<p>Network and system management has always been of concern for telecommunication and computer system operators. The need for standardization was recognised already 20 years ago, hence several standards for network management exist today. However, the ever-increasing number of units connected to networks and the ever-increasing number of services being provided results in significant increased complexity of average network environments. This challenges current management systems. In addition to the general increase in complexity the trend among network owners and operators of merging several single service networks into larger, heterogeneous and complex full service networks challenges current management systems even further. The full service networks will require management systems more powerful than what is possible to realize basing systems purely on todays management standards. This thesis presents a distributed stochastic optimization algorithm which enables implementations of highly robust and efficient management tools. These tools may be integrated into management systems and potentially make the systems more powerful and better prepared for management of full service networks.</p><p>Emergent behavior is common in nature and easily observable in colonies of social insects and animals. Even an old oak tree can be viewed as an emergent system with its collection of interacting cells. Characteristic for any emergent system is how the overall behavior of the system emerge from many relatively simple, restricted behaviors interacting, e.g. a thousand ants building a trail, a flock of birds flying south or millions of cells making a tree grow. No centralized control exist, i.e. no single unit is in charge making global decisions. Despite distributed control, high work redundancy and stochastic behavior components, emergent systems tend to be very efficient problem solvers. In fact emergent systems tend to be both efficient, adaptive and robust which are three properties indeed desirable for a network management system. The algorithm presented in this thesis relates to a class of emergent behavior based systems known as swarm intelligence systems, i.e. the algorithm is potentially efficient, adaptive and robust.</p><p>On the contrary to other related swarm intelligence algorithms, the algorithm presented has a thorough formal foundation. This enables a better understanding of the algorithm’s potentials and limitations, and hence enables better adaptation of the algorithm to new problem areas without loss of efficiency, adaptability or robustness. The formal foundations are based on work by Reuven Rubinstein on cross entropy driven optimization. The transition from Ruinstein’s centralized and synchronous algorithm to a distributed and asynchronous algorithm is described, and the distributed algorithm’s ability to solve complex problems (NP-complete) efficiently is demonstrated.</p><p>Four examples of how the distributed algorithm may be applied in a network management context are presented. A system for finding near optimal patterns of primary/backup paths together with a system for finding cyclic protection paths in mesh networks demonstrate the algorithm’s ability to act as a tool helping management system to ensure quality of service. The algorithm’s potential as a management policy implementation mechanism is also demonstrated. The algorithm’s adaptability is shown to enable resolution of policy conflicts in a soft manner causing as little loss as possible. Finally, the algorithm’s ability to find near optimal paths (i.e. sequences) of resources in networks of large scale is demonstrated.</p>

Network and system management has always been of concern for telecommunication and computer system operators. The need for standardization was recognised already 20 years ago, hence several standards for network management exist today. However, the ever-increasing number of units connected to networks and the ever-increasing number of services being provided results in significant increased complexity of average network environments. This challenges current management systems. In addition to the general increase in complexity the trend among network owners and operators of merging several single service networks into larger, heterogeneous and complex full service networks challenges current management systems even further. The full service networks will require management systems more powerful than what is possible to realize basing systems purely on todays management standards. This thesis presents a distributed stochastic optimization algorithm which enables implementations of highly robust and efficient management tools. These tools may be integrated into management systems and potentially make the systems more powerful and better prepared for management of full service networks. Emergent behavior is common in nature and easily observable in colonies of social insects and animals. Even an old oak tree can be viewed as an emergent system with its collection of interacting cells. Characteristic for any emergent system is how the overall behavior of the system emerge from many relatively simple, restricted behaviors interacting, e.g. a thousand ants building a trail, a flock of birds flying south or millions of cells making a tree grow. No centralized control exist, i.e. no single unit is in charge making global decisions. Despite distributed control, high work redundancy and stochastic behavior components, emergent systems tend to be very efficient problem solvers. In fact emergent systems tend to be both efficient, adaptive and robust which are three properties indeed desirable for a network management system. The algorithm presented in this thesis relates to a class of emergent behavior based systems known as swarm intelligence systems, i.e. the algorithm is potentially efficient, adaptive and robust. On the contrary to other related swarm intelligence algorithms, the algorithm presented has a thorough formal foundation. This enables a better understanding of the algorithm’s potentials and limitations, and hence enables better adaptation of the algorithm to new problem areas without loss of efficiency, adaptability or robustness. The formal foundations are based on work by Reuven Rubinstein on cross entropy driven optimization. The transition from Ruinstein’s centralized and synchronous algorithm to a distributed and asynchronous algorithm is described, and the distributed algorithm’s ability to solve complex problems (NP-complete) efficiently is demonstrated. Four examples of how the distributed algorithm may be applied in a network management context are presented. A system for finding near optimal patterns of primary/backup paths together with a system for finding cyclic protection paths in mesh networks demonstrate the algorithm’s ability to act as a tool helping management system to ensure quality of service. The algorithm’s potential as a management policy implementation mechanism is also demonstrated. The algorithm’s adaptability is shown to enable resolution of policy conflicts in a soft manner causing as little loss as possible. Finally, the algorithm’s ability to find near optimal paths (i.e. sequences) of resources in networks of large scale is demonstrated.

