CITP Lunch Seminar: David Johnson – Empowering and Connecting Africa with Community Owned Networks


[ Background Conversations ]>>Yes, I did. It’s recording.>>It’s recording. All right. Okay. Welcome to
another CITP lunch talk. David is, some of you
have already met him, is the visiting IT
policy fellow at CITP. He comes to us from the
University of Cape Town where he leads the Net4D lab which is Network
for Development lab. He has a Ph.D. in Computer
Science from UC Santa Barbara, and David’s research
largely focuses on developing decentralized
networks to enable communities and individuals how to
build their own services and infrastructure. Or, maybe own their own
services and infrastructure. David has requested that you
only ask clarifying questions during his presentation. He will leave time at the
end for other questions. So, with that, I’ll
hand it over to David.>>Thank you. Thanks, Tiffany. And, thanks CITP for, yeah,
for allowing me to talk about this topic which
I’m very passionate about. So, today, I’m going to talk
about community-owned networks, and the question is really around could this be a
connectivity solution for Africa. And, I’m going to
go through a lot of the building blocks required to build these community-owned
networks. My talk is going to
be a bit like a sort of water insect bouncing
over a pond. So, you know, you might, and get through this talk
a little bit dazed. So, let’s see how it goes. So, before we sort
of dive into looking at the connectivity problem
in Africa, it’s good to kind of step back and see what
does the world look like now. So, this is a picture of all
the cellular connectivity across the planet. And, you know, where it’s lit
up, there are cell towers. Currently, 88% of the world
is covered by 3G or 4G. This is according
to the GSMA in 2017. But, only 48% of the world
are actually connected to the internet. So, there’s this
interesting gap. Of course, babies
aren’t using Twitter, so we have to kind
of eliminate them. But, the rest of the adolescents
and adults are not connected for a multitude of reasons it’s
really linked to affordability of device and connectivity
and literacy and IT literacy. This is the reason there’s this
gap between the 48% and the 88%. If we looked to sub-Saharan
Africa, the picture is more bleak. So, we have 59% of
the continent covered which is actually not bad
considering it was pretty low in around 2000, 2001,
around 10%. But, only 21% are connected. So, there’s this even larger
gap, but again, we have to see that Africa has a growing and
swelling youth population. So, we have to eliminate, you
know, babies and kids up to age of four which is around 15%. But, it’s still a very
substantial gap, and again, it’s linked to affordability
of devices and connectivity and also especially in
Africa, there’s a literacy and IT literacy issue. It’s interesting
to look at Wi-Fi versus cellular connectivity
when you look at the planet. So, if I flick between these, you’ll notice Africa
gets darker, but Europe and the U.S. get
brighter with Wi-Fi. So, that just tells you that
there’s a lot more people who own personal Wi-Fi
routers and universities and public spaces
with Wi-Fi in Europe and the global north,
as we call it. So, the other interesting
thing is the growth of internet use year on year has
gone through a turning point. So, it reached a kind of
a knee around about 2012. So, we’ve connected the easy hub
of the world, and it’s now kind of starting to flatten off. And so, when you see that,
you realize, you know, the current models that we have
are starting to hit a ceiling. And, you see this also with
smartphone [inaudible]. So, if there’s less smartphones
being bought, you know, less people are using
this to connect. Of course, there are people sort
of replacing their old phones, and there’s some churn there, but when you start
seeing negative on your smartphone growth, there’s definitely
something going on. So, the question I want to
ask is will the next 50% of the world be connected
through equipment that looks more like
that or equipment that looks more like this? So, what you’re seeing over
there is my friend from Berkeley who worked with on the left
Indonesian community network that installed a cell
phone network in a tree. It’s a great idea
because as the tree grows, the signal gets better. [ Laughter ] On the right is a community
I’m working with in Cape Town that are powering their
Wi-Fi network with wind. And so. So, I want to step back
and, if you’ll bear with me, look at a possible way of
seeing what could happen with the community-owned network
movement in Africa and the rest of the world by looking
at what happened with electrification
of rural America. So, in the early 1930s, 90% of rural Americans
had no electricity. So, yeah, only 10%. It’s the same thing. Of rural Americans had access. And, at that stage,
the U.S. was 44% rural. Then, Roosevelt is
part of the New Deal in ’35 had an executive order to create the Rural
Electrification Administration. There was new regulation
to create the loans and provide engineering
support to these rural areas. Farmers in the community, then,
formed coops, paid their $5 to join, and the growth
of rural American coops for electrifying
rural America started. By 2010, 42% of the nation’s
distribution lines are owned by rural cooperatives in
the U.S. So, my question is if we look at rural
cooperatives in 1935, could we, and this is now the picture of
African cooperatives in 2017. Africa has about a
60% rural population. Not that different to the
U.S. in the early 1900s. Could this be a possible future? This is what rural
electric cooperatives looked like in 2019. Perhaps that’s what
cooperatives providing broadband in Africa could look
like in 2025. Interestingly, these rural
