Mesh networks
Mesh networks consist of wireless nodes supplied to consumers. The wireless nodes act both as a client and as a repeater. Mesh networks is not to be confused with Mesh Networks the company
www.meshnetworks.com which supplies wireless routing devices, or software. The original idea of mesh networks is a network of nodes that can communicate with each other and provide multiple routing links to each other, thus providing redundancy.
www.whatis.com has a search facility which defines the term.
In the wireless field mesh networking is more attractive in that the limitation from the expense from wiring from node to node in a variety of ways has been eliminated. There is the ability for wireless nodes to be placed in a seemingly random fashion. Connections ?make themselves? or if they are out of range they do not get formed.
The system is a multi hop system. The devices hand packets from one device to another and thus help the transmission of packets much the same as a relay racer hands a baton on to another racer. If there is another racer at the handover position that also has a route to the destination then the packets may be handed to different runners but still reach the destination.
In adverse conditions, such as loss of nodes then the packets may take alternative routes as exist to reach the final destination. In theory if the roads become congested then multiple roads may be taken by the numerous packets heading for the destination. However a packet can only take one road at a time but a data stream is broken into many packets.
In theory as these nodes are wireless and self connecting then minimal preparation is required for their deployment. They can be dropped in and will assist in the network design. The more populated an area will logically be where more nodes are placed.
It is erroneous to state that thousands of nodes may be deployed as the scalability of a mesh network has yet to be proved and may even be impossible with out some deal of modification such as the likes of Turbowave and Mesh Networks, to mention only 2.
Reliability is also a questionable as if nodes disappear and reappear in a random fashion, then what was a good quality link one day is no longer. Wireless compared to wired networks have differing speeds of links and differing bandwidth (related).
Interference has been attributed to some poor quality links. Multiple radios on the same channel can occur and more so in adverse circumstances which is just when the network is requiring more capacity.
Application
Applications of a mesh network can be for the adhoc creation of a connected community such as the Nzwireless.org. Community groups have associated so that they are able to link and share communication without the need for an ISP.
Last 10 meter application are also abound. ISPs are able to link to areas with cable and fibre. However there is the problem with distribution which can be costly. This involves wiring from the street to the user. The wiring consists at the best case of wiring from street to house to wall to computer. Worse still is wiring from the end of the street or from the edge of the suburb.
It has been suggested that adhoc applications of the mesh network may work from hardwired accesspoints. This is a basic understanding of the Mesh Networks approach. Hardwired Access points would be located near to users who may have adhoc mesh network devices to connect to the access points.
Farmers may have a use for this type of application with longer distances and higher gain antennae linking them to neighbouring farms.
Pros
Cheap and dirty wireless connections. Costs to end users may be lower as there are less infrastructure costs involved if using the ISP model of implementation.
End users are able to create their own network without ISPs and this is already happening in NZ with Nzwireless.
No infrastructure costs when end users create their own networks.
Minimal redundancy which is better than none, as there may be more than one route to the destination. This depends on density of nodes.
Cons
Bottom up creation of a network creates multiple non compatible systems.
Lack of reliability as one user may depend on a second but closer user to connect to the rest of the network. If the uncontrolled second user powers down then conection for the first is lost.
Sharing of bandwidth. The second and closer user to a destination bandwidth will decrease as the more distant users try to download. In effect the user that can connect fine to the destination is hindered by those who cannot. Those users who cannot connect through to the destination with out the second user are sharing the second user?s bandwidth.
Congestion. As the density increases so does the bandwidth usage however 802.11b has only 11 channels or so, which means there is a limit on the numbers of simultaneous radio signals without interference. Worse still if you consider there are only 3 channels that do not overlap in frequency.
The placement of nodes is usually uncontrolled and therefore coverage is patch and interference is rife.
As the numbers of sites increases the likely hood of packet storms and routing loops increases thus limiting the size or number of nodes. Without some layer 3 routing protocol the ball park figure for numbers of nodes would be less than 30 arranged in a mesh network.
Other technologies Compared and contrasted
See RoamAD for infrastructure based on modified mesh network principles.
See turbowave, mesh networks, Wave Wireless,
Community free based mesh network software:
http://www.sputnik.com/products/gatewaydownload.html
Examples how it would benefit NZ
It probably would only benefit the community based mesh networks, which were free to all. The software for this is free. It is happening already. An ISP based system would seek to charge for this service and create barriers to entry and defeat the whole purpose of a low cost network. Reliability would not be carrier class and so end users would not seek to pay for an unreliable service.
Benefits include a mini Internet for communication freely between individuals. Freedom of speech would be facilitated. Bandwidth limitations would be removed and force ISPs to follow suit. Competition would evolve from the community against the commercial infrastructure. Devaluation of the wired networks would possibly reduce consumer costs and therefore increase availability to the public.
Some would consider this the ramblings of a fool, but it is out there, and only an opinion.
research help 
ops: 
Linear Mode
