A Village Telco in South Africa
Steve Song
For the last few months Alberto Escudero-Pascual
(http://www.it46.se/), David Rowe (http://www.rowetel.com/) and I have
been informally discussing what I am calling a "Village Telco", an
out-of-the-box telephone company (with data too) built from low-cost
commodity components. This would include easy-to-use/setup/manage
technology, network design, links to suppliers, etc but most
importantly, it would provide the business model for establishing
local voice (and data) services. I write more about the idea here
http://manypossibilities.net/2008/02/village-telco/
I was very happily surprised, shortly after moving to South Africa,
when Mike Jensen introduced me to Rael Lissoos who runs a company
called Dabba (dabba.co.za). Rael already has what I consider a
working model of a Village Telco in Orange Farm south of Johannesburg.
See http://manypossibilities.net/2008/03/dabba/ for detail and pics.
When I chatted with Rael, he was right in tune with the notion of
packaging and commoditising the Village Telco model in a shared Open
Source/Hardware space. A project like this could be useful from
Patagonia to Papua New Guinea.
I love to get some reactions to this. It seems to me that there are
the seeds of an open, shared initiative here to commoditise (and
consequently drive down the cost of setup) of a "village telco",
"microtelco", call it what you like, although you will note that I put
voice first and foremost. I know there is work going on elsewhere on
WISPs-in-a-box. This idea is not so different (and very
complementary) but focuses on a) voice first and b) the business
model.
Any and all comments welcome!
-Steve
---
Steve Song
http://manypossibilities.net
Kobus Roux
I like the term Village Telco. It describes it pretty well. For me the
ultimate challenge is to see villages earning an income from
terminating calls, as is the custom between proper telcos. Everyone is
advocating the benefits of having ubiquitous access via mobile or
other forms of telecommunications, but the negative result is that the
poor, mostly rural communities, just get poorer as the national
networks charge their fees and see the money flow back into their
urban headquarters.
Another challenge is that many African countries still view carrying
of voice as the exclusive territory of the national operators, and
will be quick to stop any Village Telco by restriction or by
exorbitant license fees.
Having the enabling technology is one necessary condition, but we need
inputs and ideas on how the business model should look like to scale
up and replicate beyond those few exceptional villages. Then create
the enabling environment (regulatory, legal, etc) to make this happen.
Perhaps this is why Nokia Siemens Networks have decided to link their
Village Connection to national mobile networks - to scale and avoid
license and other restrictions?
We started working on the WISP-in-a-box as the piece of technology
that would make it easy for any entrepreneur to set up a Village
Telco, for exactly the reasons you highlighted. I would be happy for
us to join other similar initiatives, or incorporate work that others
have done. (Assumption: Everything is open source).
If there are others who are interested, please comment on this thread,
or drop me a note. We are also looking for people willing to try out
and test the platform as it evolves.
Kobus Roux
On Mar 7, 10:58 am, "Steve Song" <ssong.li...@gmail.com> wrote:
> Hi all,
>
> For the last few months Alberto Escudero-Pascual
> (http://www.it46.se/), David Rowe (http://www.rowetel.com/) and I have
> been informally discussing what I am calling a "Village Telco", an
> out-of-the-box telephone company (with data too) built from low-cost
> commodity components. This would include easy-to-use/setup/manage
> technology, network design, links to suppliers, etc but most
> importantly, it would provide the business model for establishing
> I was very happily surprised, shortly after moving to South Africa,
> when Mike Jensen introduced me to Rael Lissoos who runs a company
> called Dabba (dabba.co.za). Rael already has what I consider a
> working model of a Village Telco in Orange Farm south of Johannesburg.
Steve Song
Thanks for your thoughtful message. I am eager to collaborate with
the work that you are doing at CSIR. First let me say that yes Open
Source is the default assumption but so is the notion of a sustainable
business model built on Open Source tools. Let me try and address
some of the very valid issues you have raised.