cooperatives are now rolling out broadband in the U.S.,
so there’s around about 100 at the moment rolling
out broadband. And, it makes sense because
they have all this cable infrastructure they can run
fiber on the poles and so on. So, we need to understand
the model of building community networks, and the model that’s often
used is this network’s commons license which is similar to the creative commons
license you’ll find for media and for software. So, it basically says you have
the right to join the network and the responsibility
to extend it. You have the right to
understand the network. Nobody can say, “I’m filtering
the traffic, but I’m not going to tell you what’s going on.” You have the right to spread
the knowledge about the network. You have the right to
offer services and content on the network, and the
freedom to use it as you wish without harming anyone else. So, that last clause
is something to protect from what’s called the
tragedy of the commons. So, when you have
a commons resource, you can have the situation
where it’s overconsumed and starts to hurt other people. For example, people might all
put up very powerful antennas and just drown everyone
else out. So, that’s a very
useful framework for building these networks. Another metaphor I like to, this is my own personal
metaphor I use to look at building networks
is to see it as building a house
on firm foundations. And, you have three axes. One axis is infrastructure. These are the radios
and the technology used to build the networks. Another axis is in the
wireless world is spectrum. If it was fiber, it
would be right of way, way leaves, and so on. And then, on the z axis,
you’ve got services. So, once you put
in infrastructure, you have spectrum, you
start building services. All of this sits on top of
the foundation of regulation, and nothing can be
done without skills. You need skills to
have good regulators. Regulators need to understand
how to create regulation that is in the public good and so on. And, you need skills to
build these networks. The problem is that the
regulation has favored the very large, big national operators. So, if you think of a house, they’re building ginormous
hotels and apartment blocks because they’re getting
a lot more spectrum, a lot more of the
valuable spectrum, and the small cooperatives and community-owned networks
have been getting tiny slices of spectrum. And so, they kind of like
can only build these small, little houses. So, I’m going to kind of
use that model as I work through the presentation. So, we also need to understand
how hard it is to build networks in some of these areas. So, if you look at a
typical rural area, you have very low population
density which is very difficult for somebody putting
up a tower that needs to get returned on investment. If your tower is only going
to reach 50 or 100 people. It’s a very harsh environment. There’s power instability,
lightening, dust. A lot of equipment fails. There’s a skills
shortage often of people to maintain infrastructure. As I said earlier, low
literacy and digital literacy. Often poor regulation or
misunderstood regulation, regulation that’s not clear. And, a lot of the
technology and services that are deployed there
are often not well suited for some of the above reasons. Interestingly, I’ve found many
African people perceive the internet as a cultural threat. So, for example, they might have
grown up building wire cars, and that was their culture. And now, their kids are
playing Flappy Bird. And they say, you know, the internet is destroying the
culture that we grew up with. Probably not a unique
problem to Africa, but definitely is
something that is mentioned. Just to show you how
insane the power issue is, I monitored power usage in a
Zambian village over two weeks, and I saw multiple power
failures in the day, some power failures lasting
as long as eight hours. I saw these brown outs where
the power would drop to 40% of its normal value and
spikes where it was 100% more than its standard value. Now, put most standard kind
of equipment that you use, like your Wi-Fi router in
a situation where that kind of power is happening,
and it usually blows up. So, you need to make sure that
the equipment suits these kind of harsh environments. One of the real challenges
in South Africa, although South Africa’s
pretty well covered by 3G. It’s about 98% covered. Is the cost to communicate. So, wealthy users have
these 24-month contracts, and low-income users have
prepaid packages, data bundles. So, and we find that a typical
low-income user will be buying lots of small packages as they
get income and could be paying up to $28 for a gigabyte. Whereas, low income users are
on 24-month contracts paying for ten or 20 cents a gigabyte. So, you’ve got this insane
difference in affordability between for access
between low income users and high-income users. I mean, this is, for
me, absolutely perverse, but operators get away with it. Many low-income users
I’ve spoken or surveys that we’ve done in rural South
Africa, they spend up to 20% of their income on connectivity. And, this is income that’s
often from social grants that the government give them. Right. So, on the, I just want
to look at these three axes now and just pick through them. So, the first axis
is infrastructure, the actual radios
that you deploy. So, traditional networks
effectively look like this where you have towers and
fiber providing internet to the towers. And, you have to invest
$50,000 to $100,000 to put up these towers. And, you know, it’ll create wide
coverage for a rural community, usually accessing
through their cell phone. And, measure adult
networks on the other hand, you will have smaller radios that are placed on
people’s roofs. And, the radios discover each
other and form links that build and mesh as you add
