Traffic Optimization and Congestion Avoidance in Distributed Systems

The way users utilize the Internet has drastically changed in the last decade. Some years ago, the average user used to identify the Internet with the web browsing application. Nowadays there are multiple applications available for the users; the way people do shopping, mantain social relationships, do research and get entertainment has dramatically changed given the new bouquet of applications available on the Internet. Historically, the Internet has a best-effort nature that does not match the requirements of amny of these emerging applications. Real-time, delay-sensitive applications need a more reliable infrastructure. For this reason, new communication models have been developed and new issues arose. In general, we can identify these communication models as Distributed Systems. In this thesis, we investigate the new issues that arose from different new applications (i.e. VoIP, delay-sensitive content delivery and live streaming) and the related communication infrastructure (client/server, Content Delivery Network, Peer to Peer). For each application/infrastructure we describe the issues, we define a goal and then we describe the solution to fulfill such goal. To do so, we present three use-cases and go into details for each of them; at the end we conclude with some general findings obtained from the specific solutions applied to each of the investigated scenario.