Internet of Things
5 Validation, Results and
5.6. DEMANES Integration Scenario
The self-adaptive system was also integrated into a complex scenario for the DEMANES European research project. In this scenario SunSPOT nodes were considered for providing a backup network for the collecting of data from cargo containers. This network was dynamic by definition, is it can be formed by nodes deployed in moving cargo containers, along a cargo ship, or in the facilities of a cargo stations.
The integration was achieved by connecting a sink node to a common gateway platform running on a Beagle Bone Black board [88]. The common gateway used OSGi
[89] to interconnect the components from different partners. Therefore a specific OSGi component for interacting with the SunSPOT network was developed.
Figure 67: Cargo container and deployment area for DEMANES integration
The integration tests were conducted at the premises of TNO in The Hague. A sensor combo box were developed by one stakeholder, combining a G-Node [90]
sensor node with a BeagleBone Black to provide extra calculation power. These nodes were deployed inside the container to monitor the cargo. Another combo box was embedded in the wall of the container, providing connectivity from the inside to the outside. This combo box had another G-Node acting a as a sink for the internal WSN, as well as a SunSPOT base station and a 3G modem dongle, all connected to a BeagleBone Black board (Figure 67). An OSGi framework was running in this gateway box, providing access to all the functionalities. In our case, the access to the SunSPOT network was provided by a SunSPOT adapter OSGi component.
The normal operation mode for the scenario was to collect the data from the G-Nodes inside the container and submit them to an Internet service to storage and further processing. This transmission was done using the 3G dongle when available, or redirected to another full box through the SunSPOT network in any other case.
So that was the test performed. Once the SunSPOT network was deployed, we simulated an error in the 3G connection. The messages collected from the G-Nodes were effectively redirected to other available gateway in the network using the SunSPOT network. And SunSPOT devices were able to maintain the connectivity by
adapting the transmission power in the occurrence of any event that changed the topology, like interferences, appearance of new nodes, and disappearance of those already known (transient nodes, battery depletions and hardware failures).
Figure 68: SunSPOT Sensor node deployed for DEMANES integration tests
The SunSPOT devices were able to perform other duties, like sensing other environmental conditions in the cargo station. One of the foreseen scenarios was to include them in a cargo ship, monitoring external conditions of the cargo containers, and advising in the occurrence of unexpected events.
5.7. Summary
In this chapter we have introduced the validation analysis performed on the proposed self-adaptive system. As we introduced in the previous chapters, the system has two main goals:
• Reduce the energy consumption.
• Maintain the connectivity reliable.
The energy is maybe the most crucial resource, especially when dealing with constrained networks and devices, like the ones typical in a WSN. It is not only a goal itself in the scope of strategies like Europe 2020, but also an important constraint that imposes itself as a non-functional requirement for most WSN applications nowadays.
Acting upon the transmission power can effectively reduce the total amount of energy consumption in the network. But saving energy cannot imply a negative impact con the connectivity. If the WSN topology is not going to be changed, the optimal transmission power for each node can be estimated, although environmental conditions can introduce unexpected changes in the topology. Therefore either in those situations a self-adaptive system can provide a good way to reach the optimal configuration after variations in the topology.
The experiments conducted in this chapter collected information about the network energy consumption and connectivity reliability using the proposed self-adaptive system. In the outdoor open area scenario we have seen how a good choice of configuration parameters provides a better energy consumption rate while keeping a good connectivity. Indeed, the parameters used in experiment e05 obtained an energy savings of 11% over the consumption due to use a fixed transmission power set to the available maximum. And what is more important, this experiment also obtained an outstanding connectivity performance with a PDR close to 99%.
In these experiments we also obtained that the same experiment provided the lower number of changes in the transmission power, thus it is the one with better dynamics. Moreover, we were able to determine that the node degree, that is, the number of active neighbors at any given time, has lower effect on the performance than other parameters like the tolerance and the normalization factor for the transmission power adjustment loop.
We also have analyzed the software metrics corresponding to the implementation of the self-adaptive system in a real system. The code overhead, for instance, has a really small footprint on the devices used, requiring just a 2.06% of the available storage capacity. The same applies to the volatile memory size, representing a 19.30% of the total memory in the devices.
Finally, as this proposal was part of the DEMANES European research project, a brief description of the integration scenario and tests has been provided. The system was deployed in a small sensor network composed of a pack of SunSPOTs. The provided environmental monitoring and backup networking capabilities for the collecting of data captured inside a cargo container by other WSN deployed by other partners.