Revisiting Resource Utilization in The Internet: Architectural Considerations and Challenges

Abstract

The Internet has been a success story for many years. Recently researchers have started to deal with new questions that challenge the effectiveness of the Internet architecture in response to the new demands, e.g. overwhelming traffic growth and latency optimizations. Various proposals ranging from new application level protocols to new network stacks are emerging to help the Internet to keep up with the demand. In this dissertation we look at a few different proposals that deal with improving the speed and resource utilization in the Internet. We first discuss improving the resource utilization in the current Internet by minor changes such as adjusting various parameters in TCP. We then discuss a more radical form of resource utilization through combining the network and the available storage. Combining these two resources, which have traditionally been considered separate, could provide many new speed improvement opportunities. We discuss relaxing the barrier between the storage and the network in the context of Information Centric Networking (ICN), which in itself is an alternative proposals to the current TCP/IP style Internet. With the help of ICN, we propose different forms of in-network caching below the application layer. We argue that, although useful, the new models of utilizing network resource could show to have their own challenges. We namely discuss the resource management and privacy challenges that are introduced with ICN in general and within our proposed solutions in particular. The lack of end-host bindings and the existence of network routable data names in different data chunks make the congestion control, reliability, and privacy in ICN rather different from TCP/IP. We discuss some of these differences and propose solutions that can help addressing each issue in our particular form of ICN-based mechanisms.