We developed a formal model that supports independent reasoning about the correctness of both the underlying active network platform and the algorithms injected into it, in a manner that admits the full range of possible active network programming interfaces. We have developed a specialized form of program composition that captures the interaction between the underlying platform and the user-injected program and allows properties of each to be preserved. We have applied our results to reason about non-trivial compositions of user-programs and active networking platforms including verification of a complete mobility algorithm introduced into the network. This is the first example of formal verification of correctness of a composite service that is injected into an active network. This work was submitted to ICNP'98.
We extended our simulator, AN-Sim, to support experimentation with multicast congestion control. We have completed preliminary tests of the behavior of active multicast congestion control and comparison with traditional layered multicast as a method to deal with receiver heterogeneity. We are continuing with this work and plan a submission to Infocom'99.
We initiated investigation into several new areas, including active support for wide-area server selection, management techniques for soft state at active nodes and automated virtual topology creation and management.
We received an unrestricted gift from Novell in support of the CANEs project.
We plan a submission to Infocom'98 on multicast congestion control. This will include results from our simulator and comparisons to traditional layered methods of dealing with receiver heterogeneity.
We plan a submission to IEEE Communications Magazine for a Feature Topic Issue on Programmable Networks. This paper will focus on architectural issues and directions for active networking.
We will continue to develop support for experimentation with a richer set of ABONE topologies than are currently possible. We hope to have a preliminary version of this complete for the July DARPA PIs meeting.
We will complete a collaboration plan to allow the project to successfully continue as Ken Calvert moves to U. of Kentucky.
We will continue work with DARPA on the Active Networks PIs meeting in Atlanta in July 1998. We have established a local contact person (Linda Williams) to coordinate administrative details with Brett A. Motiff.