Adaptive semi-stateless forwarding for Content-Centric Networks

Abstract In Content-Centric Networks, users request content by issuing Interest messages, receiving Data messages in response. Interests leave state on routers on their path, to allow Data to follow the reverse path. Despite the advantages made possible by this state, its storage requirements raise scalability concerns. We propose an adaptive semi-stateless forwarding scheme where Interests are tracked only on a fraction of the routers. Between state-tracking routers, Interests gather reverse path information, which is used to deliver Data via Bloom filter-based forwarding. We show how the fraction of state-tracking routers can be dynamically adapted to reduce state while limiting communication overheads. Our scheme allows this state, along with all other required data structures, to comfortably fit into the fast memory of a practical router.