Article Preview
Top1. Introduction
A wireless sensor network is a collection of sensor nodes (small sensing devices) and a sink node connected through wireless channels. These small devices can be used for building distributed systems for data collection and processing (Wang, Cao, Li, & Dasi, 2008). Nodes in wireless sensor networks have restricted storage, computational and energy capacity. The network as a hole is characterized by a limited life time due to the energy constraint. Generally, the sensor nodes detect events, performed desired measurements, process the measured data and transmit it to the sink node (Chalak, Sivaraman, Aydin, & Turgut, 2006). However, nodes may also generate queries to find events or locate data and services they are interested in within the network (Banka, Tandon, & Jayasumana, 2005; Zhang, Yu, Zhou, Lin, & Li, 2009). Sensor networks are suitable for data collection within regions that human beings are unable or hard to reach, for example, battle fields reconnaissance, disaster (such as earthquake, fire, etc), and areas monitoring (Yu, Chen, Wu, & Jin, 2008). Other examples applications include health monitoring, intelligent building, etc. (Wang, Cao, Li, & Dasi, 2008).
Routing protocols in wireless sensor networks have to ensure good performances in terms of energy consumption, scalability and increasing network life time. Different routing and dissemination protocols were proposed for wireless sensor networks. Flooding was one of the earliest mechanisms used in routing and dissemination protocols. In Intanagonwiwat, Govindan, and Estrin (2000) flooding events and queries is used to find the shortest path, but more power consuming and bandwidth was generated as a result of broadcasting storm.
To preserve energy consumption, routing algorithms with mobile agents seems to be more suitable for wireless sensor networks (Braginsky & Estrin, 2002; Banka, Tandon, & Jayasumana, 2005). Instead of flooding technique, agents may be used to spread environmental and request information. They can cooperate to build optimal paths leading to events and queries. The problem with the agent based routing protocols, especially in environments with no geographical localisation systems, is the spiral like routing paths and the non uniform distribution of agents in the network. The proposed Fast Rumor Agent protocol is designed to fix the spiral like routing path problem. In FRA protocol, a new agent propagation approach is introduced, allowing a fast agent transfer between nodes in the network.
Our propagation approach allows agents finding new zones and avoids coming back to the neighborhood visited zones which saves energy and time during path establishment.
The rest of this paper is organized as follows: Section 2 discusses related researches on agent propagation for routing in wireless sensor networks. The proposed FRA protocol is introduced in Section 3. Section 4 presents comparison and scalability results of the proposed FRA protocol, while Section 5 concludes the paper.