User-centered control of quality of service

作为互联网的用户，我们已经开始期望以统一的费率获得广泛的服务，并拥有一个不断扩大的“数据管道”，使我们能够更快地传输更大的电影。这样的模式是否具有商业意义或理论意义并不是我们的重点。相反，我们认为仍有一些情况会达到网络容量的极限，如果不是在接入中，那么就是在网络中更深的地方。在这种情况下，区分服务模型可以帮助更多性能关键型应用程序，但哪些是关键应用程序，以及如何在不破坏互联网模型和让运营商控制我们的环境的情况下识别它们。

As users of the Internet, we have come to expect to receive a large range of services for a flat rate, and to have an ever widening ``data pipe'' allowing us to transfer bigger movies faster. Whether or not such a model makes business sense or theoretical sense is not our point. Rather, we consider that there are still occasions when we will reach the limit of the capacity of the network, if not in the access, then further inside the networks. In such a case, a differentiated-services model can be useful to help more performance-critical applications, but which are the critical applications, and how can they be recognized without breaking the Internet model and giving the ISP control over our environment.