| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 4.61 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
No decorrer dos últimos anos tem-se verificado um crescimento nas dimensões e lógica das redes móveis, o que leva também ao aumento da complexidade e sofisticação dos equipamentos da mesma. Com este crescimento, torna-se cada vez mais difícil realizar uma análise aos dados de configuração e performance de uma rede, de forma correta, centralizada e em tempo útil. A presente dissertação tem como objetivo desenvolver uma solução de correlação, análise e visualização centralizada dos dados da rede, atribuindo mais valor à informação disponível sobre a mesma e resolvendo o problema de extração do conhecimento útil a partir da informação disponível. No decorrer deste trabalho são estudadas diferentes técnicas e formas de representação para grandes volumes de dados e avaliado o seu impacto na agilização das tarefas de análise de uma rede. Foram desenvolvidos três tipos de gráficos: histograma, gráfico de barras e scatterplot. Os gráficos desenvolvidos foram avaliados quanto à sua performance com conjunto de dados de diferentes dimensões. Com o propósito de avaliar a performance da solução, bem como o desempenho dos diferentes tipos de visualização e a sua utilidade, foram efetuados inquéritos junto de potenciais utilizadores, tendo se concluído que estes melhoram e facilitam o processo de análise.
Over the last few years, there has been a growth in the dimensions and logic of mobile networks, which leads to an increase in the complexity and sophistication of the network equipment. With this growth, it becomes increasingly difficult to properly analyse the configuration and performance data of a network correctly and in due time. The purpose of this dissertation is to propose a solution that allows centralized correlation and visualization of the network data, by assigning more value to the information available on it and solving the problems of extracting useful knowledge from the available information. In the course of this work, different techniques and representation methods for large volumes of data and their impact on the performance of network analysis tasks were studied. Three chart types were developed: histogram, bar chart and scatterplot. The developed charts were evaluated for their performance with data sets of different dimensions. In order to evaluate both the performance of the solution and the different types of visualization, plusits usefulness to the user, inquiries were made to potential users and it has been concluded that the tool improves and simplifies the analysis process.
Over the last few years, there has been a growth in the dimensions and logic of mobile networks, which leads to an increase in the complexity and sophistication of the network equipment. With this growth, it becomes increasingly difficult to properly analyse the configuration and performance data of a network correctly and in due time. The purpose of this dissertation is to propose a solution that allows centralized correlation and visualization of the network data, by assigning more value to the information available on it and solving the problems of extracting useful knowledge from the available information. In the course of this work, different techniques and representation methods for large volumes of data and their impact on the performance of network analysis tasks were studied. Three chart types were developed: histogram, bar chart and scatterplot. The developed charts were evaluated for their performance with data sets of different dimensions. In order to evaluate both the performance of the solution and the different types of visualization, plusits usefulness to the user, inquiries were made to potential users and it has been concluded that the tool improves and simplifies the analysis process.
Description
Keywords
Rede Móvel Performance Visualização Telecomunicações Análise Mobile Network Performance Visualization Telecommunications Analysis