I was reading a copy of Linux Journal sent recently to me in a care package (good luck finding one in Mombasa), and one of the last articles (online link here) was about a prototype device called the Mesh Potato. I liked the concept of the Mesh Potato so much, that I thought I would share it with all of you, along with some thoughts and impressions, because I am a blogger and that is what we do.
In short, my understanding of the Mesh Potato is that it is a piece of hardware custom built by a group of people funded by the Shuttleworth Foundation to help create low cost, locally owned, telecommunications infrastructure. The group is called Village Telco actually, and it is comprised of a bunch of hardware and software hackers who are committed to providing an inexpensive, open source (both software and hardware schematics) solution to the growing telcom industry issues faced in the developing world.
I won’t go too far into the technical details of the Mesh Potato itself, but the concept, as the name suggests is that of mesh networking. First you must understand the traditional model. Basically, in a traditional networking sense you have clients (like your phone or computer) that need to somehow talk to come form of central command in order to route information from one place to another. This creates what are known as single points of failure in the system design, and oftentimes though clients may be inexpensive (such as phone handsets, which cost as little as $20 for a new one in Kenya, without even buying a contract), the central components, such as servers or cell phone towers are usually very expensive to purchase and maintain. The larger parts of the system are what make infrastructure prohibitively expensive for locals to invest in.
The notion of mesh networking (as applied here) is different. Instead of purchasing large pieces of infrastructure to blanket an area that may be larger than you need with your networking coverage (be it Internet, or phone), mesh networking devices are designed to operate within smaller zones of their own and talk to other mesh devices they detect, then cooperate with those other devices to perform the functions of the more traditional central infrastructure. The idea is that you can build your network piecemeal and eliminate single points of failure (SPFs). For example, instead of every cell phone talking to the same central tower which routes all calls, a mesh client automatically talks to the closest mesh device, which then handles all the routing to another mesh devices until it gets to the other end client using their phone. The benefit: the cost per individual mesh component may be higher than the least expensive component of the traditional model, but there is no single prohibitively expensive cost. In English, the cost of three mesh devices will be far less than the cost of a phone tower and associated supplementary devices and will cover the area you need, not the area you need plus a whole lot of land that is unnecessary (but increasing the cost of implementation).