Diagram of the routing protocol for the wireless mesh network technology (1)

With the development of computer and wireless communication technology in recent years, mobile wireless computer technology has become more and more popular and widely used. As it is no longer subject to cable laying restrictions, users with mobile computer devices can easily and freely move and communicate with others without fixed network facilities. In this case, they can form a mobile Adhoc Network or a mobile wireless mesh network. The Mobile Wireless Mesh Network is an autonomous system composed of wireless mobile routers (and connected hosts. The system can be moved randomly to automatically adapt to network topology updates without any backbone network or network infrastructure.

In addition to mobile wireless mesh networks, more and more commercial applications of fixed wireless mesh networks have emerged recently. A typical example is "community wireless network ". It is used to provide access for communities without broadband access. In these fixed "community wireless networks", each wireless router not only provides Internet access for its users, it is also part of the network infrastructure-routing data to its destination in a wireless mesh network.

A layer-3 wireless mesh network features high flexibility and inherent fault tolerance. This network simplifies line-of-sight transmission and expands the network scale and coverage with the minimal network infrastructure and interconnection costs. In real life, there are also hybrid Wireless Mesh Networks: Some mesh routers in the network are mobile, while other mesh routers are fixed.

In either case (mobile or fixed or hybrid), the wireless mesh network has some notable features, such as: high dynamic, intelligent, end-to-end Optimal Path Selection, multi-hop, generally, the bandwidth is limited and the computing capability is insufficient. There are two reasons for the high dynamic nature of the Wireless Mesh Network: first, the vro itself may move (such as in a mobile or hybrid Wireless Mesh Network) and cause rapid changes in the network topology. Second, even if the router itself does not move (for example, in a fixed wireless Mesh Network), the quality of the radio link may change rapidly due to interference, geography, environment, and other factors.

From the above features, we can know that a complete wireless mesh network protocol must have the following features:

Distributed operations

Fast Convergence (to ensure faster movement)

Scalability

Suitable for a large number of small devices

Use only limited bandwidth and computing power

Proactive operations (Reducing Initial latency)

When selecting a route, consider the quality and capacity of the radio link.

Loop avoidance

Security

Note 1: the concept of community wireless network is very popular in developed countries such as the United States and is still in the development stage in China.

In addition to traditional routing protocols (such as OSPF and RIP) designed for wired networks, there are also a large number of routing protocols designed for mobile adhoc networks. These routing protocols are generally divided into two categories:

Reactive routing protocols (such as XSL, DSR, and TORA ). This type of protocol only discovers and maintains the route when necessary. To meet traffic needs, they can use power supplies and bandwidth resources more effectively, at the cost of increasing the latency of route discovery.

Active routing protocols (such as DSDV and OSLR ). This type of protocol always maintains the route to each possible destination-the Protocol assumes that these routes may be used. In some cases, the extra delay caused by the reactive routing protocol may be unacceptable. In these cases, if bandwidth and power resources permit, the active routing protocol is more popular.

Traditional routing protocols (such as OSPF and RIP) are designed for wired networks. They are not able to handle the rapid changes in the topology and link quality that are common in the wireless mesh environment. They can be divided into two categories, according to their design philosophy: (1) distancevector; (2) linkstate ). Distance Vector routing protocols (such as RIP) are used in early English networks, such as ARPANET. Its main advantage is the simple and efficient distance vector calculation method. However, this method is prone to slow convergence and routing loops.

The connection status routing protocol (such as OSPF) features that all routers store the full network topology information and perform periodic update (link state periodic update ). In addition, changes in any stage lead to instant updates. Compared with the traditional distance vector routing protocol, the connection status routing protocol has full topology information, so the routing loop is easier and the convergence speed is faster. Unfortunately, the connection status Routing Protocol relies on the network-wide broadcast (flooding) to transmit the latest information. Therefore, especially when high mobility (or severe radio interference causes connection status fluctuations) occurs, to keep up with the rapidly changing topology, such protocols consume a large amount of network resources and vro processing capabilities and generate excessive control overhead, making them unfeasible.

Many existing adhoc routing protocols have made significant improvements in dealing with fast topology changes. For example, the put-in-line routing protocol is a reactive distance vector routing protocol. The basic idea is to discover and compute routes only when necessary, and discover the route of the destination using the query/response packet. However, the route must be found before the actual data packets are transmitted.

Using the reactive routing protocol increases the initial latency (InitialDelay ). In addition, when the mobility and load are high, this reactive routing protocol can suffer a huge protocol load. Simulation results show that these protocols have intolerable packet loss and latency under high mobility and high load. In addition, the vast majority of existing protocols, whether traditional or similar to Ad hoc, have serious problems in terms of scalability and stability when adapting to the rapid changes in the quality of common radio links.