[kictanet] Fwd: [TIER] TV White Spaces. Was: Re: Economist: On tech and Africa

Thu Feb 21 08:18:01 EAT 2013

I've been quite intrigued to see this discussion taking place. It started
about technology in Africa and has now began to bring out the fact that
many Western countries are pumping significant amounts of money and effort
into research - with universities and private sector at the forefront.

I have for a long time been a proponent of industry-academia linkages
especially for purposes of ground-breaking research and development driven
by recognition of local problems and challenges. What I find interesting in
the thread below is how military application of technology is also a key
driver of technology development in those regions (and I am sure that there
is direct funding for this from military). With the levels of conflict and
military engagements in our region (and accompanying huge expenses on
foreign developed technology) wouldn't it make sense for us to develop
academia-industry-military linkages?

Just thinking,

Brian

---------- Forwarded message ----------
From: Amanda Pratt <pratt.amanda.l at gmail.com>
Date: Thu, Feb 21, 2013 at 1:54 AM
Subject: Re: [TIER] TV White Spaces. Was: Re: Economist: On tech and Africa
To: Alexander Chemeris <alexander.chemeris at gmail.com>
Cc: TIER <tier at tier.cs.berkeley.edu>

Hello Everyone,

This is mostly directed at Claro's comment:
"There is plenty of opportunity for innovation. As software radios
have matured, radio designers observe that software is optional, i.e.
one would build a highly configurable radio that could do anything a
software radio could do but based on dynamically configurable hardware
modules, instead. Say a silicon design that implements an RF engine in
the form of an ASIC. There are reasons to believe that such an ASIC
would simplify the path toward both trusted spectrum agile and
cognitive radios."

These IC radios are what I do research on here at Berekeley.  DARPA
has put forth money for many organizations to work towards such a
'universal' or 'cognitive' radio.  Together with Raytheon and Nokia,
we at the BWRC are looking into the problem.  There are a number of
technical difficulties in this space, given the broad number of
protocols and frequencies, but it is a hot area in integrated circuit
design right now and solutions are likely to be commercialized in the
next few years.

Right now, the work is focused mostly on consumer applications (i.e.
making your iPhone work both here and in Europe) and military
applications.  But, I personally am involved with the project because
of its applications to the developing world.  I would love to speak
with any of you in the space regarding the opportunities and
challenges for getting the last 3 billion on the WWW.  Please reach
out with any insights on the the market and demand for mobile,
obstacles to adaption/demand satisfaction, areas for innovations, and
other general implementation issues in both Africa and the rest of the
developing/rural markets.

Thanks!
Amanda

UC Berkeley Graduate Student Researcher
apratt at eecs.berkeley.edu

On Mon, Feb 18, 2013 at 10:58 AM, Alexander Chemeris
<alexander.chemeris at gmail.com> wrote:
> I agree with Steve. Working on OpenBTS I'm looking for time when we'll
> see GSMWS and LTEWS. I.e. when one could install a GSM or an LTE base
> station in any place where there is an available spectrum. At this
> moment explosion of mobile networks is held by this single thing
> called spectrum regulation, which prohibit you to deploy mobile
> services until you pay a lot of money upfront. I hope that TVWS and
> initiatives like the Netherlands' low-power unlicensed GSM will pave
> the way to a wider shift in spectrum regulation.
>
> On Mon, Feb 18, 2013 at 5:42 PM, Steve Song <steve at villagetelco.org>
wrote:
>> I think it is worth considering too that TVWS is as much a regulatory
>> innovation as a technological innovation.  Secondary spectrum use through
>> the use of a geo-location database and ultimately hopefully through
spectrum
>> sensing offers a real opportunity for regulators to by-pass the typically
>> glacial process of spectrum re-allocation and re-farming.  One need only
>> look at the digital switch-over process in Africa to see how that is a
>> reality.  Not to mention the risk reduction associated with secondary use
>> versus betting a big chunk of licensed spectrum.
>>
>> Cheers... Steve
>>
>> P.S. For interest, I attach a couple of charts on propagation from some
>> BBC-funded research.
>>
>>
>> On 15 February 2013 21:45, Ermanno Pietrosemoli <ermanno at ula.ve> wrote:
>>>
>>> The free space loss is 6 dB less at 450 MHz as compared to 900 MHz.
>>> However, this is if you can clear the first Fresnel zone, which, as
Yahel
>>> pointed out, requires more antenna elevation at 450 MHz.
>>> Besides, to obtain the same gain at 450 MHz the antenna should be twice
as
>>> big as at 900 MHz. So the real world advantage is kind of iffy with
regards
>>> to 900 MHz. But the latter frequency is not unlicensed in most of the
world
>>> (outside US and some Latin American Countries), sot the comparison that
>>> matters is with respect to 2.4 GHz.
>>> I agree with Yahel that for long distance point to point links 5 GHz is
>>> more cost effective, but stick to my point for BTS coverage to both
mobile
>>> and fixed subscribers.
>>> Regards,
>>> Ermanno
>>>
>>>
>>> On Fri, Feb 15, 2013 at 6:51 PM, Kurtis Heimerl <
kheimerl at cs.berkeley.edu>
>>> wrote:
>>>>
>>>> Can you tell me how much better the propagation is for a 450mhz BTS vs
a
>>>> 900mhz BTS?
>>>>
>>>> On Friday, February 15, 2013, Ermanno Pietrosemoli wrote:
>>>>>
>>>>> Hi Yahel and group,
>>>>>
>>>>> The lower frequencies have much better propagation characteristic.
While
>>>>> it is true that the Fresnel zone size increases with the square root
of the
>>>>> wavelength, the diffraction coverage also increases with wavelength,
and the
>>>>> absorption losses in walls and foliage are also less, that is why TV
bad
>>>>> devices are touted as NLOS. Of course, this might not be the case for
very
>>>>> long distance links, but it is very significant for a multipoint
coverage of
>>>>> a base statiion.
>>>>> These frequencies have been used in commercial deployments in several
>>>>> countries in what is known as CDMA 450 systems. The improved
propagation
>>>>> properties have indeed been proved, although the commercial succes
has been
>>>>> limited.
>>>>>
>>>>> Yahel is right that the antennas are considerably bigger, and have
lower
>>>>> gain at reasonable sizes, but there is a plethora of offers in this
band.
>>>>> Regards
>>>>> Ermanno
>>>>>
>>>>> On Fri, Feb 15, 2013 at 5:32 PM, Shaddi Hasan <shaddi at berkeley.edu>
>>>>> wrote:
>>>>>
>>>>> Great points Yahel. On the subscriber density issue, there's a lot of
>>>>> demand for capacity at lower frequencies, primarily due to their
ability to
>>>>> penetrate foliage. I think 900MHz unlicensed equipment has 28MHz of
>>>>> bandwidth available; I hear in practice this is about 10MB of
capacity per
>>>>> access point for modern equipment. TVWS would augment this existing
>>>>> equipment very nicely, assuming it ever becomes available at
cost-effective
>>>>> prices.
>>>>>
>>>>>
>>>>> On Fri, Feb 15, 2013 at 2:07 PM, Yahel Ben-David <yahel at airjaldi.org>
>>>>> wrote:
>>>>>
>>>>>
>>>>> TVWS might not be the panacea for rural broadband as commonly
>>>>> believed...
>>>>> At least not for fixed-wireless (stationary roof-top antennas to
replace
>>>>> non-existent DSL solutions).
>>>>>
>>>>> Where it comes to long-distance links, we're doing just fine with
>>>>> conventional WiFi:
>>>>>
>>>>> 1. We have enough bandwidth even if all links are 40Mhz-wide, since
>>>>> these are highly directional links that do not interfere with each
other,
>>>>> and ahh... rural typically means sparse - so bandwidth scarcity is
not a
>>>>> problem...
>>>>>
>>>>> 2. We won't enjoy the "better" propagation qualities of the lower
>>>>> frequencies - we already build high capacity links on 5Ghz for as
long as we
>>>>> can see... The only limitation is line-of-sight (LoS). In-fact, LoS
is not
>>>>> the accurate term here - the limiting factor is a clear Fresnel zone
and
>>>>> that's worse the longer the wave-length. To put simply, if to get a
70%
>>>>> clear Fresnel zone for a link on 5.8Ghz between two buildings at 50
miles
>>>>> distance you'll need tower heights of 25m (at both ends), that same
link on
>>>>> 700mhz would require tower heights of 72m. To be honest, you get more
energy
>>>>> through that 70% clearance on the 700Mhz than on the microwave link,
but
>>>>> only about 12% more - so in practice 12% less in the tower height for
the
>>>>> same link quality - that's 63 meter tower for the TVWS vs. 25 meter
for
>>>>> WiFi.
>>>>>
>>>>> * BTW, the above calculation ignores the curvature of the earth - the
>>>>> real towers needed for such a link are much higher, since the "earth
height"
>>>>> in the middle of that link is about 127 meters - far higher than the
towers
>>>>> in the example. However, we assume the use of mountains or hills for
such
>>>>> links and hence the above example simply show how much taller hill is
needed
>>>>> for 700Mhz vs. 5.8Ghz.
>>>>>
>>>>> 3. The antennas are huge - if I'm to build such a link (between two
>>>>> hills or from a high mountain to a far valley) I'll choose high-gain
>>>>> antennas. Say about 30+ dBi or higher. On 5.8Ghz, that's a small
low-cost
>>>>> antenna that I could easily mount on balconies' railings by myself.
For
>>>>> 700Mhz - ohh well, most balconies won't hold the size and most
railings
>>>>> won't hold the weight, not to mention the truck and crane needed to
haul and
>>>>> lift the thing... And.. don't forget wind-load and
maintenance...etc...
>>>>> True - 700Mhz might not need as high gain as it propagates better,
but only
>>>>> about 12% better...
>>>>>
>>>>> 4. Costs and availability - I guess that's obvious... but some may
argue
>>>>> this would change with increase in demand... I doubt that - for the
above
>>>>> reasons, there won't be any demand for fixed-wireless gear on these
bands...
>>>>> Moreover, the whole manufacturing cost structure is against these
bands, not
>>>>> to mention operational complexities due to the regulatory domain.
>>>>>
>>>>>
>>>>> So are TVWS good for nothing?!
>>>>>
>>>>> These bands are promising for rural mobile!
>>>>> They may offer the increase in subscribers density per base-station
that
>>>>> may drive operators to serve the more rural areas, given them a less
risky
>>>>> chance to ROI than with current technologies, not to mention the
reduced (or
>>>>> none) license fees.
>>>>> Given the super-low gain antennas that are used in mobile-phones,
their
>>>>> physical size increase for use of these bands would not be
significant.
>>>>> * Mobile phones must not use higher gain antennas - or else we'll have
>>>>> to hold them straight or point them in the direction of the tower.
>>>>>
>>>>> So yes - TVWS are great, but might not do what many of us think they
>>>>> will...
>>>>>
>>>>> BTW - TVWS are ill-suited for dense urban areas for their strong
>>>>> propagation qualities - in the cities, we strive to contain
>>>>>
>>>>> Ermanno Pietrosemoli
>>>>>
>>>>> Presidente
>>>>> Fundación Escuela Latinoamericana de Redes (EsLaRed)
>>>>> www.EsLaRed.org.ve
>>>
>>>
>>>
>>>
>>> --
>>> Ermanno Pietrosemoli
>>> Presidente
>>> Fundación Escuela Latinoamericana de Redes (EsLaRed)
>>> www.EsLaRed.org.ve
>>>
>>> _______________________________________________
>>> TIER mailing list
>>> Website: http://tier.cs.berkeley.edu
>>> TIER at tier.cs.berkeley.edu
>>> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tier
>>>
>>
>>
>>
>> --
>> Steve Song
>> +1 902 529 0046
>> +27 83 482 2088 (SMS only)
>> http://villagetelco.org
>>
>> _______________________________________________
>> TIER mailing list
>> Website: http://tier.cs.berkeley.edu
>> TIER at tier.cs.berkeley.edu
>> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tier
>>
>
>
>
> --
> Regards,
> Alexander Chemeris.
> CEO, Fairwaves LLC / ООО УмРадио
> http://fairwaves.ru
>
> _______________________________________________
> TIER mailing list
> Website: http://tier.cs.berkeley.edu
> TIER at tier.cs.berkeley.edu
> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tier

_______________________________________________
TIER mailing list
Website: http://tier.cs.berkeley.edu
TIER at tier.cs.berkeley.edu
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tier
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.kictanet.or.ke/pipermail/kictanet/attachments/20130221/f93a6fc0/attachment.htm>