Poster Abstract: Adaptive Routing in Underwater Delay/Disruption Tolerant Networks

Underwater sensor network has attracted significant attention recently [1] [2]. On the one hand, it enables a wide range of aquatic applications. On the other hand, harsh underwater environments pose grand challenges to underwater communication and networking. Firstly, acoustic channels have high error probability, long propagation delay and limited bandwidth. Secondly, underwater sensor nodes are usually mobile with water currents. Furthermore, compared with terrestrial sensors, underwater sensors are usually much bigger, more energy-consuming and more expensive. Thus it is very costly to deploy dense underwater sensors in a large area. Due to node mobility and sparse deployment, an underwater sensor network often becomes an Intermittently Connected Network (ICN), that is, there may not be an available persistent route between a source and a destination. This kind of network is referred to as Delay/Disruption Tolerant Network (DTN). Traditional routing protocols are not practical for such scenarios since packets will be dropped when no routes are available. Therefore, new routing protocols are required for efficient message delivery in such networks. A large number of routing protocols have been proposed for DTN. One extreme is epidemic routing [3], which is a flooding scheme. Without network constraints, such as contention, buffer, battery and bandwidth, epidemic routing can always maximize the delivery rate, or minimize the average delay. However, flooding based epidemic routing will consume too much energy which is precious in underwater sensor networks. Another extreme is single copy forwarding [4], which aims to achieve the best tradeoff between message delivery delay and energy efficiency. Many other protocols such as [5] [6] and [7] are between these two extremes. They are based on either flooding or gossip. They utilize some statistic profiles to route packets to the destination using multiple copies as well as some control mechanisms. All these protocols treat packets equally and aim at optimizing a single objective, e.g., minimum average delay or maximum delivery rate or efficient energy consumption. In this paper, according to the special constraints of underwater delay/disruption tolerant networks, we propose an adaptive routing protocol for efficient routing. We assign packets different emergency levels and define several routing states. Afterwards, we adaptively choose appropriate routing state to satisfy different application requirements by utilizing an information vector (see Section II). The objective of our protocol is to achieve a good tradeoff among delivery ratio, average delay and energy consumption. Our contributions are two folds. First, we propose an efficient routing protocol for underwater delay/disruption tolerant networks with special constraints. Second, we treat packets as application-oriented to adaptively choose appropriate routing states and achieve a good performance tradeoff.