On the Performance of End-to-End Routing in Complex Networks with Intermittent Links

Emergence of IoT (Internet of Things) applications poses challenges on the networking infrastructure since those applications must accommodate a large number of end nodes (e.g., smart sensor devices), and the communication among those nodes are unreliable. In the last decade, DTN (Delay/Disruption-Tolerant Networking) has been actively studied by many researchers, which aims to provide efficient and reliable end-to-end communication in environments where end-to-end paths can not be reliably established. In DTN re-search, superiority and inferiority of several classes of routing mechanisms have been clarified. However, it is still an open question how effectively or ineffectively end-to-end routing performs in networks with moderately intermittent communication links. In this paper, we therefore address the following research questions: (1) does end-to-end routing perform effectively in a large-scale network with many nodes, each of which is connected with a few other nodes via intermittent communication links? (2) how is the average end-to-end message delivery delay affected by the degree (i.e., the total number of incoming and outgoing links) of source and destination nodes? To answer the above research questions, we analytically derive the average message delivery delay with end-to-end routing on a complex network with an arbitrary degree distribution.