radios to the network. And, when you need to
get traffic from source to destination, it
essentially routes itself through the best route from
the source to the destination. Now, there are many
positives to this. So, it’s resilient to failure. If one of the nodes goes down,
it can find another route. It’s very simple to install. A lot of routers that I’ve
used put them on the roof, switch them on, give them
a name, and they work. It’s really that simple. You don’t need to, you know, do any network training
to set these up. They are pretty low cost. you can get a lot of mesh
routers for around about $50, and it aligns well with
the community network model where the community
owns the network. So, you own the router
that you put on your roof or the cooperative could own the
router that’s put on the roof. The negatives are that it
doesn’t have the efficiency of a centrally managed system. You don’t get quite the
performance that you would with a standard, say, cell
phone network with a tower. And, single radio meshes
have some scaling issues. So, I’ve done some
studies back in 2012. I bought this indoor lab with
I called it the graveyard. With all these wireless devices that we were routing
traffic between each other. And, it scales, in
my experience, somewhere near the throughput
of one half over n. So, after five hops, the throughput
would be around about 20% of the throughput
of the first hop. Gupta and Kumar have done
some really cool studies and shown it’s not
quite as bad as that. But, that’s roughly the
type of scaling I’ve seen. There are ways to fix this, so
you can use jewel radio mesh, and all you do is you add
two radios to your device. So, for the first hop,
you might use one channel. And then, the next hop
uses a different channel. It’s a different radio. And, the advantage of that
is you can have full duplex between whereas the packet
arrives on one radio, it can be sent out the other
radio on a different channel. And, doing this, you get around
about a 5% drop of throughput, and that’s mostly due to
buffering in the device. So, I went back to the first
mesh that I installed in 2007 where I partnered
with the community in a rural area of South Africa. And, I thought I’d just show
this one because it’s kind of an interesting historic mesh in that it was the first
rural mesh in Africa. It was around about ten nodes. Internet was coming into
a clinic, and we needed to distribute this
to some homes, to a school, to some farms. Now, the interesting thing
is the satellite link was ridiculously slow, 512 kbps. And, you only had two
gigabytes per month. That was the cap. The clinic was only
using, amazingly, about 300 or 400 megs per month to
send medical information. And, they always had
the spare capacity at the end of the month. So, we went, let’s share the
spare capacity with a bunch of users in this rural village because it’s actually
forfeited every month. They just lose it. And, we set a 50-megabyte
cap per user. You would wonder what you could
do with that, but back then, people could browse
quite a few webpages. And, we were able to provide
average sort of throughput of about 2.3 megabits
per second to users. Now, they would never
feel the effect of the slower satellite link because even though there
was this scaling problem where they were getting slower
and slower per each hop, because the satellite link
was so slow and the Wi-Fi was so much faster, they
never actually felt that scaling problem. So, it was, we didn’t even
need jewel radio mesh. The single radio
mesh is quite ample. We were able to do great
things like run voice over IPs on this network and
save the clinic calls between their remote
site around $300 a month and provide internet
access to many users who hadn’t ever seen
the internet. The device that really
changed everything was in 2006. This little box. The Linksys WRT54G. I’m sure many of you had them. This was the first router that
had enough memory and flash that you could put
embedded [inaudible]. And, that opened up an
enormous world of being able to install mesh routing
firmware for $40. What we did is we stripped
the electronics out, put it in water cliff,
waterproof enclosures like you see over there. Stuck it on the outside
of a house and were able to do four miles with
a decent antenna. Even with an old can that
we used as an antenna, we were able to get three miles. A can was great because, you
know, that’s basically free, and we saved about $40
or $50 on that antenna. Since then, many
community networks that use mesh have appeared. Some of the famous
ones in Berlin, there’s a mesh that’s still
running about 380 nodes. Guifi net started off as a mesh. It’s now 35,000 and
upwards nodes. It’s more of a kind
of mesh topologically, but because it got so
huge, they’re starting to do BGP routing between nodes and use standard internet
routing protocols. This is in Barcelona. Yeah. The interesting thing is
once your mesh starts getting large, you begin to create
clusters of meshes and route with super nodes
between these clusters because of the scaling issue. Right here nearby, there’s
the New York City mesh, and that’s an interesting mesh in that they’ve had
[inaudible] donated, and you pay a voluntary
fee to connect. So, I think the minimum is $20. If you want to pay more you can, and you get unkept
internet in New York. And, it’s 401 nodes
at the moment. In South Africa, there’s
a 65-node rural mesh that’s [inaudible]. I’ve put links there if you want
to read more about the networks. So, the next axis I want
to discuss is spectrum. Specifically, spectrum sharing
which is a new technique to use spectrum more
effectively. So, it’s useful to kind
of understand how we got to this concept of
exclusive use spectrum. So, back in sort
