IP networks are built around the idea of best effort networking, which makes no guarantees regarding the delivery, speed, and accuracy of the transmitted data. While this model is suitable for a large number of applications, and works well for almost all applications when the network load is low (and therefore there is no congestion), there are two main factors that lead to the need for an additional capability of quality of service guarantees. One is the fact that an increasing number of Internet applications are related to real-time and other multimedia data, which have greater service requirements in order to be satisfying to the user. The other is that Internet usage is steadily increasing, and although the network infrastructure is also updated often, it is not always certain that network resource offerings will be ahead of usage demand. In order to deal with this situation, IETF has developed two architectures in order to enable QoS-based handling of data flows in IP networks. This article describes and compares these two architectures.
