Scott Nettles is an Associate Professor in the Electrical and Computer Engineering department at The University of Texas at Austin. He received a Ph.D. in Computer Science from Carnegie Mellon University in 1996, where his worked on high-performance garbage collection and transaction systems. From 1995 to 1999, he was an Assistant Professor of Computer and Information Science at the University of Pennsylvania, where he was the recipient of a 1997 NSF CAREER Award. While at Penn, he was part of the DARPA funded SwitchWare project, which was one of the pioneering groups in the area of Active Networking. His group developed PLAN, the first domain-specific programming language for programmable packets, and PLANet, the first purely active inter-network. Since coming to The University of Texas at Austin in 1999, he has continued his efforts on the design and implementation and application of Active Networks, including developing the first second generation programmable packet language, SNAP. His has several current focuses, including addressing the limitations of Active Packet systems in the area of resource control and applying Active Networking techniques to network security. Finally, his major current focus, is applying new architectural approaches (especially Active Networking) to the emerging area of mobile/wireless networks.
Frontiers of Active Networking Modern computer network architectures, as exemplified by the IETF TPC/IP suite, have been wildly successful. However, there are key limitations, especially in our ability to customize such networks for specific applications or network environments and in our ability evolve existing networks to support new services. Active Networking (AN) is an effort to address these limitations by creating a network architecture in which the network infrastructure is programmable. Researchers have had considerable success defining such architectures and a variety of prototype systems have been built. own efforts have focused on both of the two principle programmability approaches, programmable (or "Active") packets, and extensible routers. One of efforts, PLAN, is one of the primary Active Packet programming languages. system, PLANet, combined PLAN with dynamically loadable router extensions to create the first purely active internetwork.
My current research seeks to go beyond these preliminary "first generation" efforts to address the difficulties and limitation of those systems and to create the next generation of AN systems. Some of my recent efforts include addressing limitations on the scope of router extensions with work on dynamic updating and SNAP, an active packet system that has both qualitatively safer than PLAN and has performance competitive with Linux IP implementations. Another trust is to find new applications of AN that go beyond the simple examples of the first generation. One application area is network security, where I am working on building highly programmable firewalls and "smart" distributed security agents. The other key application area is to use AN technology to build mobile networks. Mobile network protocols are rapidly changing and AN's ability to support rapid evolution is key. Further, because Active Packets allow protocols to change on a packet-by-packet basis, AN provides technology to support adaptation to the rapidly changing wireless link conditions found in mobile networks. I expect significant collaboration with other WNCG faculty in this area.