of the early 1900s when Marconi created an arc and
saw an electron move in a piece of wire, just a couple of miles
away, that became, you know, Marconi Wireless Telegraphy. He actually said he wants to own
all the spectrum in the world because he needs it for
his equipment to work. Very glad that didn’t happen. But, and then, after
that, in 1906, they created the
radio treaty which was to protect spectrum
from interference. Actually, to protect messages
to ships from interference from hackers on the coast. And, they started
looking at spectrum from a country perspective and protecting it
against interference. If you look at the evolution
of how it was used all the way to today, you know, we’ve got
cellular appearing in 1983. The first time we thought about
spectrum based on rules but not on exclusive use was really when we had license
exempt spectrum in 1989. That became what
we used for Wi-Fi. In 2008, the FCC made
a bold decision to look at dynamic spectrum access and created rules
for TV white space. But, effectively, most of the
spectrum is exclusive use. You allocate it across the
whole country, and you lock it down for a specific
purpose or operator. But, you know, that thinking is
based on sort of 1900s thinking where we had valves and
very noisy electronics. And, you couldn’t reuse that
spectrum in other areas. Things have moved on. We’ve moved to digital. We are now, we have
software defined radios, cognitive radios, and we have
the ability to reuse spectrum. So, we can use some spectrum
in another part of the country where it’s not being used. For example, in rural areas, it’s not being used
in open areas. So, we need to move from
this concept of right to exclusive use that we
have now with operators to this concept where
they have the right to protection from interference. But, this is going
to be a long road but a battle we’re fighting. It’s also useful to look
at the Wi-Fi success story. So, in 2017, the global
picture of Wi-fi was, for smartphones was 54% of internet traffic
was being offloaded to Wi-Fi for smartphones. Cisco says by 2022 that’ll
climb to close to 60%. Now, what’s really
remarkable about this is that Wi-Fi has only
had 80 megahertz versus cellular bands
having 270 megahertz. And, they’ve achieved that
kind of level of traffic with only 80 megahertz
of bandwidth. The reason is that Wi-Fi
creates very small cells. So, you can reuse spectrum
if your cell size is smaller. So, we understand
why it happens, but it’s still incredible
that the technology that basically works
by saying shut up if I’m talking
has achieved that. So, I like to use the road
analogy for spectrum when we try to understand how
spectrum is used. If you think of bands
as lanes in a highway. In urban areas, the
spectrum is fairly well used. Cellular is mostly used. There’s quite a few
channels used for broadcast, some military channels possibly. But, if you got to rural
areas, you’ll find most of these highway
lanes are empty. The problem is our current
regulation says you have to stay out of all of these lanes even if those lanes are
actually available. So, the operators
will always talk about there’s a spectrum crunch. We’re running out of spectrum. The reality is there might
be some spectrum crunch in urban areas, but there’s no
spectrum crunch in rural areas. So, this concept of there being
a spectrum shortage is actually an artifact of the fact
that we’re still stuck with exclusive use
spectrum licensing. So, TV white space, as I said,
the FCC approved that in 2008, is effectively a concept
that says in blocks where spectrum isn’t being used
by digital or in South Africa, we still have some
analog TV signals. You have this spectrum
called white space. They should never have
called it white space, especially in South
Africa with our history. Should be called grey space,
but or something like that. But, anyway. We didn’t, we weren’t there
when these decisions were made. Interview used to
have old televisions. Well, no when you had a
TV that had snow on it. That doesn’t happen anymore. But, when you would click
through the channels and there was snow, that
was effectively white space. So, what’s so great
about white space? Well, if you had a rural
area and you wanted to distribute a signal
to all your users, if you took something
like 2.4 gigahertz Wi-Fi with an omni antenna to other
omni antennas on a roof, you might get about
two kilometers. If you use 600-megahertz
TV white space, you would get eight kilometers. So, and, you also get better
penetration through foliage and through buildings. If you look at it on a sort
of basic physics level, you have about 16 times
more area that you can get from TV white space compared
to 2.4 gigahertz Wi-Fi. If you look in South Africa at how much TV white space
there is, blue being sort of all TV white space, red
being some fairly limited number of channels. So, red, where you
see red there, it’s effectively
our urban areas. What I’ve found is if
you move into rural areas or areas just outside the
city, you get between 70% and 90% TV white
space availability. In the cities, it’s around
about between 30% and 50%. Now, the great thing is that if
you look at where are the areas that are well suited
for TV white space, and how many people live there. Doing some studies on what
area population densities that are well suited to
less than one gigahertz. I’ve found that 42% of the population are
living in those prime areas. So, that’s a great number
of people to connect where there’s currently either
only expense of cellular or very poor coverage. I did some studies
comparing South Africa to the U.S. Having studies in
California, I decided I’m going to drive around a lot with a
spectrum analyzer on my roof. And, what I found
is that in Fresno, if I can call that urban. [laughs] Nobody wants to go to