Made available in DSpace on 2016-06-02T19:46:14Z (GMT). No. of bitstreams: 1 DissCKBO.pdf: 2526592 bytes, checksum: 4e8f2461b84b8e2b8986ef676f28a287 (MD5) Previous issue date: 2004-08-30 / Universidade Federal de Minas Gerais / This study investigated responding by exclusion , which has been demonstrated as a robust behavioral process, consisting in the selection of an undefined comparison stimulus, when the sample is also an undefined stimulus (that is, both stimuli are new in the experimental setting), without any previous history that could establish the comparison as a discriminative stimulus for the selection response. Exclusion has been extensively replicated with an experimental preparation of conditional auditory-visual discriminations in which the participant learns to relate a single sample to a single comparison. Recent research has also demonstrated the occurrence of exclusion when participants learn to relate a single name to different pictures and many names to single picture. The purpose of this study was to investigate whether exclusion would occur with visual stimuli used both as sample and comparison stimuli, without a spoken word as a nodal stimulus. Three experiments were conducted using the same experimental arrangements used in previous studies, to establish the visualvisual conditional discriminations baseline for the exclusion probes: each sample related to many comparisons (Experiment I), many samples related to each comparison (Experiment II) and a combination of both arrangements (Experiment III). In each experiment, children aged 4 to 5 years were exposed to the following experimental sequence: 1) establishment of a baseline of conditional discriminations with visual stimuli; 2) exclusion probes; 3) teaching new baseline relations; 4) additional exclusion probes; 5) probes of equivalence class formation. All children learned the conditional discriminations and showed positive results on exclusion probes and equivalence probes. The regularity observed in the data supports the conclusion that responding by exclusion occurs under a baseline of conditional discriminations including only visual stimuli, both with one-to-one pairings and many-to-one (or one-to-many) pairings. These data confirm and extend previous findings and have important implications for the comprehension of relational learning, the emergence of new behavior and the symbolic function and its role in language acquisition. / Este trabalho investigou o responder por exclusão , um fenômeno comportamental robusto, demonstrado experimentalmente pela seleção imediata de um estímulo de comparação indefinido, diante de um estímulo modelo também indefinido experimentalmente (isto é, estímulos novos na situação experimental), sem uma história prévia que torne o estímulo de comparação em discriminativo para a resposta de seleção. A exclusão tem sido extensamente replicada com uma preparação experimental de linha de base de discriminações condicionais auditivo-visuais em que o participante aprende que a cada estímulo modelo corresponde um, e apenas um, estímulo de comparação, mas pesquisa recente também demonstrou a ocorrência de exclusão quando os participantes aprendem a relacionar um mesmo nome a figuras diferentes ou nomes diferentes a uma mesma figura. O presente estudo teve como objetivo principal investigar se a exclusão ocorreria se todos os estímulos fossem visuais, sem a palavra falada como estímulo nodal. Foram realizados três experimentos que buscaram replicar alguns dos estudos prévios com estímulos auditivos, quanto à linha de base para os testes de exclusão: um mesmo modelo relacionado a vários estímulos de comparação (Experimento I), vários modelos relacionados a um mesmo estímulo de comparação (Experimento II) e combinações de relações múltiplas e singulares (Experimento III). Em cada experimento, crianças de 4 a 5 anos foram expostas à seguinte seqüência experimental: 1- estabelecimento de uma linha de base de discriminações condicionais visuais; 2- sondas de exclusão; 3- ampliação da linha de base pelo ensino de novas relações; 4- novas sondas de exclusão; 5- testes de formação de classes de estímulos. Todos os participantes adquiriram as discriminações condicionais e apresentaram resultados positivos nas sondas de exclusão e nos testes de formação de classes de equivalência. A regularidade dos dados permite a conclusão de que a seleção por exclusão ocorre com linha de base constituída somente por estímulos visuais, tanto com relações de emparelhamento um para um, como muitos para um, confirmando e estendendo descobertas prévias. Os dados têm importantes implicações para a compreensão da aprendizagem relacional, da emergência de comportamentos novos e da função simbólica e seu papel na aquisição de linguagem.

This thesis proposes an Ecosystem perspective for the engineering of SoS and CPS and illustrates the impact of this perspective in three areas of contribution category First, from a conceptual and Systems Engineering perspective, a conceptual framework including the Ecosystems of System Unified Language Modeling (EoS-UML) profile, a set of Ecosystem Ensemble Diagrams, the Arms :Length Trust Model and the Cyber Physical Threat Model are provided. Second, having established this conceptual view of the ecosystem, we recognize unique role of the cryptographic credentials within it, towards enabling the ecosystem long-term value proposition and acting as a value transfer agent, implementing careful balance of properties meet stakeholder needs. Third, we propose that the ecosystem computers can be used as a distributed compute engine to run Collaborative Algorithms. To demonstrate, we define access control scheme, risk-balanced Cellular Access Control (rbCAC). The rbCAC algorithm defines access control within a cyber-physical environment in a manner which balances cost, risk, and net utility in a multi-authority setting. rbCAC is demonstrated it in an Air Travel and Border Services scenario. Other domains are also discussed included air traffic control threat prevention from drone identity attacks in protected airspaces. These contributions offer significant material for future development, ongoing credential and ecosystem design, including dynamic perimeters and continuous-time sampling, intelligent and self optimizing ecosystems, runtime collaborative platform design contracts and constraints, and analysis of APT attacks to SCADA systems using ecosystem approaches.