There are many variants of leader election algorithm in distributed networks. In this research, an agent based approach to leader election in distributed networks is investigated. Agents have shown to be useful in several ways. In the theoretical perspective, agents sometime help in reducing the message complexity of the system and sometimes help in lowering time complexity. In a more practical sense, agents perform operations independent of the processors, thereby lending a more flexible algorithm supporting different types of networks.

Network decontamination is a well known mobile agent problem with many applications. We assume that all nodes of a network are contaminated (e.g., by a virus) and a set of agents is deployed to decontaminate them. An agent passing by a node decontaminates it, however a decontaminated node can be recontaminated if any of its neighbours is contaminated. In the vast literature a variety of models are considered and different assumptions are made on the power of the agents. In this thesis we study variation of the decontamination problem in mesh and tori topologies, under the assumption that when a node is decontaminated, it is immune to recontamination for a predefined amount of time t (called immunity time). After the immunity time is elapsed, recontamination can occur. We focus on three different models: mobile agents (MA), cellular automata (CA), and mobile cellular automata (MCA). The first two models are commonly studied and employed in several other contexts, the third model is introduced in this thesis for the first time. In each model we study the temporal decontamination problem (adapted to the particular setting) under a variety of assumptions on the capabilities of the decontaminating elements (agents for MA and MCA, decontaminating cells for CA). Some of the parameters we consider in this study are: visibility of the active elements, their ability to make copies of themselves, their ability to communicate, and the possibility to remember their past actions (memory). We describe several solutions in the various scenarios and we analyze their complexity. Efficiency is evaluated slightly differently in each model, but essentially the effort is in the minimization of the number of simultaneous decontaminating elements active in the system while performing the decontamination with a given immunity time.

We address the security issue of the data which mobile agents gather as they are traversing the Internet. Our goal is to devise a security protocol that truly secures the data which mobile agents gather. Several cryptographic protocols were presented in the literature asserting the security of gathered data. Formal verification of the protocols reveals unforeseen security flaws, such as truncation or alteration of the collected data, breaching the privacy of the gathered data, sending others data under the private key of a malicious host, and replacing the collected data with data of similar agents. So the existing protocols are not truly secure. We present an accurate security protocol which aims to assert strong integrity, authenticity, and confidentiality of the gathered data. The proposed protocol is derived from the Multi-hops protocol. The protocol suffers from security flaws, e.g. an adversary might truncate/ replace collected data, or sign others data with its own private key without being detected. The proposed protocol refines the Multi-hops protocol by implementing the following security techniques: utilization of co-operating agents, scrambling the gathered offers, requesting a visited host to clear its memory from any data acquired as a result of executing the agent before the host dispatches the agent to the succeeding host in the agent???s itinerary, and carrying out verifications on the identity of the genuine initiator at the early execution of the agent at visited hosts, in addition to the verifications upon the agent???s return to the initiator. The proposed protocol also implements the common security techniques such as public key encryption, digital signature, etc. The implemented security techniques would rectify the security flaws revealed in the existing protocols. We use STA, an infinite-state exploration tool, to verify the security properties of a reasonably small instance of the proposed protocol in key configurations. The analysis using STA reports no attack. Moreover, we carefully reason the correctness of the security protocol for a general model and show that the protocol would be capable of preventing or at least detecting the attacks revealed in the existing protocols.

Thesis (M.C.S.) - Carleton University, 2006. / Includes bibliographical references (p. 115-122). Also available in electronic format on the Internet.

Thesis (M. Sc.)--University of Canterbury, 2008. / Typescript (photocopy). Includes bibliographical references (p. 104-111). Also available via the World Wide Web.

Thesis (Ph. D.)--University of Cambridge, 2000. / Cover title. "June 2000." Includes bibliographical references.

Network decontamination is a well known mobile agent problem with many applications. We assume that all nodes of a network are contaminated (e.g., by a virus) and a set of agents is deployed to decontaminate them. An agent passing by a node decontaminates it, however a decontaminated node can be recontaminated if any of its neighbours is contaminated. In the vast literature a variety of models are considered and different assumptions are made on the power of the agents. In this thesis we study variation of the decontamination problem in mesh and tori topologies, under the assumption that when a node is decontaminated, it is immune to recontamination for a predefined amount of time t (called immunity time). After the immunity time is elapsed, recontamination can occur. We focus on three different models: mobile agents (MA), cellular automata (CA), and mobile cellular automata (MCA). The first two models are commonly studied and employed in several other contexts, the third model is introduced in this thesis for the first time. In each model we study the temporal decontamination problem (adapted to the particular setting) under a variety of assumptions on the capabilities of the decontaminating elements (agents for MA and MCA, decontaminating cells for CA). Some of the parameters we consider in this study are: visibility of the active elements, their ability to make copies of themselves, their ability to communicate, and the possibility to remember their past actions (memory). We describe several solutions in the various scenarios and we analyze their complexity. Efficiency is evaluated slightly differently in each model, but essentially the effort is in the minimization of the number of simultaneous decontaminating elements active in the system while performing the decontamination with a given immunity time.

Este texto apresenta nosso projeto de implementação de um arcabouço de suporte a agentes móveis dentro de um ambiente de grade denominado InteGrade. Nosso arcabouço - MobiGrid - foi criado de forma a permitir que aplicações seqüenciais longas possam ser executadas em uma rede de estações de trabalho pessoais. Os agentes móveis são utilizados para encapsular essas aplicações com longo tempo de processamento. O encapsulamento de uma aplicação com longo tempo de processamento dentro de um agente móvel é o que denominamos como tarefa. Sendo assim, as tarefas podem migrar sempre que a máquina é requisitada por seu usuário local, já que são providas com capacidade de migração automática. Nosso arcabouço também fornece ao usuário um gerente que rastreia as tarefas por ele submetidas. Baseados no ambiente de execução de tarefas descrito, criamos um modelo matemático para efetuarmos simulações de como se comportariam muitas tarefas submetidas a uma grade com grande quantidade de estações de trabalho. Neste trabalho apresentamos também esse modelo, bem como os resultados das simulações nele baseadas. / This text presents a project which focuses on the implementation of a framework for mobile agents support within a grid environment project, namely InteGrade. Our framework - MobiGrid - was created in such a way that time consuming sequential applications can be executed on a network of personal workstations. The encapsulation of a long processing application by a mobile agent is what we call task. Hence, the tasks can migrate whenever the local machine is requested by its local user, since they are provided with automatic migration capabilities. Our framework also provides the user with a manager that keeps track of the submitted agents. Based on the execution environment described above, we have created a mathematical model which allows us to simulate how a great quantity of tasks submitted to a grid with many workstations would behave. In this text, we also present our model, as well as the results of our simulations.

No description available.

No description available.