Connect In versus Connect Out
====================
I think of this as the "connect in" versus the "connect out" model for
telcos. Traditional telcos are based on the notion of centralised
roll-out and by extension "controlled scarcity". The reason the
Internet was able to spread so quickly is in large part due to the
fact that there was no centralised control. Anyone with the energy,
resources, and expertise could and did "connect in" to the Internet.
That is how the Internet first arrived at Stellenbosch University in
South Africa and how it has spread everywhere. This is not to deny
the importance of national infrastructure rollouts but there is no
reason why the two concepts cannot co-exist.
Why a Telco and not a WISP
==================
My experience and instincts tell me (comments from others welcome
here) that voice continues to be the "killer app" that delivers
consistent fundamental value to people. I believe this not just
because I have observed it in many projects and circumstances but was
recently able to validate it personally when I moved to Cape Town a
few months ago. In getting set up here, the one thing I absolutely
needed was a phone. Internet was very handy (especially for a
net-connected type such as myself) but I could survive without it. I
couldn't have managed without a phone. So that is the mission....
deliver local, scalable voice services at very low cost. Plus
Internet access too but that is a value-added service. The technology
is pretty much the same as a WISP but it is a "what you make work
first" issue.
On Regulatory Smackdowns and the Importance of Being Local
========================================
Another more-or-less fundamental premise is that a Village Telco with
no upstream connectivity at all that offers low-cost voice within its
own network would deliver sufficient value to justify its existence.
This is based on my assumption (for which I am still looking for
solid research evidence to validate) that the majority of phone calls
are local anyway. While there will always be migrant workers and
dispersed families that want to communicate long-distance, they are a
comparative minority in most places. Even if local calls are not a
majority, I believe the percentage is large enough to make the Village
Telco concept worthwhile.
This means that a Village Telco should be able to operate in a fairly
restrictive regulatory environment, perhaps meshing/connecting to
other Village Telcos before finally finding a regulatory dispensation
or workaround to connect upstream.
In the case of Dabba, they are a Value Added Network (VAN), which
under the new Electronic Communications Act allows them (short of
"self-provisioning" which is a very woolly concept in the wireless
world) to connect their clients upstream. It even allows them to
assign real phone numbers to individual SIP handsets. Another upside
of being a VAN is that they do reap the benefits of the extortionate
interconnect fees in South Africa. Other telcos must pay to terminate
calls on their network. They also benefit from the reduced
interconnect fees applicable in underserviced areas. In my opinion,
there is a value proposition for both the Village Telco and for the
upstream providers who get many more possible callers on their
networks. This is "network effects" in action, Economics of Abundance
101. :-)
So what
=====
So what I a suggesting here is an Open Source / Open Content / Open
Hardware initiative that would build on Dabba's successful pilot,
bringing in experience and expertise from other wireless initiatives
that would make it as cheap and as easy as possible to set up a local
telco. What I am imagining is:
- very inexpensive, self-configuring wireless access points that
connect themselves to an existing network
- an elegant, no-brainer, web-based telco and client management system
that has one point of interface for all telco management
- swiss-army knife-like connectivity that can network IP, GSM, POTS,
etc connections in a least-cost-routing environment
- cheap pre-configured handsets that are as simple as a mobile phone
for people to use.
- templated business spreadsheets, pay-as-you-go cards, everything you
need so that setting up a Village Telco becomes as easy as setting up
a blog :-)
Dabba has gone further towards this goal than anyone I have ever
encountered which is why I am proposing to base the initiative on
their success. What is best for me about Dabba is that Rael sees the
potential of Open Source as complementary to both his business and
social vision. I should say no more on that score as Rael is on the
list and can speak for himself.
Lastly, while I am thinking of this as a huge leg-up for local
entrepreneurs and for interested upstream providers, there is no
reason why an Open Source Village Telco project could not be carried
out by a community or a cooperative. Let many models flourish. :-)
So, anyone interested in helping get this idea off the ground?
