A Group-Based Architecture for Wireless Sensor Networks Jaime Lloret, Miguel Garcia and Jesus Tomas Department of Communications Polytechnic University of Valencia Valencia, Spain jlloret@dcom.upv.es, migarpi@teleco.upv.es, jtomas@dcom.upv.es Abstract—Many routing protocols for ad-hoc networks and sensor networks have been designed, but none of them are based on groups. We propose to divide the network into several groups of sensors. When a sensor send data to other groups, the data has to arrive just to one sensor from each group, then they propagate it to the rest of sensors in their groups. We have simulated our proposal for different types of sensor topologies to know which type of topology is the best depending on the number of sensors in the whole network or depending on the number of interior sensors. We have also simulated how much time is needed to propagate information between groups. The application areas for our proposal could be rural and agricultural environments to detect plagues and to propagate it to neighbouring areas, or for military purposes to propagate information between neighbouring squads. Keywords-Sensor Network; Group-Based Architecture; Group- based routing algorithm. I. INTRODUCTION There are many routing protocols that can be applied to sensor networks. They can be classified into two groups [1] [2]. One group is formed by protocols based on the network topology and the other group the ones that do not take it into account. First group can be broken down into three subgroups: 1- Plane routing. All nodes in the network have the same role and perform the same tasks. Because of the number of nodes in these networks, the use of a global identifier, for every node, is not feasible. It uses a data-centric routing where the base station sends requests to some regions and the nodes from that regions reply. Some of the algorithms in this group are SPIN, Direct diffusion, Rumour routing, MCFA, GBR, IDSQ, CADR, COUGAR, ADQUIRE, and so on. 2- Hierarchical routing. It is very scalable and has an efficient communication. It has been designed for energy saving purposes, because central nodes have unlimited energy, while leaf sensors have limited energy. When the sensor network topology is formed, data can be routed. Some algorithms such as LEACH, PEGASIS, TEEN, APTEEN, MECN, Virtual grid architecture routing and TTDD are hierarchical routing algorithms. 3- Position-based routing. All data is routed through the sensors depending on their position. Distances between sensors are known because of neighbouring sensors signals. There are other protocols that base node’s situation on GPS and, using that information, route the data to the most adequate sensor. These algorithms consume more energy than others because of the need of GPS signal. Some of those algorithms sleep sensors when the network has not any activity. Some examples are GAF, GEAR, GOAFR and SPAN. Second group does not have into account the structure of the network. It can be broken into five subgroups: 1- Multipath Routing Protocols. The information could reach the destination through different paths. Because sensors have to calculate several paths, they use a main route when they have enough energy; otherwise, they use an alternative path. 2- Query-Based Routing protocols. They are based on a central node that sends a query about an event to the specific area. When the query arrives to that area, it is routed to the destination sensor, and then it will reply. A sensor from an area could be sleeping, saving energy, while there is not any query to that area. 3- Negotiation-Based Routing Protocols. Before data transmission, the sensor has to negotiate the data it has to send, so redundant data could be deleted, and resources will be available while data exchange. SPIN protocols use this type of routing, but they take into account the network structure. 4- QoS protocols. The information is routed to the sensors taking into account quality parameters such as delay, energy, bandwidth and so on. SAR and SPEED protocols are based on quality of service algorithms. 5- Data coherent/incoherent processing based protocols. These algorithms use several routing techniques taking into account the data processing of a coherent or incoherent result. None of the routing protocols aforementioned are group- based. We propose to divide the network of sensors into several groups and if a sensor has to send data to other groups, when this data arrives to one sensor from a group, it propagates it to the rest of sensors in its group. The paper is structured as follows. Section 2 examines some works related with our proposal such as neighbour selection and architectures based on groups, and explains our motivation. There is a description of our architecture proposal in section 3. Analytical model for some types of topologies of sensors are shown in section 4. The propagation time to reach a sensor from other group is analyzed in section 5. Finally, section 6 gives our conclusions and future works.