Commentary: the joy of meshBMJ 2008; 337 doi: https://doi.org/10.1136/bmj.a2500 (Published 03 December 2008) Cite this as: BMJ 2008;337:a2500
- 1KTH, Forum 100, SE-164 40 Kista, Sweden
- 2Swedish Institute of Computer Science, Box 1263, SE-164 29 Kista, Sweden
Wherever there are several XO laptops, a so called mesh can be formed. A mesh network is a highly resilient form of network in which data, voice, and instructions can transfer between network nodes through a series of hops in a wireless peer-to-peer fashion. In its simplest form, a mesh network is ad hoc: your computer can join or invite other computer owners to the network, provided that they are within range. It is enough for one computer in the mesh to have an internet connection for everybody to get access to the internet, and many other resources can also be shared, such as a printer connected to a USB port on one of the computers.1 The data rate across this network will not be high but should still easily serve a few dozen computers with good results, as ongoing tests indicate.
The XO has two ways of connecting to the internet, in addition to the simple mesh access. If you are close to a wi-fi hotspot you may connect without a mesh, and if you are close to a school server mesh network you may use that server’s internet hook-up. The school server can also take care of back-ups and, for example, build a database of homework assignments. Part of One Laptop per Child’s vision is children forming simple mesh networks in their homes and connecting to the school server when at school (or indeed, at the clinic). This is not as easy as you might think to configure, since mobile ad hoc networking means that no network manager can know exactly which computer to expect at a given time. Even if no distribution of IP addresses is needed, a fixed computer network is a lot easier to maintain and secure, and even portions of the internet itself constitute mesh networks. For the XO user, the graphical depiction of the “neighbourhood view” makes it simple to network.
Such interfaces were first developed for military applications of wireless mesh networks. So called blue force tracking—knowing where your vehicles are—is enabled by graphical views of node tracking. For the clinic, patients could interact with each other just like at school, but more importantly with the medical staff. One or more clinic servers could be set up to provide for patient interaction, such as letting waiting patients fill in forms and submit them to the server. Reasonable security could also be provided, chiefly through passwords, to let patients access parts of their health records. A clinic server can be hoisted up a flag pole each morning and be partly powered by solar cells.
In the developing world, wireless access is sometimes faster and more reliable than fixed connections, and forming a simple mesh is free. The One Laptop per Child project was not the starting point for mesh networking in the developing world. In South Africa, for instance, the Wireless Africa movement has been experimenting with mesh since 2005.2 Many developers are currently working on efficient and reliable protocols for such networks (based on the IEEE 802.11s standard and targeting the developing world’s rural areas in particular), a bottom-up development made possible by the open source Linux platform of the XO computer. Indeed, long before the start of the One Laptop per Child programme the early German Freifunk (free radio) mesh had a similar slogan (one laptop per (street) corner) almost five years ago.3
One of One Laptop per Child’s objectives is to extend the range of the XO neighbourhood view to a kilometre or more. In addition, movements like Freifunk are now contributing much downloadable content and web based applications for the XO. Solar power and other alternatives to the crank handle are also being developed. In the next few years more opportunities for the clinic will no doubt arise—for example, it may be possible to view film and graphics via a clinic server within a year. Video conferencing and other high bit rate applications, however, will remain a long term goal.
Cite this as: BMJ 2008;337:a2500
Competing interests: None declared.