Cheers.... Steve
------
Steve Song
Telecommunications Fellow
Shuttleworth Foundation
steve...@shuttleworthfoundation.org
http://manypossibilities.net
David Rowe
> As I am proposing that we build on Rael's successful pilot in Orange
> Farm, he can answer the question in far more detail than I can.
Actually Rael I would really like to hear what your (i) ideal and (ii)
minimum box should do (features, cost etc). In particular can you
please explain how the billing system should work, as all of the other
building blocks (commodity harwdare etc) seem to be falling into place.
Steve, your idea of an open source Village Telco "distro" sounds great.
I would like to work with you on the project planning - I can perhaps
help estimate the scope of the technical development tasks.
A couple of technical areas that I can offer assistance in:
1/ Low cost open hardware VOIP hardware (the IP04 and soon IP08) that
runs uClinux and Asterisk. The IP08 will have 8 ports and USB which can
connect to a low cost wireless USB key. This means an "all in one"
Wifi-VOIP box. Power consumption is just 3W.
2/ The IP0X devices can run web servers, SQL servers, PHP e.g. they can
host the billing systems.
3/ At GK3 we demoed a "headless" IP04 to IP04 link, i.e. you can set one
up following voice prompts on a phone without a screen or laptop. Ideal
for set up in a remote village. In the future perhaps we can exchange
the written documentation and just install local-language voice prompts.
4/ I am also working on a "$10 ATA" project, which allows any commodity
router to connect to an analog phone and become a VOIP device. The
hardware is simple enough to be built by anyone, parts cost around USD
$10. It can attached to virtually any commodity router to turn it into
a wireless ATA.
Early Goal
----------
Can I suggest a "release early, release often" approach? I feel there
is great value in deploying a basic system ASAP (like in a few months).
Add the bells and whistles later, and work to feedback from the field.
Cheers,
David
--
Free Telephony Project
open embedded IP-PBX hardware and software
http://www.rowetel.com/ucasterisk
suhail
I have really enjoyed reading your first post and the discussions that
followed.
I am an entrepreneur in Tanzania with a key passion for rural
communication and 'alternative telco business ideas'. There are many
options that I have been playing with that are all variations of the
WISP concept, BUT I agree with you one one thing now - VOICE first,
DATA next and many other things later.
The business principle that I come with is that 'everyone along the
chain must make money' and 'the solution must be sell-able to ANY sort
of rural entity/person'. By the second statement I mean that the
solution should be such that a rural cooperative, school, small
business or individual should be able to get involved and start
churning minutes in and out.
To the building onto this 'open source, open access, everyone makes
money' business concept/technology solution, I would like to add that:
a) Each Mesh Node be completely Solar - and there are many successful
case studies and examples where it works perfectly
b) Linking of each 'rural box' to the 'base' - this has been a
challenge for me and I can only think of low cost Ku-Band VSAT for now
c) Allocation of addressable numbers - This is a scenario where each
business would either piggy back on an existing telco's numbers or
that each business gets its own set of numbers
So I would be very keen and interested in playing a part in developing
this unique 'prototype' based on the principles mentioned and others
that previous posters share. I would also be very keep on deploying a
test site to see reaction/usage/feedback and 'appropriate-ness' of the
service offered.
There seems to be several pieces of the solution worked out well. We
can identify the areas that need more work and start testing it. I
would also love to see the dabba example. Seems like Rael is already
ahead of the game!
Would someone be able to take the lead and coordinate the building of
this concept? Which I see as becoming a 'standard' for anyone wishing
to take up the concept.
Thank you,
Suhail.
chrisfm
It is good to hear from you and I enjoyed your comments around the
need for the business model. Some recent discussions on deep rural
wireless networks in Zambia lead me wondering whether the Western
approach to business was the right way to go. You just restored my
belief in the need for business models.
I read some new research that indicated that as much as 60% of calls
are made locally. Now whether your a capitalist or not that means
someone somewhere is making a lot of money out of people who are
desperate for voice communication - despite the high cost!
