Section: New Results

Platforms and Substitution Networks

Participants : Tony Ducrocq, Lucie Jacquelin, Milan Erdelj, Karen Miranda, Nathalie Mitton, Enrico Natalizio, Jovan Radak, Priyanka Rawat, Tahiry Razafindralambo, Loic Schmidt, David Simplot-Ryl, Julien Vandaele.

Platforms

In the framework of the ANR SensLAB project, a wireless sensor testbed has been set up in Lille in order to allow the evaluation through experimentations of scalable wireless sensor network protocols and applications. All functionalities offered by the platform have then been presented in [17] , [16] , [42] . SensLAB's main and most important goal is to offer an accurate open access multiusers scientifc tool to support the design, the development tuning, and the experimentation of real large-scale sensor network applications. The SensLAB testbed is composed of 1024 nodes over 4 sites. Each site hosts 256 sensor nodes with specific characteristics in order to offer a wide spectrum of possibilities and heterogeneity. Within a given site, each one of the 256 nodes is able both to communicate via its radio interface to its neighbors and to be configured as a sink node to exchange data with any other "sink node". The hardware and software architectures that allow to reserve, configure, deploy firmwares and gather experimental data and monitoring information are described. We also present demonstration examples to illustrate the use of the SensLAB testbed and encourage researchers to test and benchmark their applications/protocols on a large scale WSN testbed. A survey of platforms similar to SensLAB can be found in [6] .

Emulation

Although some platforms like SensLAB are very convenient, they do not always fit the application requirements and setting up experimental testbed of large scale wireless sensor networks requires huge cost, space and human resources. A more affordable approach is needed to provide preliminary insights on network protocols performance. To overcome the need for significant number of sensors required to perform a realistic experiment, and/or to experiments with high density networks, we introduce in [43] a novel approach: emulation by using all available sensors as candidate forwarding neighbors of the node S currently holding the packet. Destination position is virtual. After successfully sending message to forwarding node B over realistic wireless channel, the position of virtual destination is adjusted by translating it for vector BS and possibly rotating it to change the neighborhood configuration. The same node S then again selects new forwarding neighbor. Such selection of best forwarding neighbor continues until virtual destination appears close to a real node, and the later then becomes final destination node. Compared to real testbeds, our emulation has advantages of testing networks with very large densities (which may not be possible in small scale implementations), and in unlimited scalability of our physical implementations (e.g. we can emulate network with a million nodes).

Substitution network

A substitution network is a rapidly deployable backup wireless solution to quickly react to network topology changes due to failures or to flash crowd effects on the base network. Unlike other ad hoc and mesh solutions, a substitution network does not attempt to provide new services to customers but rather to restore and maintain at least some of the services available before the failure. Furthermore, a substitution network is not deployed directly for customers but to help the base network to provide services to customers. Therefore, a substitution network is not, by definition, a stand-alone network. [36] describes the quality of service architecture for substitution networks and discuss provisioning, maintenance, and adaptation of QoS inside and between the base network and the substitution network. In the same context, [33] shows the impact of the router mobility on the QoS of such networks.