Fresno except people like me who want to study
TV white space. There is very little
TV white space. Maybe 20% channel availability. Whereas, in the middle
of sort of the valley in rural California, there’s
around about 50% availability. In South Africa, urban areas
look like rural America. They run about 50% availability
in an urban area like Victoria. In a rural area, sometimes
there is almost 100% TV white space availability. What you see there is just sort of some other interesting
information about the U.S. cleared their
TV band in 2009 for digital TV. So, TV used to go up to
850, and that was cleared for cellular operators and
sold off to AT&T and Verizon. South Africa actually
still busy clearing that. Our TV goes up to 900 megahertz,
and we are still in the process of auctioning that spectrum off. So, if you now look at this mix
of TV white space and Wi-Fi, what can you do now that
you have TV white space? Well, there’s some links that
might have trees in the way or be just out of line
of sight with Wi-Fi, but TV white space you
can now build links where there’s vegetation
in the way or where you’re just
out of line of sight. So, the ultimate device
is a device that has both. So, develop this hybrid
Wi-Fi TV white space box that has both a TV white
space in it and a Wi-Fi radio, and if it finds a link where
it’s can’t reach the other end with Wi-Fi because
there’s foliage, it’ll use the TV
white space link so, the thing to remember is that Wi-Fi generally will
give you more capacity if you can see the
other Wi-Fi radio. Whereas, TV white space
will give you lower capacity but better coverage. So, having both, you have
the best of both worlds.>>Clarifying question. Because we’re limited
to clarifying questions.>>Yeah.>>What protocol are you
running over TV white space?>>So, this is 802.11. This particular device I used
was just down converted Wi-Fi into 700-megahertz,
600 megahertz.>>[inaudible] do much
better if you were to use, like [inaudible].>>Yeah. There’s some native
ones like 82.11 AF and so on, but those radios were
way too expensive. We have this problem
with TV white space. It hasn’t reached
scale of economy yet. So, I just used a simple
down converted Wi-Fi radio. But, the routine protocols
I used ran over both radios. So, another exciting development
is that open cellular, this is now with
Facebook are working on what I would call
white space cellular. So, they’ve developed this
$1500 LTE based station which you can deploy
in a community network, and it will use either
unallocated blocks of spectrum by the government, or it can
use, similar to TV white space, spectrum that isn’t being used
in a certain geographical area. This is going to be a
hard fight because we have to somehow ask the
operators if they’re prepared to share the spectrum. And, there’s a long
regulation battle ahead to, before we actually get there. I’ll actually be in
Amsterdam tomorrow at the Telecom Infra Project
Summit to ask the operators if they’ll be kind and start
sharing their spectrum. The last axis I want
to discuss is services. So, once you have
some infrastructure, what kind of interesting
services can you deploy? I’ve done some studies, and
this is in a Zambian village to understand how do
people communicate and specifically how do they
share messages or content with each other when they
have a tool like Facebook. So, we studied Facebook
traffic in the Zambian village over a number of months. This study, we had 14,000 or
so instant message messages between users or
users in the village and users outside the village that they were communicating
with. And, we found that around about
if you look at that picture, the dark, the red lines are
instances where there is a lot of traffic between
users and the village. And, we found that up to 60% of
the instant message traffic was between users in
the same village. So, they were sending messages
all the way to California over a satellite link and
coming all the way back into the village. Which is insane. And, we found the same
thing for sharing images. So, the point is does that look
like a design for a locality of interest that we saw
in the Zambian village? We also did some questionnaires
with people in South Africa and Zambia to find out how
they’re sharing images, video, and audio. And, we found that nobody
was using online services to share video. They were sharing video
on USB sticks and sitting around a cell phone in the grass
watching each other’s videos. Which makes sense, because
the cost is exorbitant. So, to illustrate this, let’s
say a local artist wanted to upload a hip-hop
video in the village. In South Africa, their HD
video would cost them about $4 for three minutes to upload. All their fans, each time
they were trying to access that hip hop video would be
paying between $1 and $4, depending on the quality. Can’t you host that content
much closer to the user? So, we built a solution. This was back in 2012. We created a system called
village share, and we thought about it and went,
well, we don’t want to rebuild the social network. Everybody’s already got the
social network on Facebook. The problem is, the content
is living in the wrong place. So, with this app, you would,
say, you know, share your music or your video with your friend. It would, then, know that that
friend lives in the village and place the content on a
local server in the village. So, you’re still using the
low bandwidth messaging part of Facebook internet, using the
normal international servers, but putting the high bandwidth
content in the local village. Yeah.>>So, Facebook supported that? [inaudible]>>We used the Facebook API,
and we just built an app. There was one interesting trick. We had to do an DNS override to
place the content in the, well, to access it in the
