First, this chapter describes the limitations and the basic behaviors of IEEE 802.11. IEEE 802.11 uses the carrier sense multiple access with collision avoidance mechanism (CSMA/CA) to detect whether the medium is available to avoid collisions and provides distributed and centralized coordination functions to achieve medium access control. Second, the architecture of the IEEE 802.11s mesh networks is introduced. Because of the diversity of architecture between IEEE 802.11 and IEEE 802.11s, using multiple channels is necessary. Finally, the problems caused by channel switching in the IEEE 802.11s mesh networks and the issues for solving the problems are described.
2.1 Background
2.1.1 IEEE 802.11
IEEE 802.11 uses the CSMA/CA mechanism to access medium. Each node checks whether the medium is available by carrier sensing before transmitting. Virtual carrier sensing, one of the carrier sensing functions in IEEE 802.11, is provided by the network allocation vector (NAV) which indicates the amount of time that the wireless medium will be reserved. Each node computes the expected amount of time to complete its operation sequence and sets this value to the NAV and then the other nodes count down from the NAV to zero. The NAV being nonzero implies that the virtual carrier sensing function deems that the medium is occupied. On the contrary, the virtual carrier sensing function indicates that the medium is available when the NAV is zero. It protects the operation sequences from interruption by the NAV.
The IEEE 802.11 MAC layer provides two medium access coordination functions: the distributed coordination function (DCF) [1] and the point coordination
function (PCF) [1]. The DCF is a contention-based mechanism. Each node checks whether the medium is available before attempting to transmit. When the medium is available, each node accesses the medium after a random back-off time generated from its contention windows. On the other hand, The PCF is a contention-free mechanism. A point coordinator which resides in an AP uses a centralized access control method. When the PCF is working, time is divided into the contention free period (CFP) and the contention period (CP). In the CP, the DCF works and each node contends for medium access. In the CFP, each node transmits frames only when it is polled by the point coordinator. In IEEE 802.11e, the PCF has been extended to the HCCA. The basic behavior of the HCCA is similar to the PCF but the HCCA includes more mechanisms for the quality of services.
2.1.2 IEEE 802.11s mesh networks
IEEE 802.11s defines the mesh networking using the IEEE 802.11 MAC/PHY layers that support layer-2 path selection protocols and data forwarding over multi-hop topologies. Moreover, IEEE 802.11s also defines the multi-channel mesh networking to separate traffic into different channels to improve network throughput.
Figure 1 illustrates the architecture of the mesh networks. Each node which joins a mesh network is called a mesh point (MP). A MP which also plays the role of an AP is called a mesh access point (MAP). A MP which bridges wired networks is called a mesh point portal (MPP). Mostly, an user is a MP or a STA. For the MP case, an user transmits data through its neighbor MPs which forward these data to the destination.
For the STA case, the mesh networks play the role of a wireless distribution system which is extended from wired networks. An user transmits data through the MAP and then the MAP forwards these data to the mesh networks. If BSS traffic and mesh forwarding traffic use the same channel, they starve each other because the channel
can only be occupied by one side. As a result, they are usually separated into different channels.
Figure 1 Architecture of the IEEE 802.11s mesh networks
2.2 How to switch N radios among M channels? ( M > N )
In a multi-channel scenario, the radios must switch among channels if there are less radios. However, some problems caused by channel switching are given as examples shown in Figure 2.Figure 2 Examples of the problem at channel switching
Switching among channels causes packet loss in the unattended channel because no radio stays in. This is one kind of the deafness problems [13] [14] [15]. Furthermore, when the switch node returns to a channel and intends to transmit data, a collision might happen due to the channel is occupied by the other transmission and the switch node does not update its NAV. This is one kind of the multi-channel hidden terminal problems [16]. Furthermore, packet loss results in link disassociation and more overhead in the infrastructure BSS and the mesh networks. In the mesh networks, dropping mesh forwarding traffic renders all consumed resource as being wasted, especially for a long forwarding path. Moreover, link disassociation might cause that the network topology is destroyed and rebuilt. In a infrastructure BSS, a STA might be isolated or associate with another AP because of link disassociation. When the switch node returns, STAs might re-associate with the switch node due to a better link
quality. The radio switches frequently might exhaust the resource of the networks.
Therefore, to minimize the above overhead, it is necessary to notify STAs and neighbor MPs before leaving the channel to avoid packet loss.
Using shorter channel switching interval is another intuitive idea to reduce the damage of the deafness problem. However, high switching frequency causes that channel switching overhead becomes quite heavier and the multi-channel hidden terminal problem becomes more serious. On the other hand, using longer channel switching interval could decrease switching frequency. Nevertheless, allocating a long period for a low traffic load channel is inefficient in channel utilization. A suitable switching interval depends on channel switching overhead and traffic load. Therefore, in the multi-channel environments, it is an important issue that how to switch radios among channels sufficiently in the IEEE 802.11s mesh networks.