Mesh networks (meshed networks) are ad-hoc networks in which routing points are immobile. Customers are connected by a wireless network access points and the access points are connected by wireless links.
The advantage of these networks is that they can cover a large geographical area, without the need for laying cables. For example, on a large campus, the access points can be set on the roofs of various buildings without the architect of the network has to worry about connecting access points to a wired Ethernet system.
Several opportunities are emerging to make a mesh network:
• Use the same frequency as the terminals, considering that the access points are treated as terminal machines. The downside of course is to use bandwidth taken from the other terminal machines. Also, be careful that two access points are not too far apart and not require the transmitter and receiver to lower their speed. This solution is considered first generation mesh networks.
• Use different frequencies. For example, an 802.11b Wi-Fi network with three frequencies available, it is possible to use two communication cards with different frequencies. The downside is of course disturbing the frequency plan, especially if the network is large and has many access points. This solution is part of the second generation of mesh networks.
• Still in the second generation, the mesh network uses a different standard to connect the access points between them. For example, an 802.11g mesh network can use the 802.11a IEEE standard for interconnecting the access points.
• It is considered that the third generation using three frequencies in total. One to connect customers and to interconnect the two access points. In this case, the upstream and downstream connections of the same node using different frequencies. It usually uses 802.11a, which has up to eight different frequencies.
Mesh networks pose new problems for wireless networks, including: how to optimize the battery access points if they are not connected to the electric current? how to optimize routing to avoid disrupting user traffic from access points, especially if they are already saturated? what density of access points should be used, which is to ask the question of the power of access points?
The advantage of this technology is to be able to easily reconfigure when an access point fails. Guests can connect to another access point, even slightly increase the power of neighboring access points that failed.
The main problem is to manage the routing. The latter is treated in the access points that are not very powerful machines, and we must therefore avoid access points that support a lot of traffic from connected clients. Many proposals are being discussed, primarily those from ad-hoc networks.
The IEEE working group also 802.11sa this. Following fifteen proposals, the group selected two proposals the SEE-Mesh and Wi-Mesh, who have come together to form a single proposal. This proposal has become a standard in April 2007, after many discussions implementation. The access points and stations that have the 802.11s routing algorithm are appointed Mesh Points (MP). Radio links are used to interconnect them. The default protocol is HWMP (Hybrid Wireless Mesh Protocol). This hybrid protocol is a combination of a protocol from AODV, the RM-AODV (Radio-Metric AODV) and an algorithm based on the trees. A second protocol may be used when MP accepts: RA-OLSR (Radio-Aware OLSR).
The IEEE 802.11s group also defines security solutions for mesh networks. To do this, define mutual authentication MP, generate and control the session keys, enabling data encryption on the lines of the ad hoc network and detect attacks. To do this, perform authentications using the IEEE 802.1x protocol. The session keys are managed by a PKI (Public Key Infrastructure), dedicated to safety. Confidentiality is ensured by the IEEE 802.11i standard.