local village. The problem with this, this
was great, but when people ran out of phone time, they were
firewalled from their content. So, it wasn’t yet
an ideal solution. We currently have a
project called [inaudible] where we’re now looking at
deploying lots of analogs to internet services
in a local context. So, we are adding things like
Diaspora instead of Facebook, Nextcloud instead of
Dropbox, Rocket.Chat, well, not instead of, but
with WhatsApp. So, these are all services you
can deploy on a local server, and they’re something
like a docket container. And, we also creating platforms where you can have VM hosting
locally using something like eucalyptus instead
of using Amazon EC2. And, the great thing is
these are open platforms that you can innovate on. We’re not saying this is
how the service should look. We want people to innovate
and create their own content, create their own services. When we work with
these communities, we want to be coresearchers
with them, design the solution with them. This is part of the Orlando
[inaudible] research type of thinking. The key things we want to amplify the locality
of interest. So, when one looks at the global
internet today, we kind of move from our thoughts
straight onto the internet. Twitter is probably
the worst culprit. As you think, your thoughts go
onto, you know, Facebook is fine and global googled them. The way I’d like to see these
services look is more like this. We form an idea in our thoughts,
and we chew through them in the privacy of our minds. Then, we begin to share them
with our local communities. And then, some of
these ideas might move from the local communities
to the global internet. Some of them might stay. They might have a perfect,
happy resting place in your local community. So, we want to create an onion
layer for these to live in. So, just to conclude some of the current work we’re
doing is all of the above. And, we are deploying or working
with the community network to deploy in an area called
Oceanview where there’s around about 60% unemployment, a
major issue with affordability. I’m looking at an
interesting model where a [inaudible] is rolling
out a lot of Wi-Fi hotspots. We want to look at adding
individual ownership to that where if you buy a
router and you put it up, you can be rewarded
for routing traffic. And, we would like
to see if that helps to incentivize expansion
of the network. And, we’re using
block chain for that. So, this is a project
I’m currently working on over the next
couple of months. So, the way forward,
really, is we’re looking for intention amplifiers. The intention of a
community network is to close the connectivity gap,
close the affordability gap. We can, we need networking
infrastructure that’s simple to install. We need regulation
and technology to support spectrum sharing. We can use these micro cloud
localized service platforms to encourage local
content creation. And, we need to have skills
development through workshops, training programs, so that
people can build these kind of services and infrastructure. So, my question is
could we use something like the universal
service and access funds. In America, it’s called USF. As a sort of a New Deal for building these cooperative
based community networks. And, can operators
maybe be incentivized to share the spectrum by
maybe getting a discount on their USF fees. They have to currently pay
between 1% and 3% USF fees. Would that be enough for
them to be a bit kind? I will say, there was a dark or realistic side we
have to also look at. Local communities and
cooperative can go two ways. They can either flourish or
they can be relational breakdown in these communities,
and they can collapse. So, in a sense, they’re also
more, they’re less stable than a commercial network which
is coming from the outside, and there’s no particular
local interest or knowledge of local power structures. And, the next billion connected
users are probably mostly going to be using this
for entertainment or doing something we don’t
yet know, something unexpected. And, you know, they might not
be watching Khan Academy videos or watching the crop
prices, which NGOs and a lot of researchers like to have
pictures of a Kenyan farmer with his phone checking
crop prices. Maybe there’s a kid playing,
you know, Flappy Bird, and or playing online
games with his friends. That’s also fine. Whatever happens is fine, but it
needs to come from the community and not be pushed on the
community from the outside. Last metaphor. This is a communication
telecommunications chart. If we fill it up with just
the Verizons and the Comcasts, these are the big rocks. We will never fill the jar. We need these cooperative
community-owned networks, the small rocks, and perhaps
add individual ownership which could be the sand. Thank you. [ Applause ] All right.>>Questions?>>Ten minutes.>>So, in one of the
last slides on USF fees.>>Yeah.>>I think in the United States, those are passed on
to the customers. So, whether they’re
reduced or not sort of doesn’t have any
impact on the telecom.>>Right.>>[inaudible] incentive
for them to do anything one
way or another. Is it different in South
Africa or [inaudible]?>>I guess I wouldn’t say that, but that’s probably
what they’re doing.>>I mean, it’s like
a line item, like, on your phone bill [inaudible].>>Right. But, if
the fees were reduced because they share spectrum
and the fees were kept the same for the users, they would be
pretty pissed off if the knew that that was happening. Yeah.>>I was just wondering how you
approach communities [inaudible] these kind of networks in. You sort of have to talk
to federal officials, local officials,
and in terms of, I like the comparison
to the New Deal. The big difference seems to be that when you have
federal funds, that makes things pretty easy across the entire
