Each node in a complete mesh is intimately linked to every other node. This allows a message to be transmitted over many routes. A partial mesh is one in which not all nodes are directly linked to one another. Only those connected by a single link are considered neighbors. Partially connected meshes are used when you want to limit the number of hops that a message can travel.
In computer networks, a fully meshed network is said to have "full mesh connectivity". In other words, if one node in a fully meshed network fails, then any other node can still communicate with every other node by passing the message along each of its adjacent links.
Fully meshed communication networks were first implemented for military applications where they are necessary for efficient group communications. Today, they are used in wireless networking and some local area networks (LANs). However, fully meshed networks are expensive to install and maintain because each node needs multiple connections to other nodes. Also, traffic going through partially connected networks may need to be rerouted around damaged links or nodes.
Partial mesh connectivity provides some degree of redundancy in case some links or nodes fail. With this type of network, messages can still be passed from node to node even if some links or nodes are down. Partial mesh networks are commonly used in wireless networking and some LANs.
When every node in a network has a circuit linking it to every other node, the network has a complete mesh topology. Full mesh is the most expensive to install but provides the most redundancy, so if one of those nodes fails, network traffic may be routed to any of the other nodes. In general, a fully meshed network has more total links than necessary, so some of the links are redundant--that is, they connect two nodes that can communicate directly with each other. However, a fully meshed network requires that all nodes have at least one link to another node in the network, so it cannot be partitioned into subnets.
Full mesh networks are difficult to install and maintain because all nodes need to know about all other nodes in order to route packets. Also, there is no limit to the number of hops between any two nodes; therefore, full mesh networks are inefficient when used for small communities or local areas. Finally, full mesh networks require a large amount of hardware in order to function properly. For example, a typical home router does not support full mesh networking.
In summary, a fully meshed network is very expensive to install and maintain and is unsuitable for most applications.
Partial mesh networks consist of several subnetworks interconnected by partially meshed regions. These regions allow traffic to flow from some nodes in one subnetwork to some nodes in another subnetwork, while blocking traffic from others.
Having nodes placed in a mesh topology has certain advantages, including the ability to receive messages more rapidly if the route to the intended receiver is short. Messages should always get through since they can take many different paths. Multiple connections imply that no node should be isolated (in principle). A disadvantage is that attacks on one node may cause other nodes to stop functioning.
In a ring topology, every node receives all messages and forwards them onto adjacent nodes until it reaches the closest node which does not know how to forward the message. Thus, in a ring network, each node gets to see every message transmitted by any other node in the network. This is why ring networks are often used when every node needs to access the data stored at every other node because this type of network architecture provides direct connectivity between any two nodes.
Mesh networks use multiple routes from source to destination to avoid nodes that are down. Therefore, messages will usually reach their destinations even if some nodes are taken out of service. Mesh networks are useful for supporting remote offices or employees who must stay within the network but need to communicate with others outside of it. These individuals would only have access to their own local router through its local IP address rather than the global address reserved for routers on the Internet.
A drawback of mesh networks is that they are more complex to set up than ring networks and require more maintenance.
There is no central connecting point in a mesh structure. Instead, each node is linked to at least one other node, and often to several. Each node has the ability to send and receive messages from other nodes. The nodes serve as relays, forwarding messages to their eventual destination. A path may pass through multiple nodes, but only one message will be delivered to it's destination at any given time.
In a mesh network, every node is connected to at least one other node. However, not all connections between nodes are equal. Some connections may have better quality than others. For example, if node A connects to node B, and node B connects to node C, then there are three connections in total. If node A connects to node C instead, then there are two connections in total. If both nodes A and B connect to node C, then there are four connections in total.
In general, a mesh network has N*(N-1)/2 connections, where N is the number of nodes.
For example, in the diagram below, there are five nodes in total. Therefore, the maximum possible number of connections in this network is 50. In fact, this network has exactly fifty connections.
Each connection has two ways around the network, so fifty is the maximum possible number of paths through the network.
A complete mesh topology connects every node on the network to every other node on the network. This delivers the most dependable network since it is totally redundant. Changes in one part of the network will be reflected by other parts of the network almost immediately because there are no disconnections or lapses in transmission.
A bus topology connects nodes in a ring. The advantage of this topology is that it's easy to add more nodes; just add more buses! The disadvantage is that if one node fails, then all communications through it will be lost. A star topology is like a bus topology but with one node acting as a central hub. If you need to connect several clusters together, a star topology would do the job. The cluster nodes would share the hub node's connection to another cluster or destination device. The choice between a bus, star, and complete mesh topologies depends on how many nodes you expect to be connecting together.
In conclusion, a completely mesh topology is the most stable because changes in one part of the network are reflected by other parts almost immediately. A bus topology is next because it's easy to add more nodes by just adding more buses. Finally, a star topology is least stable because if one node fails, then all communications through it are lost.