It is a fundamental principle of networking that useful communication requires support for common protocols. That is, deployment of a new distributed application requires that whatever communication services it uses be implemented in the network and in all participating computers. For a moderate set of applications (electronic mail, file transfer, remote login) a small set of building-block services (as provided in the Internet by TCP/IP) suffices. However, as computation becomes more distributed, human interfaces become more sophisticated, and the number and variety of network-based applications grows, existing protocols must be stretched to provide the needed services. This is happening in the Internet today, with the growth of multimedia applications and the World Wide Web. Unfortunately, because protocols are typically complex distributed algorithms, they are expensive to develop and deploy. Thus, the need for common protocol support can become a barrier to the introduction and use of new distributed applications.
My current research approach focuses on frameworks that make it easier to deploy new protocool building blocks, and to configure existing ones to provide services better tailored to the application. The CANES and ActiveCast projects represent the application of this strategy to the shared nodes of the network, while the Tau project focuses on end systems.
| Office: | 230 Hardymon Building |
| Phone: | +1.859.257.6745 |
| Email: | calvert@netlab.uky.edu |
| Mail: | Laboratory for Advanced Networking |
| Hardymon Building, 2nd Floor | |
| University of Kentucky | |
| 301 Rose Street | |
| Lexington, KY 40506-0495 |