Uma nova política para distribuição de conteúdo de vídeo em redes CDN

Abstract

Among the most varied content currently available for access on the Internet, video is undoubtedly the most consumed media. For this to be possible, there are large CDN (Content Delivery Network) providers with a robust infrastructure capable of delivering video content with low latency and high availability. However, in regions where there are infrastructure limitations, large CDN providers are not present and there are few investments to expand distribution capacity in such locations, as is the case in the Amazon Region. In this sense, the objective of this work was to propose a methodology to evaluate the impact of combining video content distribution strategies using CDN networks, content segmentation and selective distribution considering the scenario bandwidth conditions as a means of distributing content for scenarios with limited infrastructure, in order to guarantee the distribution of multimedia content to locations with infrastructure limitations. After analyzing the conditions of limited infrastructure and the main scenarios found in the literature, three (03) distribution scenarios were chosen and implemented. Then the choice of the simulator was made, since the implementation of a CDN architecture requires a complex operational structure, which involves the use of equipment, communication links and technologies for content management. A third step was to measure the performance of the CDN network to identify its ability to serve customers with the content to be replicated, as well as to enable the monitoring and performance management of a CDN network. According to the results presented in the simulations, when distributing the videos in all versions without the use of segmentation, the times required for transmission were the highest presented, especially in the scenarios with the lowest bandwidth available. It also showed that when using other technologies in conjunction with CDN networks, the obtained metrics were better and that when the distributed videos were segmented, the distribution process got even better metrics.