Uncovering the metrics and procedures employed by an autonomous networking system is an important problem with applications in instrumentation, traffic engineering, and game-theoretic studies of multi-agent environments. This thesis presents a method for utilizing inverse reinforcement learning (IRL)techniques for the purpose of discovering a composite metric used by a dynamic routing algorithm on an Internet Protocol (IP) network. The network and routing algorithm are modeled as a reinforcement learning (RL) agent and a Markov decision process (MDP). The problem of routing metric discovery is then posed as a problem of recovering the reward function, given observed optimal behavior. We show that this approach is empirically suited for determining the relative contributions of factors that constitute a composite metric. Experimental results for many classes of randomly generated networks are presented. / Master of Science

Today`s continuously growing Internet requires users and network applications to have knowledge of network metrics. This knowledge is critical for decision making during the usage of network applications. This thesis studies application related network metrics. The major approach in this work is to examine the traffic between a simulated user and network applications. We use the historical data collected from previous usage of network applications to make predictions for future usage of those applications. We also use the historical data obtained from a given application to make predictions about another application. Prediction mechanisms require us to make parameter choices so that certain weights can be placed on historical data versus current data. We study these different choices and use the values from our best experimental results. From these studies we conclude that our data prediction is quite accurate and remains stable over a range of parameter choices. The use of shared routing paths between users and network applications are explored in the performance prediction of applications. Only some servers at the same locations show similar prediction results. The network applications studied are also varied, including web, streaming, DNS, etc. We see whether sharing information obtained from different applications can be used to make predictions of application performance. However, we observe limited success in predictions across applications.

Submitted by Franciele Moreira (francielemoreyra@gmail.com) on 2018-04-11T11:57:10Z No. of bitstreams: 2 Dissertação - Lucas Pereira Martins - 2018.pdf: 2113878 bytes, checksum: 4c38a4d425f42506d50aa62bf6661c3f (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2018-04-12T11:22:54Z (GMT) No. of bitstreams: 2 Dissertação - Lucas Pereira Martins - 2018.pdf: 2113878 bytes, checksum: 4c38a4d425f42506d50aa62bf6661c3f (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2018-04-12T11:22:54Z (GMT). No. of bitstreams: 2 Dissertação - Lucas Pereira Martins - 2018.pdf: 2113878 bytes, checksum: 4c38a4d425f42506d50aa62bf6661c3f (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2018-03-15 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / Understanding how species interact with each other is essential to advance our knowledge on community ecology. However, there are still gaps regarding how interspecific interactions affect and are affected by the variation in the composition of species that constitute the ecological networks. In particular, studies of beta diversity are interesting to understand how sets of biotic filters may affect species’ geographic distribution and persistence on communities. Another interesting question is if variations in species composition may cause changes on how ecological networks are structured. In this dissertation, we propose to evaluate patterns of beta diversity and structure on insect-plant interaction networks. Specifically, in the first chapter we aimed to test if beta diversity is affected by the trophic level to which the assemblage belongs, and by the degree of specialization of species at higher trophic levels to their host species. For this, we used a tritrophic system comprising plants of the family Asteraceae, endophagous herbivores and parasitoids sampled in remnants of Brazilian Cerrado. Our main results show that total beta diversity of parasitoids (higher trophic level) was lower than those of the other trophic levels, while there was no difference between total beta diversity of plants and herbivores. Furthermore, the degree of specialization of the assemblages of herbivores and parasitoids was positively associated to the beta diversity of these groups. In the second chapter, we evaluated if species beta diversity among sites affects dissimilarity in network structure, and if this relationship changes across time. Overall, the relationship between spatial beta diversity and dissimilarity in plant-herbivore network structure was context-dependent, thus indicating that different ecological processes (i.e., niche-based and neutral) may drive the organization of antagonistic networks across time. / Compreender como as espécies interagem entre si é essencial para avançar o nosso conhecimento em ecologia de comunidades. No entanto, ainda existem lacunas no que se refere a como interações interespecíficas influenciam e são influenciadas pela variação na composição de espécies que constituem as redes ecológicas. Em particular, estudos de diversidade beta são interessantes para entender como conjuntos de filtros bióticos podem influenciar a distribuição geográfica e persistência de espécies nas comunidades. Outra questão interessante é se variações na composição de espécies podem causar mudanças no modo como redes ecológicas são estruturadas. Nesta dissertação, propomos avaliar padrões de diversidade beta e estrutura de redes de interações inseto-planta. Especificamente, no primeiro capítulo visamos testar se a diversidade beta é influenciada pelo nível trófico ao qual a assembleia pertence, e pelo grau de especialização de espécies de níveis tróficos superiores às suas espécies hospedeiras. Para isso, utilizamos um sistema tritrófico composto por plantas da família Asteraceae, herbívoros endófagos e parasitoides amostrado em remanescentes de Cerrado brasileiro. Nossos principais resultados demonstram que a diversidade beta total de parasitoides (nível trófico superior) foi menor do que a dos demais níveis tróficos, enquanto que não houve diferença entre a diversidade beta total de plantas e herbívoros. Além disso, o grau de especialização das assembleias de herbívoros e parasitoides foi positivamente relacionado à diversidade beta destes grupos. No segundo capítulo, avaliamos se a diversidade beta de espécies entre locais influencia a dissimilaridade na estrutura de redes ecológicas, e se esta relação muda através do tempo. De modo geral, a relação entre diversidade beta espacial e dissimilaridade na estrutura de redes planta-herbívoro foi contexto-dependente, indicando assim que diferentes processos ecológicos (i.e., baseados em nicho e neutros) podem determinar a organização de redes antagonistas através do tempo.

