Storage Replication in Information-Centric Networking

Shifting the communication paradigm of the internetworking layer from endpoints to information has been the focus for a variety of research efforts. Optimized dissemination of information within transient communication relationships of endpoints is the main promise of such efforts. Efficient replication of information is key to delivering on this promise. In this paper, we address this promise by proposing a storage placement and replica assignment algorithm for an information-centric networking architecture. For this, we present a node architecture that is based on a coherent set of architectural invariants that are utilized in the storage solution. By directly utilizing the information structures provided at the new internetworking layer, we enable the differentiation of classes of information based on their popularity with the attempt to minimize the overall network traffic. We show that such an algorithm can be easily implemented in an information-centric network environment and we evaluate our implementation through experimentation in a testbed and simulations, showing the feasibility but also the potential of such solutions.