U.S. territory.>>Right.>>Is that a concern for you, how to fund these
kind of networks?>>Yeah. So, the first stage
is obviously, you know, communities need to show their
intention to do something. So, we typically
work with communities where they were saying we’d
really like to do this, and as a, well, as a researcher
organization like UCT, we’ll come alongside
you and help you. We can get grants through
the university to help. So, we’ve said we will help
you for one year, but we would like to see you be
sustainable after that one year. Which is why they have
to charge for the access. They’re charging $1 a gigabyte,
and if they’re scaled enough, they can be sustainable. I want to go beyond that and
say we could use USF fees to also seed building
towers, getting equipment with the same model
instead of university now, it’ll be USF fees to just
give people a leg up. Kind of like the rural
carpenters got a leg up from Roosevelt
to just at least get through the first
hurdle, the first year. Yep.>>I’m from the part of the
world that was really changed by the REA, Texas there’s so
many remote communities there. The 20th century didn’t arrive
in some places until about 1935.>>Yeah.>>My feeling about this
is that I’m absolutely fine if it’s the blue guys
because the experience of Texas, big cooperative, big.>>The blue guys.>>Yeah, community-owned
networks.>>Yeah.>>[inaudible] without
Roosevelt, there would have been, the farmer would have been
consigned to darkness.>>Yeah.>>Not just entertainment,
although movies and the radio were
a huge part of it.>>Yeah.>>It was washing
machines and all the things that come with society.>>Yeah.>>Which was a huge [inaudible]. In my view, without
government intervention, nobody’s going to risk this.>>Right.>>And.>>I agree.>>Yeah, that’s basically it.>>A fair point. Yeah. Thanks. Yeah.>>I guess my question is
still the same that I asked when you first presented
this over [inaudible]. You can do that with, you know, the equipment you have
at the university.>>Yeah.>>But, you need this
might be problem. So, I think the comparison
[inaudible] but, the washing machine was same
thing in city as in rural area. And, you know, [inaudible]. In this case.>>I would say the endpoint
is the devices are the same, but getting the electricity
was pretty different in rural areas and urban areas. That is different.>>[inaudible] even in the ’60s, a lot of the [inaudible]
was [inaudible].>>Sure, but sort of from the
user standpoint, [inaudible].>>Sure. Yeah.>>Here, like, image
you do, you succeed. Will it create down
the line problems if there are just two
different sort of levels of technological
development, and this one sort of inhibits what you can do
using this technology even if you have affordable access.>>So, are you, I’m just trying
to understand what you’re say. That technologically this is
more complex than electricity, so the analogy maybe
begins to break down.>>Can I try to?>>Maybe. I didn’t quite.>>I’m going to try
to [inaudible] and then slightly [inaudible]. So, I think he’s saying
that you needed the fact that everybody was buying
the same washing machines from Sears.>>Right.>>And then, somehow, you approximated the
power requirements of those washing machines. As I said, [inaudible]
in both the, and again, I’m older whatever
[inaudible] in cities as well as the [inaudible].>>Yeah.>>So, you had to have
sufficiently robust.>>Right.>>Washing machines that they
could survive [inaudible].>>Which is more
expensive, yeah.>>And so, I think,
well, I don’t know that they were [inaudible].>>Yeah.>>It just was a set of
constraints, and I would imagine that there may be something
similar with respect to the axis that the high-tech stuff we have
here [inaudible] these networks but would probably also
be at risk in the cities. So, you might just have a lot
more robust, high tech stuff. Like, using different
kinds of phones or something, but [inaudible].>>Right. Yeah. Yeah. Interesting. I mean, sometimes, actually,
the innovation that goes into making the stuff
more robust in rural areas improves
the ecosystem for everyone. And, there’s enormous amount of
scale if you can get that scale in these rural areas on
Africa and 60% is rural.>>Is that [inaudible]
working on the cheap stuff.>>Yeah.>>And then, other people
pick it up [inaudible] scale.>>Yeah. [ Inaudible Comment ] Right. Yeah. Yeah. [ Inaudible Comment ] Yes. Interesting. Google Learn. [ Inaudible Comment ] Yeah. So, on the
infrastructure side, I think it was an
abysmal failure. I mean, you know, the gliders, they basically given
up with that project. Google Learn is a kind
of a fun, you know, Google X project
that, it’s cool. But, it’s not something
I would say, you know, I would take seriously
for a village. In terms of first Facebook Zero,
that was a failure in India. Yes, they have a little bit, but
it also hasn’t really taken off. So, you get this Facebook
Zero where they strip out all the images and
they give you a very low bandwidth Facebook. And, we find most people
just didn’t like it, and they wanted the full
experience and were prepared to, rather than just run out
of data and [inaudible]. It’s fascinating. I would say that the
most interesting thing for me is Starlink by Elon
Musk with the satellite. Starlink. So, he’s deploying
a low earth orbit satellites across the whole planet to. [ Inaudible Comment ] I have no idea who, no. I mean, he would just say I’m
trying to [inaudible] the data. But, the idea is that you’re
going to get full coverage with low, the latency has
always been the same problem with satellites. But, he’s going to get
less than fiber latency, he claims, for undersea cables. But, these very low
earth orbit satellites, and I don’t know how he’s going
