8 thoughts on “The Mesh Potato Part 2”

  1. David,

    Great/exciting work! While several folks have tried to solve the “last mile” problem, there is no killer app to drive it, perhaps packaging VoIP will.

    I understand that your primary concern for the village telco project is to have something available and operational asap.

    But, what are your long-term “freedom” plans for the software and hardware?


    1. Will the hardware designs be available just as you did with IP04?

    2. Yes, Atcom is clearly benefiting from the open/free hardware community. But is their contribution only limited to making the finished product available? What about the “freedoms”? For example,

    3. Perhaps, efforts must be made to release software under AGPL (the “Affero” variety), the nature of software being a network service and all.

  2. Hi John,

    1. Yes

    2. Atcom are most definitely partners in this project – the benefits run very strongly in both directions. Atcom are one of the few companies who “get” open hardware and have been instrumental to the success of the IP04. I can’t see any specific mention of “freedoms” on the OpenMoko link you supplied? Do you mean availability of schematic and CAD files? Yes – the Potato schematic will be available as per the IP04.

    – David

  3. David,
    I wonder if anyone at Vilage telco is familiar with the “aloha” and slotted aloha protocols? Basically without pretty clever packet detection throughput is limited much more than you would expect. In the context of digital satelllite communications it has be quite throughly studied.

    The whole ad-hoc ahdemo discussion prompts me to ask the question.

    I suppose an operating network to some extent answers this question.

    warm regards,

  4. By a curious coincidence a Professor (Mike Miller) who I studied under worked on ALOHA for his PhD in the early 70’s. I have also done a lot of work in sat-com, and there are some parallels with Wifi. Small world……

  5. David

    I have read through the web sites and have to say wow, as a technical person this is wonderful however I do not think this project can ever be successful or in fact address the real needs of people in Africa.

    This project unfortunately does not appear to offer anything to the end user that does not already exist. I struggle to find any new feature that the project will provide that is not already on offer by commercial enterprises.

    The people living in a small township are not going to buy the hardware, they are not going to buy vouchers to call people who live relatively closely. Why would they, they can buy a prepaid cellular phone cheaply ( and I mean cheaply on the second hand market ) and for very little cost send an sms. I cannot see a person attempting to set up this network in his township as there is little they would offer that is new and given the very limited subscriber base highly unappealing to the end consumer.

    You wish to address the needs of the people in Africa then here are a few of my comments.


    African families are often dispersed over large distances as they often leave children with grandparents in rural areas and look for work in urban areas. Many domestic workers are unable to have the children living with them.

    First Requirement : Communication over longer distances


    The requirement is for communication of information, not immediate voice communication, that is demonstrated by the very high sms ( short message system of about 140 characters sent via the gsm mobile network ) utilisation of the three GSM networks in this country.

    Second Requirement : Information transmission – not Voice


    People in Africa cannot always answer a message immediately and often reply after hours or during work hours, many companies dislike lower paid employees having cell phones during work hours. The result is that many times an offline / delayed message via a voice recording system or sms is utilised. Delivery receipts are often used to verify a message has arrived at the destination and additional what time and day it was delivered , the mobile phone may have been off.

    Third Requirement : Storage of messages / delivery notification


    The project quotes hardware as being about 50 or 60 $ , with an exchange rate of over 10 rand to the dollar we are talking about over R500. For that price I can purchase a second hand cell phone , easily available, and a sim card and a 10 Rand voucher which I could use for SMS. Every now and then I buy a R10 voucher to remain connected to the network. And I can talk or send sms to any of the landline or gsm networks. Does this system beat that. NO
    Forth Requirement : Cost COST COST

    Given that these are the real issues here the technical / financial and legal restraints need to be addressed.

    First Issue : Distances

    There are in South Africa a few metropolitan areas, i.e. the major cities such as Cape Town, Johannesburg , Pretoria, Durban etc. Given that in these areas the population density varies considerably from affluent areas to power areas the distances need to be considered. A Mesh wifi network in my opinion could cover the vast majority of these urban areas. Which then brings up the issue of the backbone.

    Legal Requirements for long range communication come into play here and need addressing before any development continues. Wifi could comfortably transmit short messages via a mesh network, a similar concept exists in packet radio and has worked for twenty or more years. Voice over IP over a wifi mesh I fear will deteriorate rapidly and congest a network rapidly.

    Second Issue : Hardware and Cost and access Cost

    The hardware needs to receive data, not necessarily have the ability to process audio. So the device needs a keyboard and a screen. The TV is a very common appliance throughout the socio groups and a cheap PS2 keyboard will do. The device would have a cable to the TV , a port for a keyboard and a wifi transmitter. Such a device would be very cheap to build and the unique wifi transmitter id could be utilised as a unique ID for the device.

    So in theory a network in urban areas of Wifi transmitters relaying messages could work. Some software could be developed that does this task and since many middle class suburbs have wifi devices they could act as relay stations. A Mesh could then be built in urban areas very quickly as these pc based users would not require any internet connectivity and would only be donating wifi bandwidth. So your cost reduces dramatically to create a mesh network.

    So the device needs to be on to receive messages and store them in case a battery runs flat, if a message sent cannot be delivered it would return after a while to the sender saying the message has not been delivered or the time it was delivered. A led would light up when a message is received and the device could then be plugged into the tv to read and reply, hence not requiring a TV to be on all the time.

    This hardware would not be difficult nor expensive to develop as most of the devices are external. The cost to remain on the network would be free as it is connecting to a Mesh network, and the cost of sending messages would effectively be free.

    Third Issue : Legal

    This would need to be addressed with the regulatory authority but in this simple format it would not intrude on laws I am aware of as it is mostly open spectrum based. Transmission over long distances needs to be considered tho.

    So lets now address the requirements.

    First Requirement : Communication over longer distances – well for the urban areas it solves the problem as the device uses a wifi MESH, intercity communications would be more problematic but solutions can be found. This WIFI Mesh could be built a lot quicker than the projects current proposal. In areas where the wifi network does not reach then wifi relay station utilising solar power could be erected. Who would pay for these ? Why not get the blessing of the Lions, Rotary or any number of charity type organisations. Get them involved in the local community. They could also subsidise the hardware to get the system going.

    Second Requirement : Information transmission – well the system would work very well for transferring information and would effectively be running on a free network.

    Third Requirement : Storage of messages – the system would store limited messages and offer delivery / non delivery of messages, does the current system offer that ? Nope

    Forth Requirement : Cost COST COST – well apart from the initial cost of the device and a transformer / batteries the device is not incurring usage fees .

    To conclude the proposal above might not be perfect and would allow a network to be easily constructed very quickly and very cheaply. The experience getting it going would enable future development of protocols allowing for future development of VOIP. If you read the articles about the African experience of wireless communication, refer http://www.economist.com etc., then the greatest need is information transfer – not voice ….

    Additional services could be developed for the system, and being text based would not make the hardware expensive. Advertising over a local system or many other commercial uses could be developed utilising the network, isn’t that how the current internet works?

    I like the project, but I fear it is not going to work, it is a great technology, but not one that is suited for the poorer people in Africa. I love technology but even in billion Pound / Dollar / Rand projects I have seen that the development team develop a wonderful technology only to discover that the end user actually doesn’t like what you have delivered and doesn’t use it .

Comments are closed.