Arquitectura para el manejo de congestión en una red de datos corporativa con participación del usuario, basado en inteligencia computacional
Cargando...
Archivos
Autores
Ardila Triana, Erick
Director
Tipo de contenido
Document language:
Español
Fecha
Título de la revista
ISSN de la revista
Título del volumen
Documentos PDF
Resumen
En este trabajo se realizó el diseño e implementación de un modelo para la detección de congestión para una red de datos de área local con participación del usuario y usando herramientas en inteligencia computacional. Para adelantar esta investigación, primero fue necesario identificar las variables a utilizar como base para la medición de congestión. Para esto se hizo una recolección de datos en la red LAN de la sede Bogotá de la Universidad Nacional de Colombia y se realizó un análisis estadístico de dichos datos. Luego se planteó un modelo para la detección de congestión, basado en la interacción del usuario final de la red. Particularmente se desarrolló un sistema multi-agente distribuido para adelantar el proceso de detección. Para comprobar el funcionamiento del modelo se desarrollaron pruebas experimentales sobre un ambiente controlado en la red de estudio. Los resultados obtenidos corroboraron el adecuado funcionamiento del modelo propuesto.
Abstract. This work presents the design and implementation of a model for detecting congestion in local area networks. The proposed approach is based on the user’s participation and used computational intelligence methods. First, it was necessary to identify the variables used to measure congestion. Some data were collected at the LAN from the Bogota campus of the National University of Colombia and some statistical analysis on such data was performed. Then a congestion detection model based on end-user/network interaction was developed. Specifically, a distributed multi-agent system was proposed. In order to validate this model, some experimental tests on a controlled environment were performed on the network under study. The obtained results showed the proper operation of the proposed model.
Abstract. This work presents the design and implementation of a model for detecting congestion in local area networks. The proposed approach is based on the user’s participation and used computational intelligence methods. First, it was necessary to identify the variables used to measure congestion. Some data were collected at the LAN from the Bogota campus of the National University of Colombia and some statistical analysis on such data was performed. Then a congestion detection model based on end-user/network interaction was developed. Specifically, a distributed multi-agent system was proposed. In order to validate this model, some experimental tests on a controlled environment were performed on the network under study. The obtained results showed the proper operation of the proposed model.