With the increasingly high speed of the Internet and its wide spread usage, the current Internet architecture exhibits some problems. The publish/subscribe paradigm has been developed to support one of the most common patterns of communication. It makes “information” the center of communication and removes the “location-identity split” (i.e., that objects are at specific locations to which you must communicate with to access the object). In this thesis project a publish/subscribe network is built and then used in the design, implementation, and evaluation of a publish/subscribe proxy for today’s HTTP based communication. By using this proxy users are able to use their existing web browser to send both HTTP requests and Publish/Subscribe Internet Routing Paradigm (PSIRP) requests. A publish/subscribe overlay is responsible for maintaining PSIRP contents. The proxy enables web browser clients to benefit from the publish/subscribe network, without requiring them to change their behavior or even be aware of the fact that the content that they want to access is being provided via the publish/subscribe overlay. The use of the overlay enables a user’s request to be satisfied by any copy of the content, potentially decreasing latency, reducing backbone network traffic, and reducing the load on the original content server. One of the aims of this thesis is to make more PSIRP content available, this is done by introducing a proxy who handles both HTTP and PSIRP requests, but having received content as a result of an HTTP response it publishes this data as PSIRP accessible content. The purpose is to foster the introduction and spread of content based access. / Med allt högre Internetshastighet och dess utbredda användning, uppvisar den aktuella Internet-arkitekturen vissa problem. Publicera / prenumerera paradigm har utvecklats för att stödja en av de vanligaste mönstren för kommunikation. Det gör att "information" blir centrum av kommunikation och tar bort "plats-identitet split" (dvs att objekten är på specifika platser som du måste kommunicera med för att komma åt objektet). i detta examensarbete byggs ett publicera / prenumerera nätverk och sedan används i utformningen, genomförandet, och utvärdering av en publicera / prenumerera proxy för dagens HTTP-baserad kommunikation. Genom att använda denna proxy kan användare kan använda sin befintliga webbläsare för att skicka både HTTP-förfrågningar och publicera / Prenumerera Internet Routing Paradigm (PSIRP) begäran. En publicera / prenumerera överlagring är ansvarig för att upprätthålla innehåll av PSIRP. Fullmakten gör det möjligt för klienter av webbläsare att dra nytta av publicera / prenumerera nätverket, utan att kräva dem att ändra sitt beteende eller ens vara medvetna om det faktum att det innehållet som de vill komma åt tillhandahålls via publicera / prenumerera överlägg. Användningen av överlägget kan en användare begäran som skall uppfyllas av en kopia av innehållet, eventuellt minskande latens, vilket minskar trafiken stamnät, och minska belastningen på det ursprungliga innehållet servern. Ett av syftena med denna uppsats är att göra mer PSIRP innehåll tillgängligt och detta görs genom att införa en proxy som hanterar både HTTP och PSIRP förfrågningar, men har fått innehåll som en följd av en HTTP-svar offentliggörs denna data som PSIRP tillgängligt innehåll. Syftet är att främja införandet och innehållbaserade åtkomsten.