to get the terminal cost down. But, time will tell. But, if we have these,
perhaps these are simply going to be gateways into these
cooperative community networks in some of these remote
regions where there’s no fiber. Or, if the terminal price
comes really far down, maybe everyone’s just going to have a Starlink
box in ten years.>>You sort of brought
this in at the very end, but I think it’s
[inaudible] important, this idea of the localization.>>Yeah.>>And, some people, some of the big [inaudible]
are paying some kind of lip service to
that [inaudible]. But, I’m just wondering because
a lot of people are trying to figure out what’s going
to happen with these things. Again, [inaudible]. But, they know that
there’s a trust issue. They know that [inaudible]
American or Canary Island or whatever company’s issues. So, I wonder are they
willing, I don’t know if you would trust them anyway,
but are they, is anybody talking about [inaudible] support
you with this kind of process and keep the data truly local into your communities
and whatever. Or, is that something maybe
you think is [inaudible]?>>Yeah, I haven’t had a
lot of people interested in keeping the data local. [ Inaudible Comment ] I mean, data is really
the new uranium. Not gold. But, there’s
definitely been support from infrastructure. Yes, I mean, Facebook
wants to look at building low cost cellular
base stations and that sort of thing, but on the
data side, it’s quiet. [ Inaudible Comment ] Okay. Thanks. Yeah.>>Is there room for
standards or open source types of standards to [inaudible]?>>Absolutely. Yes. So, I mean, all the
standards that we are developing for one example would be, so
we have an [inaudible] standard to access white space. So, right now, TV white
space is something called the [inaudible] standard. It’s an IETF standard. We want to move that into other
spectrum like cellular spectrum. So, we’re working on
something called [inaudible] which is a standard to
access cellular spectrum, white space spectrum. And, we absolutely believe
in building open standards so that anyone can
innovate solutions with a common standard, yeah.>>I was interested. I’m not sure if I
fully understand, but you said you were going
to a conference to talk to the providers about being
nicer about sharing spectrum.>>Right.>>So, when you’re talking
about that, are you talking about Comcast and those
kind of providers, or who are you talking
to and how to you try to sell it to them? Like, why they should
do this beyond out of the goodness
of their heart?>>Yeah, so we’ll be talking
to, so it’s in Europe, so there’ll be Vodaphone
and Orange and those kind of operators. It’s an interesting thing. So, they’re interested in this because they realize they
are reaching saturation. So, they are looking
for solutions that can provide much lower or
can create a low-cost solutions in some of these rural areas. Like a $1500 LTE base station
is very interesting to them because they’re investing,
you know, $50,000 to put up
a base station. So, they’re interested
in this project on this low-cost device side. On the spectrum side,
I’m kind of saying, well, there are these areas
you don’t want to go. What do you have to lose
by sharing a spectrum with a database and maybe get
some discount on USF fees. We’ll see where that goes. But, you’re not going
there anyway. The database is built in such
a way that if you do go there, you just switched that channel
off, and they can take over. But, you’re not actually
losing customers because, especially if it’s an area
where you’re not even going. That’s kind of my
line of argument. But, yeah.>>Just following up on that, I’m wondering what
you’re [inaudible] role of the regulators in
the local countries. Do they have, in these places, they tend to be semi-government
organizations.>>Yes.>>They tend to have a lot of
control over what’s happening in the airways in
their countries. Is there [inaudible]?>>Yeah.>>And, opportunity for
opening up the space and compelling the
operators to do so? And also.>>Yeah.>>Part of that, they
generally oversee the USFs which I’ve seen is basically
invested in laying a lot of fiber and [inaudible].>>Yes.>>Is there, I haven’t
really looked closely. Is there any effort to deploy
these kinds of solutions through those [inaudible]?>>Yeah.>>Mentioned that
earlier, but [inaudible].>>Right. So, in South Africa,
we’ve worked very closely with the regulator, and they’ve
actually been very innovative and allowed, so we were
one of the first countries to have TV white space. All the operators and the
broadcasters pushed against it. We went to the regulator,
and the regulator said, “No, this is a good thing. We don’t care what
the operators say. This is in the public good. Do it.” So, it was definitely
the right strategy to either go to the regulators when it
comes to opening spectrum. It’s a weird dance between the
regulators and the operators. The operators are
extremely powerful and have very powerful sort
of lobby groups that come to Parliament and sort of show
the government how wonderful they are and how they’ve, you
know, connected so many people. So, you know, we often need to
be a voice of reason coming in and making sure we balance this
megaphone on the operators. But, definitely the regulators
are the right place to go, and there’s a group
called [inaudible] which is a southern African
group of that looks after sort of all the regulators, always
a sort of advisory body. We looking at approaching
them to sort of create policy for the whole region
rather than trying to go to every single regulator
in every country. On the USF fees, it’s
new territory for me, so we’ll see where that goes. But.>>All right. It’s time. David. [ Applause ]>>I guess I should stop this.

Leave a Reply

Your email address will not be published. Required fields are marked *