Bibliographic citations
Gonzales, R., (2021). Diseño de una red de datos flexible utilizando SDN para reducir el consumo de energía en un centro de datos [Tesis, Universidad Nacional de Ingeniería]. http://hdl.handle.net/20.500.14076/22359
Gonzales, R., Diseño de una red de datos flexible utilizando SDN para reducir el consumo de energía en un centro de datos [Tesis]. PE: Universidad Nacional de Ingeniería; 2021. http://hdl.handle.net/20.500.14076/22359
@mastersthesis{renati/711892,
title = "Diseño de una red de datos flexible utilizando SDN para reducir el consumo de energía en un centro de datos",
author = "Gonzales Figueroa, Renatto Gustavo",
publisher = "Universidad Nacional de Ingeniería",
year = "2021"
}
The energy efficiency of a data center is a ratio of the total energy consumed by the data center and the energy consumed by the computer systems, known as PUE. In this project, an alternative is developed to reduce energy consumption in a data center, keeping on only the network resources that are used to carry the traffic that crosses the network at any given time. To accomplish this task, we focus specifically on connectivity devices: like switches, their ports and the links that connect them; and not the rest of the infrastructure. Today network technologies are embracing a new paradigm: software-defined networks. With them, the possibility of centrally controlling network devices is opened through applications that can be developed by network administrators, automate tasks, and customize network behavior. In this project, we will take advantage of the flexibility and programmability that SDNs offer us to manage and keep on or off network resources of the data center, in order to reduce its energy consumption. The use of a route search algorithm is necessary to identify the possible route that can be assigned to a traffic flow and to ensure that the network resources on that route are available. A * is the algorithm selected for the project, it is an informed heuristic search algorithm that not only uses the limit as a heuristic function of the way to go but also considers the cost of the way already traveled. Therefore, an ElasticTree solution based on SDN has been developed, programming a Python application that uses the A * algorithm to select a route for new incoming traffic. Once the route is selected, the controller updates, through the OpenFlow protocol, the flow tables of the switches that will be part of the selected route. Using a calculation tool such as OCTAVE, we apply an energy model to calculate the variation in energy consumption, keeping only the necessary equipment on, achieving savings of up to 30% compared to a traditional rigid infrastructure.
This item is licensed under a Creative Commons License