Private Domains in Networks of Information

The networking research community has started looking into so-called information-centric networks, which regard information objects as the primary elements in the network. Users store and retrieve the information objects, denoted by object identifiers, from the network without knowing or bothering which host serves the actual request, comparable to earlier initiatives in networking that separate node identifiers from network locations. The capability to address content directly takes many optimizations, such as content- optimized routing or transparent caching, to the next level, while at the same time enabling novel services. In order to prevent unauthorized access to the information objects that roam around freely in the global network, objects shall be encrypted. We argue that for some use cases, e.g. when dealing with sensitive company-internal information, a stricter partitioning of the network is required. We propose and discuss two solutions to this problem, one focusing on securing the name resolution procedure, the other concentrating on the actual retrieval. Both mechanisms complement existing proposals by separating information objects in the private network domain while maintaining global reachability.