Projects in Kenya or highlands of Panama

[Steve Smith]

I am building a small testbed in a rural farm area of northern NM with an eye to helping two friends possibly bring a VillageTelco based mesh into the Marakwet region of Kenya and into the highlands of Panama (near Rio Serreno).  

I would like to find anyone already operating in those regions.  My main questions at this point have to do with:

1. Backhaul
1. Terrestrial Fiber
2. Cellular Uplink
3. Satellite
   
2. Locally "appropriate technology"
1. How does this impact locals?
2. What is the right order of deployment?
3. How does it fit with existing mobile devices?
   
3. Deployment and Maintenance
1. What skills are available locally?
2. What models for D&M have been tried?
   
4. Power sources
1. Mains
2. Solar
3. Charging Kiosks
   
5. Economics
1. Who can afford?
2. How will it improve economics for individuals?
3. How will it improve economics for regions?
   

[Kent McNaughton]

Hi Steve,

I'm about 70km (as the crow flies) from Rio Sereno (RS) in Western Panama.

In terms of your questions:

Backhaul: is expensive. There are, AFAIK, just three suppliers. The
phone company; the electric company; and a wireless phone and internet
supplier.

Terrestrial fiber: I believe the cable TV company and the phone company
has fiber running the length of the country.

Cellular Uplink: two companies I know of have a fair presence in most
towns of Western Panama, possibly two more who are relatively new to the
area might serve smaller towns like Rio Sereno. Rio Sereno is fairly
remote. I don't know who serves them, but I'd bet at least one of the
former does.

Satellite: I checked some years back. It was prohibitively expensive.

Local "appropriate technology": In the towns, wired service is provided
by one of the entities mentioned above. Not at all far outside the towns
people rely on wireless from one of three companies. It's expensive due
to backhaul and maintenance costs.

Impact on locals: Hard to say. Almost all have cell phones. I'd be
surprised if WiFi wasn't widely available in RS. I can't speak to home
Internet, thinking it's probably available from the phone company in RS,
but unlikely outside it.

Deployment: This is really hard to say. When you're closer to knowing
you want to do it, I think a visit would be in order.

Skills: Are available in the province. No idea whether in RS.

D&M: Sorry. I'm not familiar with the acronym.

Power sources: Power is 120V, 60Hz. Actual power supplied runs from
about 102V to 135V (so says a UPS I was testing for someone). Power is
generally available with very short outages (.5-10sec) pretty much
daily, longer outages (30m-several hours) monthly to quarterly.

Mains: see above.

Solar: Dry season (mid-Dec--early-May) sees sun every day. Wet season
sees morning sun for two to four hours most days--not much in the PM.
Varies somewhat year-to-year.
Wind: Dry season is windy. Wet season has little.

Charging kiosks: For batteries?

Economics: Western Panama is agricultural. the minimum wage is between
$12 and $15 a day (I'm not sure). A home phone w/o bells/whistles, last
I knew, was $10/month. Outside of town, wireless and cell phone are the
only options. Wireless runs $60 for 1/2Mbps, $85 for 1Mbps and $150 for
2Mbps. This is after one pays for installation ($125 to $150).

Affordability: Easiest if one lives in town. Outside, as you can see,
one would rely on WiFi, mostly.


Wishing you the best,
Kent McNaughton

On 09/16/2015 01:01 AM, Steve Smith wrote:
> I am building a small testbed in a rural farm area of northern NM with an
> eye to helping two friends possibly bring a VillageTelco based mesh into
> the Marakwet region of Kenya and into the highlands of Panama (near Rio
> Serreno).
>
> I would like to find anyone already operating in those regions. My main
> questions at this point have to do with:
>

> 1. Backhaul
> 1. Terrestrial Fiber
> 2. Cellular Uplink
> 3. Satellite
> 2. Locally "appropriate technology"
> 1. How does this impact locals?
> 2. What is the right order of deployment?
> 3. How does it fit with existing mobile devices?
> 3. Deployment and Maintenance
> 1. What skills are available locally?
> 2. What models for D&M have been tried?
> 4. Power sources
> 1. Mains
> 2. Solar
> 3. Charging Kiosks
> 5. Economics
> 1. Who can afford?
> 2. How will it improve economics for individuals?
> 3. How will it improve economics for regions?
>
>

[T Gillett]

Hi Steve

Some high level questions to illustrate the overall aims of the projects.

What sort of outcomes do you want to achieve from these deployments?

E.G. Are you seeking to set up a community telephone network and/or data network?

How big a geographical area do you want to cover?

How many nodes would you anticipate being in the meshes?

