Activecast Project, University of Kentucky

Recent Active Networks research has produced programmable network platforms that support an unprecedented level of user customization of network services to enhance and enable new forms of application. Presently, however, the power of active networks is mostly ``untamed''. To take advantage of it, applications must be written in a particular language and/or must invest in a substantial programming effort to interface with the features offered by the various active network execution environments. The next step in the evolution of active networks --the one that will support radical new uses of the network and increases in scalability-- is packaging the power of a programmable platform into services that are easy for applications to use, yet customizable. The ActiveCast project will develop a set of services that will not only dramatically enhance the ``application-friendliness'' of active networks, but will also improve the scalability and usability of networks in general.

Examples of these network services are: These application-level services are being implemented using a set of lower-level building blocks, including the execution environment we have developed in the Composeable Active Network Elements (CANEs) project. The building blocks developed for the project will also be useful in themselves. For example, the Network Recon service allows users to probe and collect information about the interior of the network; such a service is likely to be useful in a wide variety of active and non-active contexts, including congestion avoidance and control.