This is also a strong motivation for researchers to develop
alternative technology solutions that empower communities and
challenge the established service providers. I recently had to defend
why we were pushing technology for rural development and I used the
mobile phone scenario as my counter-argument.
Hence our belief that we can minimise the outflow of revenue by
aggregating demand at the local level through community owned wireless
networks and stimulating the demand by offering VoIP services.
Our approach is to develop the voice services first, however, in order
to maximise the benefits of internet, we belive the WISP in a box is a
logical follow on to the VoIP. This needs to be substantiated by
demand side studies that capture data, server logs etc. to understand
and develop the value-add services required in communities.
So What?
I think if we can scale and replicate these wireless networks,
eventually we could have enough interconnected nodes/cells to start
building a FON like network of African villages. Howeverm we need to
do more on RENEWABLES as Suhail pointed out.
Chris
Steve Song
Sorry for the slow reply.
[snip]
> Steve, your idea of an open source Village Telco "distro" sounds great.
> I would like to work with you on the project planning - I can perhaps
> help estimate the scope of the technical development tasks.
>
> A couple of technical areas that I can offer assistance in:
>
> 1/ Low cost open hardware VOIP hardware (the IP04 and soon IP08) that
> runs uClinux and Asterisk. The IP08 will have 8 ports and USB which can
> connect to a low cost wireless USB key. This means an "all in one"
> Wifi-VOIP box. Power consumption is just 3W.
I can see lots of scope for creative thinking on scale and
scalability. It may even be that there are small, medium, and large
villagetelco models. Having a standardised way for them to
interact/self-configure would be important. Perhaps creating a kind
of standardised protocol for a node to connect to a villagetelco hub
would be worth exploring.
> 2/ The IP0X devices can run web servers, SQL servers, PHP e.g. they can
> host the billing systems.
So, an IP0X coud be a node or a hub. Keen to see how far it can go.
> 3/ At GK3 we demoed a "headless" IP04 to IP04 link, i.e. you can set one
> up following voice prompts on a phone without a screen or laptop. Ideal
> for set up in a remote village. In the future perhaps we can exchange
> the written documentation and just install local-language voice prompts.
A voice prompt driven install is a very cool idea.
> 4/ I am also working on a "$10 ATA" project, which allows any commodity
> router to connect to an analog phone and become a VOIP device. The
> hardware is simple enough to be built by anyone, parts cost around USD
> $10. It can attached to virtually any commodity router to turn it into
> a wireless ATA.
How about a $15 dollar ATA+WiFi device :-)
> Early Goal
> ----------
>
> Can I suggest a "release early, release often" approach? I feel there
> is great value in deploying a basic system ASAP (like in a few months).
> Add the bells and whistles later, and work to feedback from the field.
Totally agree on the "release early, release often principle".
Standby for more news shortly!
-Steve
Steve Song
Really appreciate your message. You are just the sort of person who
needs to be involved in this project!
On Sat, Mar 15, 2008 at 1:48 PM, suhail <suh...@laiki.co.tz> wrote:
>
> Hi Steve et all,
>
> I have really enjoyed reading your first post and the discussions that
> followed.
>
> I am an entrepreneur in Tanzania with a key passion for rural
> communication and 'alternative telco business ideas'. There are many
> options that I have been playing with that are all variations of the
> WISP concept, BUT I agree with you one one thing now - VOICE first,
> DATA next and many other things later.
So pleased to hear that affirmation. Mike Best of Georgia Tech also
affirmed this to me recently. He had a similar experience of the
importance of "voice first" in India.
> The business principle that I come with is that 'everyone along the
> chain must make money' and 'the solution must be sell-able to ANY sort
> of rural entity/person'. By the second statement I mean that the
> solution should be such that a rural cooperative, school, small
> business or individual should be able to get involved and start
> churning minutes in and out.
Right on!
