A Scheme for Delay-Sensitive Spatiotemporal Routing in SDN-Enabled Underwater Acoustic Sensor Networks
In underwater acoustic sensor networks (UASNs), the sensors are deployed at different areas of the ocean, which perform information collection and delay-sensitive routing to the data center for further processing or industrial computing. However, in UASNs, the network states have spatiotemporal characteristics due to tides or autonomous underwater vehicles. To steadily route the traffic especially when the spatiotemporal characteristics of the UASNs are considered, a network architecture with intelligent traffic engineering or routing policies is indispensable. In this paper, we employ software-defined networking (SDN) technology and propose an SDN-enabled distributed architecture for UASNs. Based on the proposed architecture, we propose a scheme DSR-SDN for delay-sensitive spatiotemporal routing in SDN-enabled UASNs. The DSR-SDN includes three phases: First, topology awareness; second, spatiotemporal characteristics estimation; and third, routing computation. Particularly, with SDN features, DSR-SDN provides topology awareness based on a proposed software-defined beaconing scheme. Based on the detected topology, the spatiotemporal characteristics of the network states are estimated based on a proposed SDN-based hierarchical node localizing approach SDN-HL. Lead by the SDN controllers, SDN-HL makes use of the proposed 'minimum weighted gap' formulation and Adam algorithm to optimize the localization and builds the indirect links to increase the localization rate. To route the traffic through the network with spatiotemporal characteristics, we adopt the time-expanded network approach, based on which a spatiotemporal route decision can be made before the routing starts. The simulation results demonstrate that the proposed scheme, i.e., DSR-SDN, can conduct accurate spatiotemporal characteristic estimation for the network states and provide delay-sensitive spatiotemporal routing for the sensed data. - 1967-2012 IEEE.
- Computer Science & Engineering [598 items ]