Peer Agent Negotiation and Deployment of Adapters

Welcome to the Panda project. Panda (Peer Agent Negotiation and Deployment of Adapters) was a multi-year DARPA-funded research effort  (June 98 to July 02) undertaken by the Laboratory for Advanced Systems Research (LASR) (formerly the File Mobility Group) in UCLA's Computer Science Department.

Panda is a middleware support system for active networks that allows both aware and unaware applications to benefit from the superior adaptability and performance of active networks. It performs planning and negotiation of active network agent deployment, adapts to failures and other major changes in network characteristics, and handles automatic deployment of agent code to active network nodes.

Panda is built on the base of an existing adaptive-agent framework developed at UCLA under our DARPA-funded Travler contract. An existing active network execution environment called ANTS was used as the platform for the middleware services that Panda provides.

In December of 2000, the value of Panda was successfully demonstrated at the Active Networks PI meeting in Atlanta, Georgia.  A pair of network video and audio streams were intercepted from legacy applications (which were unaware of the the active networks environment) and redirected through a series of related active networks software packages. We were able to selectively prioritize one pair of streams based on a determination of who the speaker was.
 

Project Members:

LASR team members have published several important papers that directly relate to this research.
DARPA/ITO Contract No. N66001-98-C-8512