Youtube traffic from the perspective of a developing country: the case of Uruguay
Diego Kiedanski, Mateo Nogueira, Eduardo Grampín · 2019
Understanding the interconnection of developing countries with the global Internet has both practical and theoretical relevance. On the one side, practitioners such as service providers and network managers would be particularly interested for both commercial and technical reasons, while on the other side, revealing general knowledge about the insights of such interconnection may be of interest for both academia and policy makers. In this study we concentrate in our country Uruguay. Studying the network topology is one of the most intuitive ways to start, but unfortunately this approach does not include traffic over the network or the behavior of users. According to the Cisco Visual Networking Index, the OTT (Over-The-Top) services are one of the biggest responsible for global traffic, with 75% of Internet traffic being video. Therefore, it is natural to expect that when studying where, when and how the OTT content comes from, we would shed light on how our country makes use of the Internet. Due to its relevance, we propose to study YouTube traffic. The content of YouTube (and OTT in general) is provided from CDNs (Content Delivery Networks), large distributed networks which store and distribute content opaque to external users. This makes the problem of studying the origin and traffic routes much more complex. There are several works that have tried to locate the servers and caches that provide content to a specific ISP, mostly combining traffic analysis techniques with diagnostic tools (such as ping and traceroute). The existing bibliography has explored these questions especially for actors in Europe, but little or nothing has been done in Latin America and Uruguay. We seek to reproduce (and hopefully improve) these results in our country and geographically locate (as far as possible) the different servers in which the content we consume is found, and as these change over time, which routes are used to reach our country, among other questions. On the other hand, once you have the information of the servers of origin, many questions arise naturally: is there any correlation between the proximity of the servers and the popular content in Uruguay?, and among the videos recommended to users? In this paper, we report experimental results revealing some interesting aspects of this YouTube's content distribution hierarchy and cache policies, looked from the perspective of an Uruguayan user.