D-Cube: A Low-Power Wireless Networking Benchmark
|D-Cube is a full-fledged benchmarking infrastructure that provides a consistent way to evaluate the performance of low-power wireless systems.
Hosted at Graz University of Technology, D-Cube supports the automated testing of the reliability, timeliness, and energy consumption of low-power wireless communication protocols in a variety of settings. The primary audience of D-Cube are academic researchers and industry practitioners creating low-power wireless networking solutions that would like to quantitatively assess and compare their performance with that of other systems addressing the same class of applications.
Differently from other public low-power wireless testbeds such as FlockLab, Indriya2, FIT IoT-Lab, and TWIST, D-Cube is not meant for the general testing of low-power wireless protocols, but rather to evaluate protocol performance in pre-defined scenarios. To this end, D-Cube allows developers to test and optimize their solutions for one or more benchmark suites, i.e., for specific application scenarios and hardware platforms. Detailed performance results of individual test runs are only visible to the protocol developers. Nevertheless, D-Cube also allows developers to provide a binary file that is perpetually tested in a best-effort fashion on specific benchmark suites to accurately profile its performance. Such perpetual performance measurements are publicly available and serve as a reference for academics and companies alike about the performance of consolidated protocols.
D-Cube currently supports two popular hardware platforms: the TelosB (Tmote Sky) mote and the Nordic Semiconductors nRF52840. Hence, D-Cube enables the benchmarking of solutions based on IEEE 802.15.4 and Bluetooth Low Energy technology. With about 50 nodes for each available platform and the ability to generate harsh RF conditions, D-Cube is well-suited for testing the performance of mesh networking and multi-hop solutions in the presence of radio interference.
D-Cube is an effort supported by: