Fundamental Design Issues for the Future Internet

This paper considers whether the Internet should adopt a new service model to satisfy the traffic characteristics and service requirements of the variety of applications that are growing in demand among the general public. One example of the problems posed by such a variety of applications is that real-time multimedia applications not only suffer from the delays and dropped packets typical of the Internet architecture, but also do not back off in the face of congestion - in the process, data applications that adhere to congestion control receive very little bandwidth. Essentially, the primary goal in designing a network architecture is to optimize the efficacy, or total utility to various applications, of the architecture. This can be achieved by supplying more bandwidth or by extending the service model to accommodate applications requiring different classes of service. The authors condense the trade-off between increased bandwidth and extended services to the fact that while delivering real-time suitable service to all Internet applications is not feasible, extending the service model for real-time applications would automatically free up bandwidth for data applications. To maintain the service interface and the layering model, the particular class of service should be "explicitly requested" by a flow rather than detected and "implicitly supplied" by the network.

The authors add further commentary to the trade-off between bandwidth and mechanism. Extensions to the service model could be rendered superfluous if incorrect assumptions are made about application service requirements. On the other hand, additional bandwidth may not be extremely inexpensive and this would result in a suboptimal architecture that cannot be feasibly provisioned. The authors clearly state that their intention is to open a discussion and present the issues rather than to present definite solutions. This paper didn't seem to add as much value as some of the others since it simply touches upon issues that are somewhat obvious. However, it does present a picture of the future direction of the Internet and the issues that must be resolved to make it flourish even more. This is not essential but it's worth reading and keeping in the syllabus.