> To the building onto this 'open source, open access, everyone makes
> money' business concept/technology solution, I would like to add that:
>
> a) Each Mesh Node be completely Solar - and there are many successful
> case studies and examples where it works perfectly
Agree. There is a lot of solar energy technology and supply chain
expertise on this list. Mark Summer, Alberto Escudero-Pascual, and
Rael come to mind but it sounds like you are pretty plugged in too.
Creating a space for sharing and comparing solutions is a very good
idea.
> b) Linking of each 'rural box' to the 'base' - this has been a
> challenge for me and I can only think of low cost Ku-Band VSAT for now
Interesting question and this comes back to the issue of scale. What
sort of distances are you talking about linking?
> c) Allocation of addressable numbers - This is a scenario where each
> business would either piggy back on an existing telco's numbers or
> that each business gets its own set of numbers
I imagine some of this will be determined by national policy and
regulatory frameworks and the legal viability of getting POTS phone
numbers that can be mapped onto a villagetelco network. I was amazed
to discover that Dabba have succeeded in getting their own set of
numbers although I gather it took a lot of time and no small amount of
bureaucracy to achieve. It will be interesting to see to how many
elements of this can be generalised across countries.
> So I would be very keen and interested in playing a part in developing
> this unique 'prototype' based on the principles mentioned and others
> that previous posters share. I would also be very keep on deploying a
> test site to see reaction/usage/feedback and 'appropriate-ness' of the
> service offered.
That is brilliant. Stand by. In a couple of days, we should have a
shared web space up where Rael can begin to document the work he has
been doing and everyone can share in the prototype development.
> There seems to be several pieces of the solution worked out well. We
> can identify the areas that need more work and start testing it. I
> would also love to see the dabba example. Seems like Rael is already
> ahead of the game!
>
> Would someone be able to take the lead and coordinate the building of
> this concept? Which I see as becoming a 'standard' for anyone wishing
> to take up the concept.
For my part, I am committed to seeing entrepreneurs such as yourself
and Rael succeed and to ensuring the maximum amount of transparency
and knowledge sharing about how to go about it.
Cheers... Steve
--
David Rowe
> I can see lots of scope for creative thinking on scale and
> scalability. It may even be that there are small, medium, and large
> villagetelco models. Having a standardised way for them to
> interact/self-configure would be important. Perhaps creating a kind
> of standardised protocol for a node to connect to a villagetelco hub
> would be worth exploring.
I was thinking of a way to use existing protocols, for example:
1/ Unpack a node, and power it up. Lets say the basic node has an
antenna, a box, and a telephone handset. A light comes on or dial tone
tone appears in the handset when it associates with a wifi mesh. For a
several km link you gently rotate the antenna around on the mast, and
listen for the change in tone which indicates signal strength.
2/ You can then just dial a number on the phone. Like to talk to the
node on 10.0.0.1 you dial 10001. If your node is 10.0.0.2 then your
number is 10002. Unique IP's could be hashed off Ethernet MACs, or I am
sure there are many ways to allocate numbers. Note we are simply using
existing IP routing technology to route calls.
3/ To make life easy the nodes just talk SIP. That way any commodity
SIP phone can be a node.
4/ By default u get free calls with any node on the village telco mesh.
If u want to make an off-mesh call (GSM or PSTN) a billing system sits
at the gateway. So the user experience is (i) dial the gateway IP, say
100003 (ii) work thru the gateway's IVR system with a calling card to
access the GSM/PSTN. Anyone can put up a gateway, encouraging
competition. The gateway could be any Asterisk type box (embedded or
even x86). Open calling card software is available today for Asterisk.
5/ In a cyber cafe environment, the calling card system could be added
to the node, so the cafe can make money from ad-hoc users. In that case
a calling card is used to access the village telco mesh.
> How about a $15 dollar ATA+WiFi device :-)
Actually this is possible. That's about the manufacture cost of a
WRT54. If we can cover R&D costs and a volume run it is totally
possible to build a "Meraki with a phone" device for $20-30 in maybe
quantity 10k to 100k.
We could give 100,000 people phones for maybe $3M. You know with a
network _that_ big the phone companies/governments would be coming to
Village Telco, not the other way around.
chrisfm
I like the idea of the $15 ATA plus wifi phone. What are your thoughts
on an Asterix based authoring tool for creating community based
information services? We have developed something called Open Phone
(diallogue pallet) see www.fmfi.org.za
Chris
carolune
What an interesting project! I am not involved in any technology
development any more but being an engineer I can follow your
discussion quite well. I would love to be kept informed about this
because I already have an interest in what you're discussing.
I wouldn't want to be completely off topic so I'll keep it short. I am
the manager of an organisation called Universe Awareness for Young
Children (www.unawe.org) . We use astronomy to broaden children's
minds, stimulate their curiosity in science and world citizensip. In
order to achieve that we organise - among other things - events
whereby young underprivileged children across the world talk to each
other and compare their lives and environments, usually based around
some astronomical event, e.g. an eclipse (www.unawe.org/
eclipse2008/) . Because of the people involved (poor communities
worldwide with minimal if any techno-literacy and very young
children), things have to be kept as simple as possible, and as cheap
as possible.
The Village Telco concept you are describing would be an ideal medium
for us to reach out to those underprivileged communities and connect
them with others. So far we have mostly used skype because it has most
often taken place via 3G connection on a laptop (we've had quite
remarkable video up- and down- stream quality in South Africa) but any
other protocol that could be interfaced with any other would be
suitable - keeping the cost low and the software free, if software is
necessary. I mention video but we've done a lot voice-based too, of
course.
So if you would like to try something out at any point we would be
very keen to test it and if we can contribute to a business plan with
profiling demands from communities for this sort of educational
activities, we'd be very happy to do so!
Sorry for the off-topic post - I really look forward to following-up
on this.
Cheers,
Carolina
Steve Song
UNAWE sounds very cool! I think there are all sort of potential knock-
on benefits that the village telco concept could provide to education
initiatives. Actually, if you saw Rael's earlier post on education,
he has done that in adding education services to the Dabba
installation in Orange Farm.
-Steve
Alberto Escudero-Pascual
Sorry for the silence. :) Good to see the discussion. What I am really
interested is to have communities willing to provide continuous feedback
including wish-lists to our development.
I see the challenge of capturing users needs and be able to re-act
quickly adapting the technology or the business models. Orange Farm
sounds like a good place to bridge users-business-development.
Anyone can suggest any other active communities where users are eager to
work together with tech-developers.
There are many models to explore
VT: Village Telco
VOT: Village Own/Operated Telco
VOOT: Village Own and Operated Telco
/aep
> > The Village Telco concept you ahttp://www.voip-info.org/wiki/view/Asterisk+codecsre describing would be an ideal medium
Alex Gakuru
I am passionate about this idea! Below relevant story, earlier posted
to this list >> excuse me for re-posting, was published early last
year ion Kenya. I am currently attending Idlelo 3 at Dakar, Senegal
<http://www.aiti-kace.com.gh/idlelo/?q=node/16> and where I mentioned
it (and this thread!) in the context of possible "Open Telecoms"
business models.
Regards,
Alex
PS:
I am ok with the new list reply settings being in line with "3.0
One-to-Many Communication (Mailing Lists, NetNews)"
http://www.ietf.org/rfc/rfc1855.txt
AG
------
[Business Daily]
"Door opens for new non-profit telecom firms"
Licence-free band spectrum, courtesy of the airwaves' regulator, could
allow non-profit organisations to own and operate telecom companies.
The Communication Commission of Kenya's offer of the ISM Band 2.4 and
5.8 spectrum to registered community groups is on a
first-come-first-served basis, and already organisations from
Mukurweini, Khwisero, Limuru and Rangwe have expressed interest.
Countries like Tanzania, Namibia, Bangladesh and India have used the
concept of free frequencies in efforts to bridge the 'digital divide'
with the West.
--<snip>--
<http://www.bdafrica.com/index.php?option=com_content&task=view&id=363&Itemid=5810>
regards,
Alex
Mike Jensen
Alberto Escudero-Pascual wrote:
> I see the challenge of capturing users needs and be able to re-act
> quickly adapting the technology or the business models. Orange Farm
> There are many models to explore
> VT: Village Telco
> VOT: Village Own/Operated Telco
> VOOT: Village Own and Operated Telco
I think Alberto's hit the nail here, i suspect that the technical
implementation of the Village Telco won't be as hard as figuring out the
optimal 'business' models to maximise quick uptake. Clearly there
probably won't be just _one_ model, so it will be good to make sure the
full range of potential options are considered, and where each might
best be suited. It seems there are a wide range of possibilities, from
'Community owned/Cooperative' telcos, to village SMEs and possibly
municipal authorities, to more traditional regional or national businesses?
Mike
Ian Howard
more readily than the business model. This is also the case in sites
that I have worked on in Congo, Tanzania, Lebanon and in Ethiopia. The
business model is always difficult, but there are good models for it. In
my opinion the design must start with the economic model that has to
support it, unless a benevolent institution can support it for long
enough for the village telco to be meaningful (most projects have too
short a horizon and spend too much on "stuff" and not enough on keeping
the thing running). Great thread, thanks!
Ian
> --
Ian Douglas Howard
co-Founder, Innovator
Adapted Consulting Inc. - http://adaptedconsulting.com
iho...@adaptedconsulting.com
Toronto: +1 416 848 4130 x1
Washington: +1 202 292 4242 x1
skype: iandouglashoward
Candidate International MBA, Schulich School of Business
Chapter Leader, Schulich NetImpact
Steve Song
Ian wrote the chapter on Economic Sustainability in the new version of
Wireless Networking in the Developing World. You can download that
chapter here http://wndw.net/pdf/wndw2-en/ch10-economics.pdf or the
whole book from http://wndw.net/pdf/wndw2-en/wndw2-ebook.pdf
Cheers... Steve
--
sebastian
Hi all,
Sorry for the silence. :)
and the same apologies from me - the reason for the silence has in fact been that we are very busy working with what i assume is our common goals - but i will try to contribute more from here on again.
I am currently in South Africa,, at the Meraka institute, working with the Wireless(ISP)-in-a-box project ( .... yes, we will find a better name ;) )
Anyone can suggest any other active communities where users are eager to
work together with tech-developers.
Spontaneously adding FADECO
http://www.fadeco.org/
which i find an impressive example of an organically grown bootstrapped network project.
I am confident, the people there with whom i have worked in out November 2007 Wireless Roadshow would be willing to share experiences and feedback.
best to all,
sebastian
Ian Howard
I very much concur with Sebastian (and great to hear of the work on the
WISP in a box work Sebastian!), FADECO is a great example of what we
hope to see more of -- a true social entrepreneur at work.
Here is my interview with Joseph at FADECO last summer:
http://www.youtube.com/watch?v=ErXK29u4Pf8
Sante Sana!
Ian
Mike Jensen
> 3/ To make life easy the nodes just talk SIP. That way any commodity
> SIP phone can be a node.
Talking to Rael yesterday brought up the issue of Codecs - apparently
most of the big operators (to which the village telco will need to
interface) want to talk G.729 which is a proprietary protocol, and
increases the cost of each handset by $4-$5.. At the gateway G729 could
be converted to a free/open codec, but even then it will probably be
necessary to standardise on one codec for the phones/handsets, otherwise
folks inside the village telco with different codec handsets won't be
able to talk to each other..
Mike
Graham Todd
At least with asterisk their is a "pass thru" mode where connections
between end points (i.e. phones which may already have g729 builtin) can
occur without a per channel license free and sans transcoding.
For trunks between asterisks IAX works very well ... I have a vague
sense that a network topology could be dreamt up that would minimize
g729 license requirements for the village network operator (VNO) :)
Alberto Escudero-Pascual
codec like GSM vs G729. If GSM is used with IAX (natively supported in
the IP04) the traffic is only 12% more than using G.729. There is
already a version of IP04 with G.729 but it is also subject to license
issues.
In the networks i am designing, i am running G.711 inside of the wired
network and transcoding to G.729a for international connections.
G729 has a few drawbacks when it comes to packet synchronization.
Asterisk IPP based codec for G729 at the moment has not implemented VAD
and noise suppression.
What really surprises me is the requirement of G729, as this codec
requires more heavy investements from the Telco side. In Sweden, G729
has been dropped as processing is more expensive that bandwidth.
/aep
David Rowe
A g729 licence from Sipro was recently quoted at USD$50k plus $1.50 per
channel (e.g. port/phone). It would be possible for us to buy a "site"
license for village telco if reqd, say by funding from an NGO. For a
good cause such as this we might get a discount or even waiver.
Speex (an open source/free codec) is starting to get a lot of traction
(e.g. people are choosing it over g729). It has various bit
rate/quality trade offs (4-32 kbit/s) and can even do wideband audio.
At 15 kbits/ it is far superior to GSM at 13 kbit/s and g729 in speech
quality. At 8kbit/s roughly the same as g729, but far far less pain to
license and use. A notable problem is lack of support in ATAs or
commodity phones, at least at the moment.
A general comment - only a proportion of village telco calls will be
long distance calls terminated by ITSPs that only support g729. Many
will be "on net" (i.e. village telco to village telco), so we can use
any codec we want - others will go straight to GSM or the PSTN so once
again the codec is up to us. An easy fallback is GSM - virtually all
Asterisk based ITSPs will support this.
As well as for practical reasons (like cost and overhead to manage
licenses) it would be nice if Village Telco makes a moral stand against
closed software like g729. Just seems like a backwards step to me.
Alberto Escudero-Pascual
products?
/aep
Mike Jensen
> A general comment - only a proportion of village telco calls will be
> long distance calls terminated by ITSPs that only support g729. Many
> will be "on net" (i.e. village telco to village telco), so we can use
> any codec we want - others will go straight to GSM or the PSTN so once
> again the codec is up to us. An easy fallback is GSM - virtually all
> Asterisk based ITSPs will support this.
I guess what seemed to be more of an issue is will any commodity SIP
handset work on the network, or do they have to be ones with the VTelco
'approved' codec..
> As well as for practical reasons (like cost and overhead to manage
> licenses) it would be nice if Village Telco makes a moral stand against
> closed software like g729. Just seems like a backwards step to me.
Agreed!
Mike
Rael
The idea of engineering a wifi-ATA or wifi-sip-ip-phone unit is great.
I wish we could underwrite a 10k order.
The wifi device could also run a version of ROBIN (BATMAN) so each
phone itself becomes a node.
On the issue of phone numbers we are trying to use real routable phone
numbers so each village phone can receive calls. In South Africa we
have managed to get our own number range.
In many developed countries you can easily purchase numbers. In the so-
called less developed countries the telco's are often monopolies or
pseudo monopolies (one of the main reasons for the high prices). When
we began our project in South Africa we were on the verge of issuing
U.S. numbers to the rural villages as these were easier to get and are
internationally routable, and ironically would have been cheaper(call
costs) than the local mobile network.
Kindest Regards
Rael
Graham Todd
....
> Unique IP's could be hashed off Ethernet MACs, or I am
> sure there are many ways to allocate numbers.
Sounds like ipv6 :)
--
Sent/Read with my dovecot.org | postfix.org servers.
Alberto Escudero-Pascual
David Rowe
I have just done some work on reducing the power consumption of the
IP04. Four FXS ports (driving 4 telephones) has an idle power of 3W,
compared to say an x86 PC at 34W. This dramatically reduces the size
and cost of the solar power system:
Details on this blog post:
http://www.rowetel.com/blog/?p=61