Improved content addressability through relational data modeling and in-network processing elements

Recent realizations of the ICN principle organize content in a hierarchical namespace. We argue that this addressing mode has shortcomings because a single document could be part of several data collections. For instance, Joe's record of his New York Marathon run might be published as a content object with the name /repo/events/NYmarathon/record1234 but would also fit into /repo/users/Joe/record1234. Even further, the content of documents can be very multifaceted such that not all details -e.g. spacial coordinates, timestamps, rankings- can be made available on the name surface. In this paper we show that CCN-style networks enriched with active elements, i.e. content processing/producing entities, can overcome these problems. Exemplarily, we adopt relational data modeling concepts to organize named data and deploy Named Function Networking to implement content addressability which goes beyond the scope of pure, i.e. passive, CCN.