This dissertation work specifies an architecture to reliably send sensors’ data from sense units to cloud UrbanSense server using a vehicular network. The proposed architecture relies on the IBR- DTN protocol utilisation as a framework, in order to developed applications communicate through the DTN, as well aspects an application running on the cloud UrbanSense server that accepts TCP socket connection, receives sensors’ data, uploads it to the cloud UrbanSense database and answers acknowledgement data about sensors’ data uploading success or failure.
Metadata information is added to sensors samples, in order to understand the overall system performance. Statical metadata analysis, permits to understand the delay since bundles leave sense units until reach the cloud UrbanSense database, the retransmitted sensors samples percentage, the amount of sensors’ data transferred in a period of time or in opportunistic contacts and how bundles amount of sensors’ data varies.
The proposed architecture was validated in a real-world environment single test, conducted during one hour, thirty two minutes and forty five seconds. This test metadata analysis permitted to conclude that sense units to cloud UrbanSense database sensors’ data transmission through the vehicular network opportunistic contacts can be a reality at an urban-scale and that the developed applications perform very well. Moreover, the bundles delay varies in accordance with the roads traffic conditions. The minimum, mean and maximum bundles delay were 27 s, 140 s and 257 s, respectively. The percentage of sensors samples transmitted in first, second and third transmission attempts was 75%, 20% and 5%, respectively. However, the sensors samples transmissions that failed had been attempted to be transmitted over TCP, before the sending over the vehicular network test started. To sum up, the test had a duration of 5565 s, 11 opportunistic contacts between the sense unit and the OBU and were transferred 250 bundles with a total 149.2 mil characters of sensors’ data.