Real and Simulator Testbeds for Content Dissemination in High-density Large-scale WANET

Yu-Jen Lai¹    Youyang Ng¹    Takeshi Sakoda¹    Yosuke Bando^1,2    Arata Miyamoto¹
Masahiro Ishiyama¹    Ken-ichi Maeda¹    Yusuke Doi¹

¹TOSHIBA Corporation            ²MIT Media Lab

A wireless ad hoc network (WANET) is an effective approach to disseminating digital content without the need for infrastructure networks. It is not only useful in the absence of commodity connections such as cellular networks, but also useful for avoiding overloading those connections in crowded scenarios, an example of which is delivering movie clips throughout spectators in a stadium. However, experiments of WANET have been oftentimes conducted using only up to 1,000 nodes in a simulator or dozens of real devices, both typically scattered across a wide area at a density of around 10^-4 node/m². This paper focuses on the problem of disseminating common content to all the devices in the same area via 802.11-based WANET, and presents a pair of testbeds, one based on real devices and the other a simulator, both capable of high-density and large-scale experiments with roughly matched performance. We have found naive implementation of content dissemination protocols alone not to work in high density, and we identify necessary adjustment of wireless parameters to allow for experiments even with an extreme density of 20 nodes/m². We show that 2.5 MB of data can be delivered to 100 real devices and to 10,000 simulated devices in a few minutes on our testbeds, demonstrating the feasibility of orders of magnitude larger-scale experiments.

Paper (3MB, PDF):
Proceedings of the IEEE Consumer Communications and Networking Conference, 2017.

100-node real-device testbed (left) and 10,000-node simulator testbed (right) capable of disseminating common content from a single node to all the other nodes in high-density large-scale scenarios. With identical experimental conditions and parameters, the performance of the two testbeds roughly matches within a factor of two. The dots in the simulator image show nodes (red waiting for content and green finished receiving it), while the blue lines show wireless links.
Real-device testbed:

The testbed consists of 100 single-board computers. Each 50 nodes form a group and are connected to one central PC via USB, over which we input commands, update content dissemination protocols, monitor network statuses, and collect experiment logs. These wires are for control purposes only, and the devices disseminate content wirelessly.
Simulator testbed: video (93MB, WMV)

The simulator testbed is implemented with NS-3. We modified the primitive modules controlling the behaviors of access points (AP) and stations (STA) in MAC layer, so that nodes can autonomously switch between AP and STA to perform multi-hop content dissemination. The simulation parameters are chosen according to the measurements from the real-device testbed for consistency.

Evaluations:

	Three representative content dissemination protocols were implemented and evaluated on the real testbed at an extreme density of 22.2 nodes/m². Protocol 1: Unicast + RSSI-based relay Protocol 2: Reliable multicast + flooding Protocol 3: Reliable multicast + RSSI-based relay The best-performing Protocol 3 delivered 2.5 MB of data to 100 nodes in 115 sec. The high-density solutions described in the paper were key to enabling the evaluation. Without them, content dissemination did not advance due to severe wireless congestion.
	The real and simulator testbeds have similar performances up to 100-node experiments (Protocol 3, same density). The throughputs match within a factor of two, with similarly non-linear increasing trends.
	Simulation beyond the limits of the real-device testbed, where the number of nodes was increased from 100 to 10,000, and the maximum number of connections per node was increased from 7 (coming from the wireless hardware limitation) to 50. Delivery of 2.5 MB of data to all of the 10,000 nodes completed in 190 sec, suggesting the potential for achieving this level of performance with real devices if such parameters were available.

home