A topological and performance metrics approach to the decision making of content delivery networks
Alberto del Rio Ponce, Javier Serrano, Alvaro Lopez, Luis M. Contreras, David Jiménez, Federico Álvarez · 2022
This paper presents experimental work for collecting both topological information together with associated performance metrics observed in a realistic lab setup. The network emulates a Content Delivery Network (CDN) deployment, in a manner that the CDN could take informed decisions about from where, out of the available Points of Presence, deliver the content to any subscriber taking into account not only the own CDN status (e.g., level of consumption of computing resources of the caches) but also information of the underlying connectivity network (e.g., topology and performance). In this situation, when similar topological information is provided, the performance characterization of the different paths provides a valuable point of decision. The paper also proposes an optimizer based on the topological and performance data collected to select the most appropriate path for content delivery.