DADI is an integrative research project at Princeton on discovery, analysis and dissemination of information on the Internet. It focuses on both systems and driving applications for a new model of communication and coordination on the Internet: the asynchronous, event-based model, with particular emphasis on decoupled, content-based publish-subscribe (pub-sub) communication at Internet scale.
Many existing communication systems, involving applications on the Internet, are coupled in 'space' and in time. That is, senders specify receivers by address, and communication largely follows a synchronous, request-reply model. This works well for clients accessing information at known locations, and at request time. However, many emerging applications, such as customized information dissemination, scalable coordination of dynamic applications that are data-driven or composed of autonomous components, and information management and sharing, can be better served by a more decoupled, asynchronous and symmetric model. In the pub-sub model assumed in DADI, temporal decoupling (asynchrony) can be achieved by event-driven communication, and spatial (address) decoupling further achieved by content-based event routing. Decoupled, asynchronous pub-sub communication is increasingly recognized as an important core technology to enhance distributed systems and enable key applications, including information sharing and coordination across disparate agencies, componentized Web Services, and dynamic data-driven applications.
DADI adopts a layered approach, including research in as well as across all communication, system, algorithm, and application layers. All works are integrative and application-driven, with research at each layer driven by requirements from the layers above and below, and interactions among layers studied to exploit synergies.
Ongoing research projects can be found here.
Last modified: Tue Mar 16 22:03:30 Eastern Standard Time 2004