Tussle Networking
This presentation was the output of some thought process between Karen Sollins (MIT) and myself on information networking and its vision beyond the mere observation that more and more traffic in the Internet relates to information rather than point-to-point communication.
With that, we formulated the vision that an information-centric design of the Future Internet could drive us towards an architecture that would allow to execute (including reasoning and mediating) policies as an expression of concerns that players within the Internet have concerning certain services. With that, the Internet becomes more of an execution environment than a delivery environment, as it is today.
What is the promise of this? Not only does it shed a different and interesting (to me at least) light onto information networking, but it also promises to reduce the costs for development and deployment of communication solutions in the future by virtue of executing policies (and therefore resolving tussles in the marketplace) rather than following today’s approach of developing and deploying parallel yet rather similar architectures, each of which reflect the policies and tussles of the players within (think of the ‘mobile’ Internet as defined in 3GPP and the IETF way, isn’t IMS and vanilla SIP really similar?).
Lots of the work required for this vision has been ongoing for many years. Starting from the provisioning plane, publish-subscribe has been around in research for a while, getting new uplift through recent work on, e.g., DONA (Data-oriented Networking Architecture) that attempts remove engineering roadblocks to the vision of flat label routing. In the information space, the vision of the Semantic Web has been fuelling research for quite a while now. In particular semantic mediation is crucial for the vision of tussle networking and recent work in this area is more than just promising. One of the biggest challenges to me is the end user experience research. With the community only relatively recently focussing on this branch of research, there is still lots to do to come to solutions that will not only help end user navigate the vast information space but also express concerns (translated into policies) semi- or fully-automatically.
As a side note, the tussle networking vision also sheds an interesting light on the vision of the Knowledge Plane, as formulated by Clark et al. some years back. The policy focus of the tussle networking vision makes, so I believe, easier to understand how one could ever implement the vision outlined in the Knowledge Plane by virtue of detecting, mediating and resolving policy conflicts among parties involved in a service that does not work well. And all this initiated by an end user that observes mis-behaviour. It would put an end to the endless questions we often have as to why my Internet service does not work, my webpage does not load or the connectivity isn’t there.
Send me an email in case you’re interested in the presentation slides.