mobility, interoperbility and autonomy capabilities. Our first approach was based on XML serialization and cryptographic primitives, in order to ensure a persistent mobility of agent as well as a secure communication with hosting platforms. In the second approach, we have conceived an alternative to the first approach using binary serialization and Identity-based cryptography. Our third approach was proposed to introduce anonymity aspect to the mobile agent, and provide him with a tracing mechanism to detect intrusions along its trip. The fourth approach was developed in order to restrict the access to the resources of the agent platform, using a well-defined access control policy based on threshold cryptography. At this stage, we find it interesting to experiment the utility of mobile agents with security features in preserving the security of other technologies such as cloud computing. Thus, we have developed an innovative cloud architecture using mobile agents endowed with cryptographic traces for intrusion detection and a revocation protocol based on trust threshold for prevention.

Recent research has indicated that although security systems are developing, illegal intrusion to computers is on the rise. The research conducted here illustrates that improving intrusion detection and prevention methods is fundamental for improving the overall security of systems. This research includes the design of a novel Intrusion Detection System (IDS) which identifies four levels of visibility of attacks. Two major areas of security concern were identified: speed and volume of attacks; and complexity of multistage attacks. Hence, the Multistage Intrusion Detection and Prevention System (MIDaPS) that is designed here is made of two fundamental elements: a multistage attack engine that heavily depends on attack trees and a Denial of Service Engine. MIDaPS were tested and found to improve current intrusion detection and processing performances. After an intensive literature review, over 25 GB of data was collected on honeynets. This was then used to analyse the complexity of attacks in a series of experiments. Statistical and analytic methods were used to design the novel MIDaPS. Key findings indicate that an attack needs to be protected at 4 different levels. Hence, MIDaPS is built with 4 levels of protection. As, recent attack vectors use legitimate actions, MIDaPS uses a novel approach of attack trees to trace the attacker¿s actions. MIDaPS was tested and results suggest an improvement to current system performance by 84% whilst detecting DDOS attacks within 10 minutes.

As the networks become faster and faster, the emerging requirement is to improve the performance of the Intrusion Detection and Prevention Systems (IDPS) to keep up with the increased network throughput. In high speed networks, it is very difficult for the IDPS to process all the packets. Since the throughput of IDPS is not improved as fast as the throughput of the switches and routers, it is necessary to develop new detection techniques other than traditional techniques. In this thesis we propose a rule-based IDPS technique to detect Layer 2-4 attacks by just examining the flow data without inspecting packet payload. Our approach is designed to work as an additional component to existing IDPS as we acknowledge that the attacks at Layer 5 and above require payload inspection. The rule set is constructed and tested on a real network to evaluate the performance of the system.

It is well-established that machine learning techniques have been used with great success in other domains and has been leveraged to deal with sources of evolving abuse, such as spam. This study aims to determine whether machine learning techniques can be used to create a model that detects vulnerability scanning attacks using proprietary real-world data collected from tCell, a web application firewall. In this context, a vulnerability scanning attack is defined as an automated process that detects and classifies security weaknesses and flaws in the web application. To test the hypothesis that machine learning techniques can be used to create a detection model, twenty four models were trained. The models showed a high level of precision and recall, ranging from 91% to 0.96% and 85% to 0.93%, respectively. Although the classification performance was strong, the models were not calibrated sufficiently which resulted in an underconfidence in the predictions. The results can therefore been viewed as a performance baseline. Nevertheless, the results demonstrate an advancement over the simplistic threshold-based techniques developed in the early days of the internet, but require further research and development to tune and calibrate the models. / Det är väletablerat att tekniker för maskininlärning har använts med stor framgång inom andra domäner och har utnyttjats för att hantera källor till växande missbruk, såsom spam. Denna studie syftar till att avgöra om maskininlärningstekniker kan tillämpas för att skapa en modell som upptäcker sårbarhets-skanningsattacker med hjälp av proprietär data som samlats in från tCell, en webbapplikationsbrandvägg. I detta sammanhang definieras en sårbarhetsskanningsattack som en automatiserad process som upptäcker och klassificerar säkerhetsbrister och brister i webb-applikationen. För att testa hypotesen att maskininlärningstekniker kan användas för att skapa en detektionsmodell, tränades tjugofyra modeller. Modellerna visade en hög nivå av precision och sensitivitet, från 91% till 0,96% och 85% till 0,93%, respektive. Även om klassificeringsprestandan var god, var modellerna inte tillräckligt kalibrerade, vilket resulterade i ett svagt förtoende för förutsägelserna. De presenterade resultaten kan därför ses som en prestationsbaslinje. Resultaten visar ett framsteg över de förenklade tröskelbaserade teknikerna som utvecklades i begynnelsen av internet, men kräver ytterligare forskning och utveckling för att kalibrera modellerna.

No description available.