Changing the Internet to support real-time content supply from a large fraction of broadband residential users

Abstract

Adding quality of service (QoS) to the Internet raises a number of issues as to why we would want it, as well as how we would do it, and whether we could subtly change the user experience compared to, say, public switched telephone service voice. Much of QoS work today is focused on improving the service over the access pipe between the end user to the ISP. But this does not adequately cover end-to-end QoS experiences for the majority of user interactions involving real-time content over the Internet. This paper explores these issues and raises the possibility of ‘mass content provision’ as a key advantage of such end-to-end Internet QoS. It allows a large proportion of broadband users to become real-time content providers. Examples cited are talking/explaining sequences on a home movie while pausing and rewinding, as well as selling real-time content. This paper discusses a proposed solution to adding QoS to the Internet that can be fairly easily added on to what already exists. Furthermore it is ‘mobility friendly’. It automatically tracks and minimises flows affected by poor QoS as users move. Simulation results are presented showing the power of the method. A number of open issues and standards issues are highlighted.