Robust and Adaptive Protocol Design for Multimedia Wireless Networks

A Stanford Networking Research Center Project


The main theme of our proposed work is to design a robust and adaptive protocol suite to insure reliable and high-quality end-to-end performance in wireless multimedia networks. We propose a three-fold approach to achieve this goal: distributed cross-layer adaptive protocols, network diversity, and network scheduling. Our previous work has shown that the variability and volatility of wireless networks can be exploited by adapting to the wireless link, network, and application dynamics. Such adaptation is typically only feasible if it can be done in a distributed manner. We have also shown transmission scheduling to be a powerful mechanism to insure a given performance level and optimize network resource utilization given underlying network dynamics. However, wireless networks inherently maintain some uncertainty that cannot be adapted to or scheduled around. We propose to conquer this uncertainty by developing diversity techniques for each layer of the protocol stack. Network diversity is a new and innovative research thrust for network design. Diversity in wireless links - using multiple antennas to reduce the random variations due to multipath fading - has been exploited for several decades, yet the basic concept of diversity has not typically been used in other layers of the network protocol stack. We propose to develop network diversity techniques that include access diversity, path diversity, content location diversity, application diversity, and resource diversity. These diversity techniques will provide robustness to random variations at each layer of the network protocol stack, and will also be integrated with adaptive techniques at each layer. Diversity also provides enhanced network security and reliability in addition to robustness. We will also continue work on adaptive cross-layer protocol design to jointly adapt and optimize the behavior of individual protocol layers to each other. Specifically, we will explore link adaptation, dynamic allocation and management of network resources, and application adaptation. Our work in this area will mainly focus on distributed adaptive techniques. We will also study the integration of network diversity with adaptive protocol design. Our prior work has shown that scheduling provides robust and efficient performance in wireless networks. Scheduling insures sufficient resources will be available in the network to meet the performance requirements (e.g. rate and delay) of a given application. It can also be used to minimize the energy requirements in the network. We plan to extend these results to develop and analyze new and innovative scheduling techniques that maximize efficiency and robustness in multimedia wireless networks. Our proposed work will have a high impact on wireless network design. It is now well-recognized that cross-layer design is critically needed to insure continuity, robustness, and good end-to-end performance in multimedia wireless networks. Yet, this is a highly complex design problem that requires multiple levels of expertise across the network protocol stack. We believe the depth and breadth of our research team makes us uniquely positioned to develop and analyze strategies for cross-layer design techniques. In addition, our ideas for scheduling and diversity in network design are new, innovative, and have the potential for significant impact on network design and performance. Our team's collaboration over the last two years has yielded important concrete research results and, perhaps more importantly, substantial cross-fertilization across disciplines resulting in broad new ideas for network design. The next phase of the project will place even more emphasis on this cross-fertilization and on the exploration of the resulting innovations. This project is a continuation of the SNRC project "A Multilayer Approach to Mobile Networking" (2000-2002).
Last modified on November 7, 2003.
Comments to webmaster@wsl.stanford.edu