What will be served from a node - a single device or a collection of devices in a home or business?

Are these 'last mile' Internet distribution initiatives, or more like local community data/telephony networking initiatives?

Regards
Terry

--
You received this message because you are subscribed to the Google Groups "Village Telco Development Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to village-telco-...@googlegroups.com.
To post to this group, send email to village-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/f209bb6f-5ecc-4179-9640-70af2d9231cf%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Steve Smith]

Kent -

thanks for weighing in...  Are you in the Boquette region, or closer to David district?
Do you have a mesh/telco running?

Rio Sereno *does* have cellular in the area.   As you indicate, reports are that most folks, even if they have phones can't really afford to use them much if at all.  My contacts living in the hills there get service but even with expatriate assets can't really afford it as "normal" internet.  Performance is also very limited.  

There is apparently one very weak Internet Cafe in RS... one computer and a very spotty/slow WiFi.

The elementary school *just* got power back after maybe more than a year (someone stripped the mains for copper).   They have a couple of ancient computers.   But without internet (or power) their utility is small.   There have been others who have put in "caching proxy servers" in situations like this so that redundant access to materials is not re-retrieved...   for example, study plans, reference materials, etc.

"D&M" was just an abbreviation for Deployment and Maintenance.     Mostly wondering how far and few between someone might be who can evaluate the condition of a battery on a solar charger or reflash/configure a MP, etc.

Charging kiosks: For batteries?

 

There has been some effort to "rent" battery/inverter packs from centralized kiosks in the third world where you "trade an empty for a full" if you can't otherwise get to mains to recharge...   similar to our "propane exchange" systems here.

 

Economics: Western Panama is agricultural. the minimum wage is between
$12 and $15 a day (I'm not sure).

I'd been given the number of $10/day... maybe because it is more rural yet!
 

A home phone w/o bells/whistles, last
I knew, was $10/month. Outside of town, wireless and cell phone are the
only options. Wireless runs $60 for 1/2Mbps, $85 for 1Mbps and $150 for
2Mbps. This is after one pays for installation ($125 to $150).

thanks!

- Steve

[Steve Smith]

Terry -

Some high level questions to illustrate the overall aims of the projects.

What sort of outcomes do you want to achieve from these deployments?

E.G. Are you seeking to set up a community telephone network and/or data network?

 

In both cases, my interest is to create a very small system based on opporunistic criteria.  It would be working demonstration system to start.  Probably serving 1 school and a few dozen households at best. to start.
 

How big a geographical area do you want to cover?

At first it would be just what can be covered in a single mesh...  but the point is to have the backhaul and the tech (ubiquiti-class point-to-point?) to extend this to several such meshes.  
 

How many nodes would you anticipate being in the meshes?

Ideally a few dozen, growing to whatever the topographical and other practical constraints impose.
 

What will be served from a node - a single device or a collection of devices in a home or business?

One or a very few devices.  Possibly mostly mobile devices.  In both areas there is a modest penetration of mobile devices (vanity symbols?) despite limited use due to coverage, performance, and cost.
 

Are these 'last mile' Internet distribution initiatives, or more like local community data/telephony networking initiatives?

There may be significant added value for "community telephony" over the system,  but with the penetration of cell phones, it seems as though even telephony will be via that route.

If these end up "school centric" then "community data" will be a significant aspect...   a cacheng proxy server somewhere for regularly used data...  a 'library" of  locally useful information, etc.

My "local" testbed will be used primarily to provide internet access where it is prohibitively expensive/weak/vacant.   We have lots of small farms distributed along dry rivers with irrigation ditches, etc.   The phone/cable companies have never done well and now with expired tariffs, they no longer bother to maintain/upgrade things when they fail, leaving lots of folks with little or no phone, much less data.    Wireless is available in about half of the cases, but the companies providing are starting to demand long (2 year contracts) and fairly expensive installs.  

Many people simply give up on internet access at home and use the library, etc.  when many need no more than  e-mail and maybe some incidental web browsing.    Those who want streaming video can and do pay the weight for wireless... though if they are shaded from the towers (lots of tall cottonwoods) the ISPs generally just refuse to try to serve them.  

We have a good public-paid fiber backbone nearby but their economic model is to resell to commercial ISPs, I'm trying to establish an effective example for community broadband/coops in this area as well, despite the existing "competition".   With even a small mesh in place, I think the fiber "owners" will talk... right now they are only talking to established commercial ISPs.   

- Steve

[T Gillett]

Hi Steve

I do some work in rural SE Asia, and the situation sounds similar to what you have described.

As you say, cellular has largely taken over the telephony side of things, although in some situations wireless mesh telephony has something to offer in providing fixed telephone points around a community or campus as the ongoing call costs are zero.

"Last mile" Internet access is also useful in some situations as long as the usage is reasonably modest.

One nice aspect of mesh solutions is that they can avoid the use of Ethernet cabling on a campus, which can reduce cost and maintenance issues (rats enjoy chewing on CAT 5 cable!)

For a school or community organisation situation, you may be interested in the digital library idea whereby content is stored and served locally - along the lines of what you have described as a 'caching server'.

There are several projects (eg Pirate Box, RACHEL, Internet-In-A-Box, XSCE School Server) that have developed solutions for this. Some are quite complex and relatively expensive, and may be problematic to maintain in remote areas.

The RACHEL project has used RPi devices, but I have found that in their basic form they are not robust enough to survive a year's unattended service, which is my yardstick for success. But they have an excellent set of content intended for use in schools and remote communities.

We have developed VT SECN firmware for providing essentially RACHEL content on VT SECN devices, including the MP02 and various TP Link routers. In simplest form, these are standalone classroom micro-servers.

See these links for some additional info, and for an on-line preview of the standard RACHEL content:

    http://wiki.villagetelco.org/RACHEL_on_MP2 
    http://www.worldpossible.org/rachel/    
    http://schools-wikipedia.org

The RACHEL web content, plus whatever else you want to serve up, is stored on a 32/64GB USB/SD memory device which is plugged into the VT SECN based router and served up via the built in web server.

The standard RACHEL content on a 32GB memory includes Wikipedia for Schools, Khan Academy, Gutenberg books, Medline and Hesperian medical references, and a selection of other resources. But you can add and subtract whatever content you like.

The nice thing about this approach is that it is quite economical (~$50 basic), and the devices are simple and robust in terms of both software and hardware, being completely solid state.
The devices are portable, so can be taken to a classroom as required, and stored safely when not in use.
Obviously the solution is limited to static web content, but that actually gives a lot of scope.

The performance is quite surprising from standard routers.
For example, an MP02 can support 30+ concurrent wifi connections, so you can have a classroom full of students browsing content from the device.

A 5MB Khan Academy video is served up in about a second and then runs for 5 mins on the client device, so you can easily have the classroom all watching videos concurrently, which is not at all intuitive from such a small server.

We have tested with 20 students hitting Wikipedia for Schools page based content and changing pages as fast as possible and seeing no reduction in sub-second response times.

There is OS based memory caching in operation, so requests for recent pages are served from RAM rather than Flash, which is obviously faster. The cache hit rate in a classroom is statistically good, given a common area of interest. The more RAM a device has, the more caching is available.

Local content can be added through the web interface, or by directly loading files on to the USB/SD device. You can use any standard file type supported by the client browser - html, pdf, jpg, ogg, mp3, mp4...

If there is even limited upstream Internet access available, you can connect the device via Ethernet cable (or mesh) and it will distribute the access via wifi or Ethernet. Thus you can serve bulk content locally, but still have classroom access to the Internet for low bandwidth usage like email.

Hope this is of interest.

Regards

Terry

--
You received this message because you are subscribed to the Google Groups "Village Telco Development Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to village-telco-...@googlegroups.com.
To post to this group, send email to village-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/cdc4ca8c-5161-4e59-b7ee-b8958f1a7d2a%40googlegroups.com.

[Steve Smith]

Terry -

This is all very interesting and relevant, thank you!  More questions as I follow your references more.

Thanks!

- Steve

[T Gillett]

Hi Steve

Happy to discuss further as I think this is an area with a lot of potential.

Providing access to useful information for schools and community organisations particularly in rural areas of developing countries is probably one of the better ways to provide "aid".

As I see it, there are two critical challenges that have to be met by these systems: 

- they have to be robust and reliable in a situation where they exposed to harsh conditions and where there is little or no maintenance available, and

- and they have to provide content that is really useful to the local community. It has to be relevant to the community and delivered in an appropriate language.

The first is largely a technical problem which can be dealt with by good design and a focus on simplicity / minimising complexity.

The second is arguably a much tougher problem.

Just providing Internet access is generally not a viable answer, as the available bandwidth is not sufficient to support multiple users at anything like a reasonable cost. And there is the issue of providing safe material for use in schools rather than raw Internet access.

Locally stored content can be part of the solution, but it is still not a trivial problem. 

Deciding what content to store, how to keep it updated, and how to keep the system complexity low are still significant challenges.

And locally stored content essentially has to be static as there is no dynamic content engine. (So forget apps like ReadingEggs, unfortunately.)

Putting RACHEL and other static content onto a memory device (USB / SD or HDD) plugged in to a router device can be simple, robust and cost effective, but doesn't deal with the issue of updating the content.

The IIAB project (http://internet-in-a-box.org) has a local store of 1-2TB which includes RACHEL content, Wikiipedia, maps of the world and lots of software development resources. So a very large amount of content, but no built-in update mechanism.

The idea of trickle feeding data to a local content store looks promising. 

The Outernet project (https://outernet.is) is using this approach by feeding data from satellites at rates of 100MB to 1GB per day. 

The basic ground station cost is around $200 using their Lighthouse device and a satellite dish. There are no ongoing data usage costs.

The range of content available is obviously constrained and must be selected by the Outernet management team for broadcast.

Looking at their on-line preview (https://librarian.outernet.is/en/) shows that much of the content is a duplicate of the RACHEL library, but this may change over time. It is not obvious how a user would navigate this content.

There is also the concept of a more real time feed for news, weather warnings and the like.

The same trickle feed approach may be possible using cellular (or other) networks in some areas, with the advantage of being simpler, and including a two way data connection, which can support low bandwidth applications like email. 

I have seen cellular accounts in SE Asia where unlimited data is available at a shaped speed of 256kbps for $5 per month, which could obviously provide a good trickle feed ( ~1 GB per day) with a set up cost of ~$50 and ongoing cost of $60 per year.

The design problem is then how to implement a simple and robust local server, and an application on it that allows the community to browse the Internet and select appropriate material to be downloaded to the local store, and a to allow users to subsequently navigate the library to find material that they want.

All that said and done, I think that the second basic challenge remains - finding content that is really useful and appropriate to local communities. 

I suspect this is not a problem that can be solved in a general way - it requires lots of work to be done on a per-community basis, assuming a good platform is in place. 

There seems to be lots of effort directed to platforms, but not so much on developing approaches to deal with the content side of the equation.

Shared libraries such as RACHEL are a start.

As we have all seen, just dropping in even good technical solutions seldom results in useful outcomes unless the other challenges are dealt with effectively.

Regards

Terry

 

To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/978ec8f3-1c71-45e2-a650-5b1dc498606b%40googlegroups.com.

[Anish Mangal]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi,

Following this discussion with keen interest. I am a community
volunteer for the XSCE (School Server Community Edition) project, and
we deal with the content/media/networking side of things and have
pilots in a few places around the globe.

An FAQ is maintained here

http://wiki.laptop.org/go/XS_Community_Edition/FAQ

Happy to help if this sounds promising.

On the flip side, I'm looking to do pilots to connect villages in the
Himalayan mountains remotely and within themselves through wifi
meshes, so - following this keenly ;-)

Cheers,
Anish

<https://groups.google.com/d/msgid/village-telco-dev/cdc4ca8c-5161-4e59-
b7ee-b8958f1a7d2a%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>> .

>>>>
>>>> For more options, visit https://groups.google.com/d/optout.
>>>>
>>>
>>> --
>> You received this message because you are subscribed to the
>> Google Groups "Village Telco Development Community" group. To
>> unsubscribe from this group and stop receiving emails from it,
>> send an email to village-telco-...@googlegroups.com.
>> To post to this group, send email to
>> village-...@googlegroups.com. To view this discussion on
>> the web visit
>> https://groups.google.com/d/msgid/village-telco-dev/978ec8f3-1c71-45e
2-a650-5b1dc498606b%40googlegroups.com
>>
>>

<https://groups.google.com/d/msgid/village-telco-dev/978ec8f3-1c71-45e2-
a650-5b1dc498606b%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .

>>
>> For more options, visit https://groups.google.com/d/optout.
>>
>

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJV/lccAAoJEBoxUdDHDZVpnK4H/1BtCj7EyhHhsP06HyQT/XOa
z7QsGX26WyRA72D2E3WMpWPfA7z7RQukAb8KbyyAg2bpgSF7IwrnVI+hcTczR1RP
Ok1S0SF4BVM9//xEXvzrhw02qeCtRcZcnI8VdMaIoj/9iSogX8Kdd947d4el6eRG
/qlula3Rxpjx1GGWzM0/queKpJLSrrGaAT2CSaHUbvTKawL4Cpf1hvfzkLHdvnew
cEeoG+X2cp2CdS+zAMJeoaVT6KIh+bvb60xZEtBW1isEg7JJoX448LIJ1h/ddcmb
oeGZyIgc4NrRZu63efeBGqIAbTcBUUqLtocTfF5FhtejvNK+mZ0B7jbaib7IfI8=
=FCdh
-----END PGP SIGNATURE-----

[T Gillett]

Hi All

Just for interest, here are some photos of some recent tests we did where we had 35 laptops (XO and Ubuntu) connected to a VT-RACHEL classroom micro-server. 

The micro-servers are set to accept only 35 wifi connections to prevent excess connections causing overload conditions.

Each laptop was running a browser with several tabs open with a self-refreshing weather map loop from the Internet, a Khan Academy video and pages from Wikipedia for Schools, both served locally from the micro-server.

The micro-server had a 32GB digital library (RACHEL content +) installed on a USB/SD memory, and an upstream connection to a router with Internet access.

Using an MP02 device as the microserver, the upstream connection can use Ethernet, WiFi, Mesh or USB Modem.

On smaller devices the upstream connection is limited to Ethernet due to memory constraints.

The attached screenshot shows the upstream connection via Mesh.
In the top section of the screen you can see two other mesh nodes, and in the bottom section, two client devices that are attached to the AP.

Using this sort of configuration you can easily set up standalone classroom micro-servers with a digital library on USB/SD, and have them connected wirelessly into a school campus mesh to get upstream network connectivity to additional servers or Internet.

Regards

Terry

[Andrés Pacheco]

Terry,

This is great! Better yet, is there a "package" with BoM of all required components, & base documentation / how to / project plans? That could be the basis for putting together grant proposals to set up such environments in schools that qualify!

--
You received this message because you are subscribed to the Google Groups "Village Telco Development Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to village-telco-...@googlegroups.com.
To post to this group, send email to village-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/CAD58Ekz9fE40O1ROUNVcNuXucPjEaq%2Bq0A9jn%3DDRhzW%2BLN%2BwfA%40mail.gmail.com.

For more options, visit https://groups.google.com/d/optout.

<20150921_110303.jpg>

<20150921_122538.jpg>

<VT-RACHEL-Test.png>

[T Gillett]

Hi Andres

There is some documentation on the VT Wiki page here:

         http://wiki.villagetelco.org/RACHEL_on_MP2


The implementation of VT-RACHEL is pretty simple. The typical steps are as follows:

1. Select Device

We have produced firmware for a number of devices including the VT MP02 and various TP-Link devices including MR3020, MR3040, WR842 and WDR3500/4300

The first step is to decide which device best suits your application.

- The MP02 is a good all-round solution, being simple and robust in construction. It has a single radio and no external antennas in its base form. It has more RAM memory (64MB) that most of the other devices and so has more caching capability. It runs from a 12V supply.

- The MR3020 and MR3040 are basic travel routers that run from a USB supply. The 3040 has its own battery that will power it for several hours.

- The WR842 is a more conventional router with external antennas.

- The WDR3500/4300 have more RAM memory (128MB) for caching and are dual band (2.4 and 5GHz) radio devices and so suitable for advanced set ups where you can run a backbone mesh on a different band to the local AP function.

The cost for these devices ranges from $25 to $75

2. Select Content
The next step is to determine what content you want to use and to prepare a suitable SD/USB memory device.

We have used 32GB devices to serve the standard RACHEL library including Wikipedia for Schools (in four languages ~7GB), Khan Academy (En ~13GB), Gutenberg Books, Medline etc.

If you just want to serve just WfS, you can use 8GB memory devices, or 16GB devices for just Khan Academy.

Khan Academy is also available in other languages including French Spanish and Portuguese. These three will fit on to a 32GB device.

Costs for memory devices are typically $5, $10, $20 for 8, 16 or 32GB devices.

Downloading all the content and loading it on to the memory devices can be challenging, and so if there is sufficient interest, we will make pre-loaded MP02 devices and SD/USB memory cards available from the factory through the VT on line shop.

3. Networking

Once you have installed the firmware on the router device and loaded the memory card, the system is ready to use as a stand-alone system in a classroom.

Simply connect the client devices to the VT-RACHEL device via wifi in the normal way, then start a browser and enter the URL:
           http://vtrachel

By default the Ethernet port is configured as a WAN port for connection to an upstream router for Internet access if required.

If you want to use mesh networking to provide a backbone network then it is a matter of planning the network and configuring the devices to suit.

This is generally pretty straightforward, but will obviously vary from case to case.

A simple hub-and-spoke network would have one central mesh node connected to the upstream router, then each of the satellite nodes just set to Mesh WAN mode in order to connect with the central node. The default settings will work out-of-the-box.

Regards

Terry

To view this discussion on the web visit https://groups.google.com/d/msgid/village-telco-dev/51B0AAE7-B2A4-4DA7-A90D-6D9F4B9B77BE%40gmail.com.