Originally, the Internet was conceived by DARPA as a mesh network and designed to be one. One of the characteristics of such networks is that if you destroy a node, you only lose the node. The rest of the network continues to function quite happily. Similarly, if you lose a connection between two nodes, the network still functions quite happily. Technically a mesh network is defined to be a network where there are at least two pathways to every node – as illustrated below. In practice it means that the network has a lot of redundancy built in and it is therefore resilient.

A Mesh Network

Only recently has the IT industry and the telecoms industry started to take note of mesh networks. The design of Nicholas Negroponte’s One Laptop Per Child (OLPC) device incorporates a WiFi based mesh network. Thus if you have a number of such devices within a relatively small area it creates a usable and resilient network. The idea is that in, say, a village in a developing country there would only need to be one or two Internet connections and then every OLPC user would have access to the Internet and a usable free telephone service through Skype. All that was necessary to achieve this was to replace the standard WiFi (broadcast and receive) capability with a WiFi broadcast, receive and relay capability.

The point of this posting is to draw attention to the power of a mesh network. Mesh networks are disruptive in the business sense and can drive new business ideas. The mesh suddenly becomes very powerful if every node in the network has computing power. Then the mesh network is, potentially, a very powerful parallel computer or a very powerful distributed network of computers.

All peer-to-peer business ideas ride on this fact. If Skype had to provide its own computer power to provide its service, then it wouldn’t be a free service. It would cost too much. However, Skype works by distributing most of its computer activity amongst its population of users. The routers and cache points for directory information are the users’ computers. That’s why the Skype service is banned in most large organizations, it can generate unacceptable levels of network traffic within the corporate network.

Another use of mesh architectures is in the streaming and downloading of video. If one central server had to manage all the traffic, then as soon as the numbers of users grew large, the service would tank. But if you spread the workload across all the receiving PCs then the load can be nicely balanced and the need for power in the center diminishes.

Mesh networks could be viewed as a natural consequence of Metcalfe’s Law, but it is worth viewing them as a separate phenomenon, because it is both the power of the computers and the power of the network that deliver their impact. Maybe we could think of it as the marriage of Moore’s Law and Metcalfe’s Law.

Note: This is a key IT Trends posting that I will probably refer back to, to discuss various business events as they occur in the IT market. A list of such postings can be found here: