It seems like ages ago but it was less than a year ago that Google sponsored a forum on TV White Spaces in Africa in Dakar. I wrote an article on the state of TV White Spaces in Africa based on that event. Last week I attended the Dynamic Spectrum Alliance Global Summit in Accra and that offered an ideal opportunity to reflect on what has happened in the last year and where things are going.
From White Spaces to Dynamic Spectrum
One of the first things you’ll notice in news about TV White Spaces (TVWS) is a shift in terminology to the more generic term dynamic spectrum. This is a recognition that the same approach that is being applied in TVWS projects could be applied to any frequency band. And indeed things are heading that way in the United States with the FCC’s emerging regulations on the use of 3.5GHz. So expect to see more use of “dynamic spectrum” in the future. And indeed the event had sessions on dynamic spectrum that went beyond television spectrum. Investor and philanthropist, Jim Forster, and researcher Shaddi Hasan spoke about the potential of applying dynamic spectrum approaches to GSM spectrum.
This last year also saw the formation of an industry body called the Dynamic Spectrum Alliance (DSA). This alliance was announced last year by Paul Mitchell of Microsoft at the Google TVWS Forum and came into being legally at the end of last year. It is in full swing now with members from around the world. The Accra event was their second global summit and the first event attended by its newly appointed Executive Director, Prof. H Nwana, former head of spectrum for OFCOM.
Facebook joins the party
Last year I marvelled at having both Google and Microsoft on the same stage advocating for a common outcome. This year was even better with Facebook joining the DSA and participating in a pilot with Microsoft in Ghana. To have these three giants, fierce competitors in any other market, standing together on dynamic spectrum is a powerful message. Microsoft is going gang-busters with dynamic spectrum pilots in Africa. They now have at least five pilots that are either active or in the works. Google has played things more cautiously. They have continued to support the Cape Town TVWS trial as well as playing an active role in the DSA. They don’t have much interest in supporting more pilots however; they are rather waiting for another significant point of leverage to move dynamic spectrum forward. Facebook, through its Internet.org initiative, is still mostly gathering information as it decides its strategy for improving access.
Pilots and Trials
The Google-sponsored TV White Spaces Trial in Cape Town is now complete and work is now underway to develop a regulatory framework for dynamic spectrum use in South Africa. The trial has generated a wealth of new data to better inform regulation. In particular, the trial established that TVWS radios could operate in television channels directly adjacent to television broadcast channels without interference. This is in contrast to the FCC regulations which insist on a gap channel between television broadcast and TVWS use. The trial also validated the importance of taking terrain into account when calculating the range and scope of available spectrum for a geolocation database. Integrating terrain modelling into spectrum propagation models gives a much more realistic picture of available spectrum and typically reveals more available spectrum. If the FCC were to consider adopting adjacent channel rules as well as terrain-based spectrum propagation modelling, there would be a lot more spectrum available for dynamic use in the U.S.
The South African regulator, ICASA, was represented at the event. They announced their timeline for producing TVWS regulation. It looks like the full process of public consultation to final publication of regulations will take until the end of 2015. They are considering taking a combined unlicensed and “protected mode” approach to regulation which would provide for WiFi-like use of TVWS spectrum and a quasi-licensed approach which would provide some protection from interference for “protected mode” users of TVWS spectrum. If South Africa goes ahead with this dual-approach to regulation of white spaces, this added complexity may turn out to be a disservice to other countries in Africa who may look to South Africa for a simple model they can emulate.
When questioned about a South African Department of Communications report on the Second Digital Dividend which appeared to indicate that the regulator would wait for the completion of the Digital Switchover before allowing any commercial use of TVWS spectrum, the representative for ICASA denied that this was the case.
Elsewhere in South Africa, the Microsoft-sponsored TVWS trial in Limpopo is underway but little was said about it at the event.
The Mawingu project in Kenya has been in full operation for some time now. After some early friction with the regulator and some partnership challenges, the project infrastructure is now well-established. Microsoft’s focus now is on working out what business models can best support these kinds of networks and towards that end they have contracted the University of Southampton to carry out research into the economic impact of Microsoft’s white spaces projects in Africa.
The TVWS pilot in Malawi which was undertaken by the University of Malawi and the Malawian regulator (MACRA) in partnership with the International Centre for Theoretical Physics (ICTP), has been in operation in September last year. Malawi was represented at the DSA event by Jonathan Pinifolo, Deputy Director of Spectrum Management for MACRA. Malawi appears to be on track to allow commercial use of TVWS spectrum by the end of June 2014. That would make it the first African country to regulate the commercial use of TVWS spectrum.
At the event last week, Microsoft and Spectralink announced a partnership to deliver connectivity to Koforidua Polytechnic and All Nations University College in eastern Ghana. Similar to the Mawingu project, they will combine TVWS backhaul with local WiFi infrastructure access for students on campus. SpectraLink will provide a managed connectivity service to the university including both local access and upstream connectivity.
Other pilots appear to be brewing locally with the Ghana-India Kofi Annan Centre of Excellence in ICT exploring a partnership with Nigerian company Wavetek to do a pilot. The University of Ghana Legon may also explore dynamic spectrum technologies from both a research and application perspective.
in May 2013, Microsoft announced a partnership with the Tanzania Commission for Science and Technology (COSTECH) and ISP, UhuruOne to carry out a pilot there but not a lot has been heard since the announcement and I didn’t meet anyone from Tanzania at the event.
While dynamic spectrum has been brewing for some time in Nigeria, not much seems to have happened yet. At the Google event in Dakar last year, Wavetek announced that they were the official distributors for Carlson Wireless in West Africa but to date no trials have taken place. Apparently a commission comprising the ministry of communications, the telecom regulator and the broadcast regulator has been convened to consider a trial but time will tell when that will move forward. Having separate regulation for broadcast and for telecoms is particularly inconvenient for TVWS spectrum which must embrace both.
This year Wavetek announced plans to establish a dynamic spectrum R&D centre at the University of Ilorin in Kwara State, Western Nigeria. They also have plans to establish a campus WiFi network there.
In yet another 4Afrika initiative, Microsoft are partnering with MyDigitalBridge Foundation in Namibia to deliver last-mile connectivity via TVWS technology. The pilot will be located in Omusati, Oshana, and Ohangwena, connecting 34 sites, most of which are schools. In something of a parallel with Kenya, the project enjoys strong support from the Namibian government but the Communications Regulatory Authority of Namibia (CRAN) appears to be reluctant to allow a trial to proceed. Hopefully the regulator will see the potential that this project represents and allow it to proceed soon.
Mozambique and Uganda
Although not formal pilots yet, I am aware of dynamic spectrum projects that are just getting off the ground in Mozambique at the University Eduardo Mondlane and in Uganda at Makarere University. I would be surprised if there were not more trials/pilots in the planning stage elsewhere across the continent.
Wireless Standards and Manufacturing
Last year it was completely unclear as to what might emerge as the dominant standard for TVWS spectrum. The two big competing standards were the 802.22 standard also known as Wi-Far which has evolved out of the WiMax standards group and the 802.11af standard also sometimes known as White-Fi which has evolved out of the WiFi standards group. The weightless standard, proposed by Neul is no longer a factor as Neul are no longer strategically focused on TVWS. For rural broadband, one might imagine that the 802.22 standard would have an edge being purpose-built for rural communications but it would appear that designing equipment to conform to the 802.22 standard is both expensive and difficult. As a result, I was unable to discover any companies currently manufacturing 802.22 equipment.
On the other hand, 802.11af appears to be on the cusp of a breakthrough. Mediatek, a semiconductor company that Wired magazine refers to at the “Taiwanese giant you’ve never heard of” is a founding member of the DSA. They are they third-largest manufacturer of WiFi chipsets in the world and they are also white-label manufacturers for a vast array of low-cost smartphones. They are on the verge of releasing a tri-band chipset in partnership with Aviacomm that would support traditional WiFi in the 2.4GHz and 5GHz range as well as TVWS using the 802.11af standard. The potential of this cannot be overstated. A low-cost integrated WiFi / TVWS platform at large manufacturing scale could open up a new platform for connectivity innovation. Apparently trials of this platform will be carried out in Glasgow in a few months with next year being a target for mass manufacturing. As the Accenture folk like to say, “now it gets interesting”.
Wireless standards like 802.11af will become more and more important as the market evolves. No one wants to be handcuffed to a single vendor over time and having a standard like 802.11af will mean that devices from any compliant manufacturer will behave well together.
So Where Can I Get My Kit?
As things currently stand, there are three companies with TVWS equipment deployed in Africa. Carlson Wireless with its RuralConnect platform is used in the Cape Town trial as well as the Malawi pilot. Microsoft has partnered with Adaptrum and 6Harmonics as the technology providers for their projects in Kenya, South Africa, Tanzania, Namibia, and Ghana. Carlson has recently withdrawn from the Dynamic Spectrum Alliance so I was unfortunately unable to get an update on their development plans. Adaptrum have released their ACRS 2.0 platform which offers a robust point-to-multipoint solution. At the summit last week they announced that a base station plus two client stations could be purchased for less than USD 5000. Carlson Wireless’s RuralConnect technology is in the same ballpark. The real buzz at the event though seemed to be around 6harmonics’ next generation technology which would be based on the new Mediatek chipset. According to Mediatek, with their new chipset, TVWS technology could even find its way into handsets within a very few years.
On Geolocation Databases
Geolocation databases remain a bit of a fuzzy problem for TVWS regulation in sub-Saharan Africa. For a country like Malawi, they are a bit like using a sledgehammer to tap in a nail. Malawi has vast tracts of unused spectrum in the UHF and could easily proceed with TVWS commercial licensing without requiring a full geolocation database support. They can build in provisions in the TVWS licenses to ensure that there is a process to evolve to geolocation authentication as the demand and market grows in Malawi.
For countries like South Africa that may be more ready to adopt a geolocation database-enabled approach to TVWS regulation, the challenge is rather what database to use and how to manage and pay for it. There are a number of geolocation database service providers in the United States including Google, Spectrum Bridge, and others. While I think Google would be happy to offer this service to the regulator in South Africa, the post Snowden-revelations climate has made many countries think twice about putting strategic assets in the cloud. It may take some time to re-build trust for this. Geolocation service providers are emerging elsewhere too. Finland’s FairSpectrum was represented at the DSA event and has one or more dynamic spectrum projects in South-East Asia. They were actively looking for African partners at the event.
It is also an open question as to who will operate and pay for a geolocation database service. In the United States, the emergence of a competitive market for geolocation services is predicated on the growth of a huge Internet of Things (IoT) market that would be the mainstay customers of such a service. In Africa where rural connectivity is likely to dominate as the primary use of TVWS spectrum in the near future, a geolocation database service would likely need to be subsidised by the regulator or the government.
Finally, there are also academic debates that continue as to what standards-based algorithms should be used for the terrain-based modelling that will drive the geolocation database. Hopefully geolocation databases will be implemented in such a way that terrain modelling standards can be applied and upgraded as both the math and processing power evolves.
The only real disappointment for me at what was otherwise a very exciting event was how the ITU representative at the DSA summit positioned the ITU with regard to dynamic spectrum. At the World Radio Congress in 2012, where collective agreement on future spectrum allocations are made among ITU member countries, it was formally decided that:
“the current international regulatory framework can accommodate cognitive radio systems, without being changed. The development of these systems, such as TV white spaces, is therefore in essence in the hands of national regulators in each country.”
This should be a clear message to any national communication regulator that they have a green light to develop TVWS regulation. And yet, that is not the message we heard from the ITU at the DSA summit. The ITU speaker communicated a sense of uncertainty and caution with regard to moving forward with TVWS regulation. He encouraged further consultation and engaging in ITU Radiocommunication Study Groups to establish best practices, all of which perhaps sounds reasonable. Unfortunately, this position completely ignores the fact that one of the key issues that dynamic spectrum regulation addresses is the slow pace of traditional spectrum management approaches. This is most evident in the fatally slow pace of the switchover to digital terrestrial broadcasting where it appears that 50 African countries will miss next year’s deadline. Dynamic spectrum is a low-risk approach to secondary spectrum use that doesn’t require a long wait for best practice. It can evolve dynamically as technology and demand changes. The ITU is needlessly spreading fear, uncertainty, and doubt about dynamic spectrum and it is undermining efforts in sub-Saharan Africa to get TVWS pilots off the ground. At the DSA Summit, it took persistent questioning from the audience to get the ITU representative to acknowledge that there was nothing stopping African regulators from proceeding with TVWS pilots and regulation.
On one level you could argue that not much has changed in the last year. There are some new pilots and new partners but no really big news like the first commercial regulation of TVWS in Africa. Beneath the surface, however, there was a sense of an impending revolution. The news from Mediatek and their partner Aviacomm may well mean the mainstream integration of TVWS into a huge range of devices within just a few years. If they go into mass production around an open standard like 802.11af, how long will it be before companies like Broadcom and Qualcomm/Atheros are obliged to get into the game? Having WiFi and TVWS in a single device is a sign of the future where you won’t care or know what spectrum you are using, your device will automatically optimise your connection across a range of frequencies based on local regulation.
Update 22 May 2014: I want to qualify my remarks about geolocation database services in the cloud. My comment about concerns about hosting strategic assets in the cloud might have been too much of a leap of intuition on my part. There was nothing publicly said about this at the summit.