Analysis of Wi-SUN FAN Network Formation Time

Abstract

The Wi-SUN FAN (Wireless Smart Ubiquitous Network Field Area Network) standard is attracting great interest in various applications such as smart meters, smart cities and Internet of Things (IoT) devices due to the attractive features that the standard offers, such as multihop and mesh topologies, a relatively high data rate, frequency hopping, and interoperability between manufacturers. However, the process of connecting nodes in Wi-SUN FAN networks, which includes discovering, joining, and forming the network, has been shown to be slow, especially in multihop environments, which has motivated research and experimentation to analyze this process. In the existing literature, to measure network formation time, some authors have performed experiments with up to 100 devices, which is a costly and time-consuming methodology. Others have used simulation tools that are difficult to replicate, because little information is available about the methodology used or because they are proprietary. Despite these efforts, there is still a lack of information to adequately assess the formation time of Wi-SUN FAN networks, since the experimental tests reported in the literature are expensive and time-consuming. Therefore, alternatives such as computer simulation have been explored to speed up performance analysis in different scenarios. With this perspective, this paper is focused on the implementation of the Wi-SUN FAN network formation process using the Contiki-NG open source operating system and the Cooja simultor, where a functionality was added that makes it possible to efficiently analyze the network performance, thereby facilitating the implementation of new techniques to reduce network training time. The simulation tool was integrated into Contiki-NG and has been used to estimate the network formation times in various indoor environments. The correspondence between the experimental and numerical results obtained shows that our proposal is efficient to study the formation process of this type of networks.

Keywords: Contiki-NG; Wi-SUN FAN; node connection process; simulator.

Figure 1

Wi-SUN FAN profile.

Figure 2

Unicast frequency hopping.

Figure 3

Broadcast frequency hopping.

Figure 4

Discovery and joining process.

Figure 5

Process of sending or receiving PA or PC packets that allow for the synchronization of new nodes.

Figure 6

Process of sending and receiving PAS or PCS packets that allows nodes to throttle the rate at which their neighbors transmit PA or Ps packets.

Figure 7

Contiki-NG stack with new functionality incorporated.

Figure 8

Network topologies. (a) Linear. (b) Fully connected. (c) Mesh scheme.

Figure 9

Experimental scenario. (a) Testbed. (b) Connection scheme.

Figure 10

Topologies configured in Cooja: (a) Linear. (b) Fully connected. (c) Mesh.

Figure 11

Network with 20 nodes.

Figure 12

Network with 101 nodes.

Figure 13

Network formation time considering a linear topology.

Figure 14

Network connection time considering a fully connected topology.

Figure 15

Network connection time considering a mesh topology.

Figure 16

Network connection time considering a 20-node mesh topology.

Figure 17

Network connection time using configurations for small-, medium-, and large-scale networks.