Today the use of the social networking technologies is becoming more common in users daily lives, both their professional and personal lives. Recently the application of social networking technologies’ communication power has begun to be leveraged by enterprises in order to gain competitive advantage in terms of productivity and employees’ efficiency & job performance. However, there exist some gaps in the communication and coordination processes within a company between the different employees and groups involved in their professional activities. Overall, we can see that large enterprise employees lack identity with their companies due to their lack of voice and the small impact that they have within the large number of individuals that compose a large organization. In this thesis we deeply analyze the factors that directly affect communication and collaboration between employees within a company and how we can improve this communication and collaboration using social network technologies. In this work we focus on the state-of- the-art and compare one of the most used social networks (LinkedIn) with the most used corporate network (Yammer) providing a tangible measure with the key performance indicators. Afterwards we measure with Node XL software and its metrics samples of monitored corporate networks, introduce an innovation within the networks and show the improvement of the key performance indicators related to the software  metrics that have changed in order to enhance some social network parameters that are the tangible measure of plenty of social corporate network issues. / Idag blir användningen av sociala nätverk teknik vanligare i användarnas vardag, både deras yrkesliv-och privatliv. Nyligen har börjat tillämpningen av sociala nätverk teknik kommunikation makt utnyttjas av företag för att skaffa sig konkurrensfördelar i form av produktivitet och anställdas effektivitet och arbetsprestation. Men det finns vissa brister i kommunikation och processer samordning inom ett företag mellan olika medarbetare och grupper som deltar i sina yrkesverksamhet. Sammantaget kan vi se att stora företag anställda saknar identitet med sitt företag på grund av deras brist på röst och liten inverkan som de har inom det stora antalet individer som utgör en stor organisation. I denna avhandling analyserar vi djupt vilka faktorer som direkt påverkar kommunikationen och samarbetet mellan de anställda inom ett företag och hur vi kan förbättra den här kommunikationen och samarbetet med sociala nätverk teknik. I detta arbete fokuserar vi på state-of-the-art och jämföra en av de mest använda sociala nätverk (LinkedIn) med det mest använda företagets nätverk (Yammer) ger en konkret med centrala resultatindikatorer. Efteråt mäter vi med nod XL programvara och dess statistik exempel på övervakade företagsnätverk, införa en innovation inom nätverk och visar en förbättring av de centrala resultatindikatorer relaterat till programvaran statistik som har ändrats för att stärka vissa sociala nätverk parametrar som är konkret mått på många sociala företag nätfrågor.

La Science des réseaux est apparue comme un domaine d'étude fondamental pour modéliser un grand nombre de systèmes synthétiques ou du monde réel.La découverte du graphe petit monde et du graphe sans échelle dans ces réseaux a révolutionné la façon d'étudier, d'analyser, de modéliser et de traiter ces réseaux. Dans cette thèse, nous nous intéressons à l'étude des réseaux ayant ces propriétés et souvent qualifiés de réseaux complexes.A notre avis, les recherches menées dans ce domaine peuvent être regroupées en quatre catégories: l'analyse, la structure, le processus/organisation et la visualisation.Nous abordons des problèmes relatifs à chacune de ces catégories tout au long de cette thèse. (...) / Network science has emerged as a fundamental field of study to model many physicaland real world systems around us. The discovery of small world and scale free propertiesof these real world networks has revolutionized the way we study, analyze, model andprocess these networks. In this thesis, we are interested in the study of networks havingthese properties often termed as complex networks. In our opinion, research conducted inthis field can be grouped into four categories, Analysis, Structure, Processes-Organizationand Visualization. We address problems pertaining to each of these categories throughoutthis thesis. (...)

This work reviews the current literature and application methods of strategic KPI frameworks, KPI definition and measurement, target setting and benchmarking for social networks based on journals, books and whitepapers while including an interview with a social media agency. In a second step, the acquired knowledge is applied on a case study for a salon haircare company. The conclusions are based on in-company experts in the field, external social media agencies and an internal survey. Based on a behavioral survey, the requirements and gaps in the organization are identified. This leads to the application of a theoretical framework on the case company and a suitable definition of KPIs. The final delivery includes an analytical tool measuring all KPIs. Finally, all KPIs are related to targets and benchmarks from industry, social network sites, competitors and historical data. The final results are easily transferrable to other players in the industry.

The purpose of this diploma thesis is to provide a complex overview for the network status visualization topic. The diploma thesis is structured into four parts. The first one is dedicated to the definition of the most important terms, together with the description of several visualization methods, network metrics and last but not least to the discussion of visualization risks and benefits. In the second part, an own visualization process is at first proposed and then the particular steps of this process are discussed. Easy to understand examples are applied through this part. In the third part, several method of data collection protocols and visualization tools are briefly described. The analyzed tools are graphic libraries, as well as standalone applications and Network Management Systems. The last part is dedicated to the analysis of the current visualization tools used on the VŠE network. A short analysis of the university's network is also made. According to this analysis, an own solution is proposed for a particular part of network.