Afrimesh, Simple & Easy setup of wireless mesh networks for rural Africa
One of the fundamental issues with connectivity in Africa, is the lack of fiber or copper cables in rural areas where a large portion of the population live. The mobile operators have stepped up to the challenge with their mobile broadband offerings. However the costs of mobile broadband are still comparatively high for the rural areas. Additionally, the roll out of a GSM based mobile network is costly due to the infrastructure costs. Almost all hardware and software systems used by mobile operators are propriety and costly.
Could wireless adhoc mesh networking be the answer?
A wireless mesh network (WMN) is a communications network made up of radio nodes organized in a mesh topology. Wireless mesh networks often consist of mesh clients, mesh routers and gateways [1]. The mesh clients are often laptops, cell phones and other wireless devices while the mesh routers forward traffic to and from the gateways which connect to the Internet.
– From Wikipedia
Packets in a wifi mesh network travel in exactly the same way messages get passed on through communities using the oldest form of communication: Word of mouth. There is no master node or central node, and each node is responsible for routing as well.
In South africa, there are several wifi networks in operation such as the Wireless user group, who have chapters in Cape Town, Johannesburg, Pretoria. However, the advantage of wifi mesh networks are the low costs. A mobile phone, a laptop or even a cheap wifi router could be used as a node. This would be useful in rural areas or developing countries in driving down the costs of access to communications. Wireless mesh networks can also be ADHOC networks, which would be useful in cases of disaster relief when the normal communications network such as GSM is down.
So why don’t we use more wifi mesh networks?
Right now, the setup of a wireless mesh network is complicated and takes specialized technical skills, Unix/Linux experience and some hacking about with code. There have been steps taken by the Open Source community with the B.A.T.M.A.N project (No not the dark knight) but the Better Approach To Mobile Adhoc Networking which is a routing protocol for Mesh networks.
However, this still takes some specialized skills to setup and manage. You are also faced with the issue of increasing complexity as the network grows. This makes the network much more difficult to manage and configure.
The solution?
A new South African project called Afrimesh aims to solve this problem of complexity by making setting up a wireless mesh network as simple as buying a WIFI router and switching it on, with a few minutes of configuration.
Afrimesh will be a simplified dashboard for network management and configuration. The project was founded by Antoine Van Gelder with assistance from the Meraka institute at CSIR and is based on the work done with the B.A.T.M.A.N project. I had a sneak peek at the PRE-Alpha release (See screenshots below), and being in the very early stages of development it’s still very rough. However i can say the project shows a lot of potential, and if it makes managing and rolling out a WIFI mesh network as simple as switching on a WIFI router i see a great future for Afrimesh.
Popularity: 59% [?]
You write “You are also faced with the issue of increasing complexity as the network grows.”
This is true, but it isn’t about technology, its about math. Networks get complex faster than they get large. A 5-node network has 10 potential connections between any two nodes, while a 10-node network has 45 possible connections, and a 20 node network has 190 connections. This makes calculating useful data routes through even small meshes a certifiably hard problem.
I’d love to believe that this will work, but when you say that Afrimesh “solves the problem of complexity”, you are talking about user experience complexity, not actual mesh complexity. Their UI could be simple and clear, but it won’t make any difference if the meshing itself doesn’t work.
Have you seen a field test? That would be a big hurdle, and worth reporting.
@clay, thanks for the response. And yes i agree, the problem of routine is actually a mathematical problem. However if the dashboard can hide the complexity of the setup to the user this would be a huge increase in user experience. I also believe work done on routing protocols and autoconfig protocols should assist with this. The dashboard could most likely sit on top of these protocols shielding the user.
Regarding the field test i do not believe the system is ready for that as this release cannot even be regarded as Alpha. However maybe Antoine the founder can comment on that?
Dear Clay,
Great question!
Yes, because the number of potential routes in a network increases exponentially with the amount of nodes in that network there are hard limits on how large you can grow a mesh before your routers are spending all of their computational capacity crunching routing tables instead of forwarding packets!
To give an idea of the scale of existing networks we can look at the grand daddy of mesh field tests which is the Freifunk network (http://www.freifunk.net)
Freifunk have been very active in the development of the OLSR (http://www.olsr.org/) mesh routing daemon which forms the backbone of their network and to give an idea of how that approach scales: the Freifunk Berlin network had more than 600 observed live nodes during a 29 period in March/April 2007.
I am myself involved with a OLSR-based community mesh in Cape Town, South Africa which comfortably supports around 150 nodes.
A highly innovative alternative to OLSR has arisen recently which is the B.A.T.M.A.N. (http://www.open-mesh.net/) routing daemon. (Disclosure: Making it easier to build and sustain mesh networks using B.A.T.M.A.N. is the main focus of the Afrimesh project.)
Rather than attempting to crunch the routing table for the entire network on each router B.A.T.M.A.N. employs an elegant algorithm which, in effect, distributes the computational load across the entire network!
For an overview I can highly recommend this short article:
http://www.open-mesh.net/wiki/BATMANConcept
For more formal evaluations of mesh technology I can suggest the work of the Meraka Institute of the CSIR (http://www.meraka.org.za/wireless.htm)
Their work is particularly valuable and unusual in that they are not only engaging pure research questions but are simultaneously tackling live deployments in the real world.
A nice example of the work they are doing is this evaluation of mesh technology in a rural environment:
http://www.fmfi.org.za/wiki/images/3/3e/Peebles_mesh_ictd_india.pdf
(Disclosure: Meraka Institute are a primary motive force behind Afrimesh
)
I hope this answers your questions Clay, thank you for the interest you have shown!
– antoine
Antoine, thanks for the comment.