Application of Self-Organizing Network in IOT Devices

Labs Guide

In recent years, IoT devices have been widely used in homes and industries. Network access is one of the key technologies that determine the stability, reliability and cost of IoT device connections. When a device needs to be wirelessly connected to the Internet, the most instinctive idea is to connect directly to the gateway/router through Wi-Fi, cellular, etc. to achieve IoT device networking.

In addition to being simple and effective, Wi-Fi and cellular connection methods also have the advantage of large bandwidth, but the coverage of a single-point gateway/AP is limited. Especially in scenarios such as homes and industrial areas, IoT devices have signal coverage dead spots, such as in the pool, in the corner of the sofa, and even outside the house. In this case, either expand the coverage of the gateway/router or change to another networking method, such as using a router with Mesh function. That is, multiple routers are formed into a group of MESH to expand the wireless coverage, thereby partially solving the above problems. However, it is not completely a self-organizing network, and a group of MESH routers is not only expensive, but also still has coverage dead spots.

The problem of lack of signal coverage can be effectively solved by using self-organizing network (AdHoc) type network protocols, such as Zigbee and Bluetooth mesh, which are a type of self-organizing network. Self-organizing networks have the characteristics of spontaneously generating network topology and automatically healing the network after the node fails.

So what is an ad hoc network? The biggest difference from traditional networks is that the nodes of an ad hoc network are in an unstable topology state. Due to the movement of communication nodes or changes in channels, it is difficult for nodes in the network to have a stable access to the AP. Nodes in an ad hoc network can become forwarding nodes for other nodes to access the network, and there is no need for manual intervention during the topology change process, which is the origin of its name "ad hoc".

From the planar structure, the drone network with no more than 10 nodes, to the Zigbee network with a theoretical maximum of 65,536 nodes, they are all self-organizing networks.

Due to its advantages such as no need for basic network, strong anti-destruction, low power consumption and wide coverage, self-organizing networks are widely used in military, emergency relief, sensor networks and many other fields.

From Bluetooth and ZigBee under the IEEE802.11 standard to various private protocols in the military field, the core of the self-organizing network is still how the routing of distributed nodes maintains the routing relationship to cope with the changing topology. In short, when a node sends data, how does it know which node is the next one?

Based on the solution of channel access at the MAC layer, there are two main solutions at the network level:

1. Table-driven method: In this method, all nodes in the network broadcast the routing information of the node regularly or irregularly. Other nodes forward the information after receiving it and update their own routing tables. When updating the routing record, a self-incrementing id is generated, and the routing table is eliminated by comparing the id size. Through the above steps, the full routing table can be maintained in all nodes in the network. The advantage is that the forwarding delay is small; the disadvantage is that the cost of maintaining the routing table will be relatively large, and it will increase sharply with the increase in the number of nodes.

2. On-demand routing: In this method, the node does not maintain a real-time, full routing table. When there is a need for forwarding, a routing request will be initiated, forwarded by other nodes, and reach the destination node; the destination node returns a confirmation message, which is forwarded back to the source node, thus establishing a route. The advantages and disadvantages of this method are exactly the opposite of the previous one: low maintenance overhead but large delay.

So, compared with traditional point-to-point networks, what are the advantages of self-organizing networks in IoT application scenarios?

Self-organization does not require high network coverage.

It can effectively solve the network coverage problem in scenarios where inherent network facilities are insufficient, such as indoors and outdoors at home, and in the field.

The network environment is highly adaptable, and when obstruction or node failure occurs, the network will automatically adjust.

At the same time, it is also extremely convenient for users of IOT devices. The installation process of the equipment does not require deliberate adjustment of the network environment of the device node. When the network environment changes, manual intervention is generally not required.

Large amount of network connections.

Traditional Wi-Fi methods generally have a connection limit of 40 to 80, which cannot support industrial-grade IOT application scenarios. Self-organizing networks can greatly increase the number of nodes supported in the network through network segmentation. For example, the ZigBee protocol supports 65536 nodes. Modern smart buildings and factories often require tens of thousands of sensors, so self-organizing network access is the best choice.

The power consumption is lower than that of WiFi and cellular. Through timed wake-up and monitoring, self-organizing network devices can remain in sleep mode for a long time, reducing power consumption.

This feature determines that IoT devices can rely on batteries to work for a long time, greatly reducing the maintenance work of the equipment.

As mentioned above, self-organizing network protocols such as ZigBee, Z-Wave, Thread, and Bluetooth mesh have become indispensable connection methods for the IOT industry. They will also cover a wider range of scenarios and play a more important role in the future development of the Internet of Things.

Author: Guo Guanhua, Unit: China Mobile Smart Home Operation Center