Shared Spectrum Framework for Community Networks for Kenya online discussion

Thanks Barrack and Mwendwa for your excellent comments.

I believe extending power for Wi-Fi will lead to more interference, not less, and that the problem isn\’t so much the lack of spectrum (Wi-Fi already has more than IMT yet IMT is fulfilling 90 or 95% of the internet currently I guess), but the way Wi-Fi works through using unlicensed spectrum. Letting the CNs use IMT would be better, whether they are licensed directly or lease from an existing licensee who\’s not using their spectrum in that location.

I believe the CA\’s framework does suggest exploring that option and I\’d recommend some pilots in these various options to see what\’s technically feasible as well as whether existing licensees are willing to enter into these agreements for reasonable prices (if they\’re not using/planning to use the spectrum in that area, then any revenue for them could be better than nothing, even if little, though CA could mandate it to be made available for a very small admin fee, I.e.g. Use-it-or-lose it concept) or make some specific spectrum available for CNs to use, e.g. 700MHz that would be really good for fixed wireless access in rural areas and for which there\’s some unused spectrum currently but which currently requires 2.5bn to get…

Out of interest I am not sure if Mawingu counts as a CN since its quite big now, but the friend I\’m staying with at the moment pays them 5k a month and a 15k set-up fee for the router and installation for 5mbps, whereas Safaricom charges the same for same speed using 4G (basically unlimited data, though after 400GB they\’ll reduce to 1mbps) with router about 10k according to their website. So in this case the pricing is basically the same (obviously depending on actual coverage being available or not).

Interesting to note that Safaricom fiber is cheaper than 4G but of course availability is limited to some urban areas.

M: +254-790985886
Deputy CEO, Government Affairs
Huawei Kenya
From:Mwendwa Kivuva <>
To:Adam Lane <>
Cc:KICTAnet ICT Policy Discussions <>;Heiko Rehm <>

Subject:Re: [kictanet] Shared Spectrum Framework for Community Networks for Kenya online discussion

On Fri, 28 May 2021 at 18:52, Adam Lane <<>> wrote:
Hi Mwendwa

Everything has gone silent, even though this is a controversial topic!

Probably it is a hard topic, and of course, it has a level of political weight behind it.

Maybe I can stir up some debate…

Firstly, it would be helpful to clarify what the problem is before trying to identify the solution. For example, is it a lack of spectrum or is it the cost of spectrum or is it something else? Actually there is quite a lot of spectrum available in Kenya that is completely un-used (for example, in 700 Mhz band that is good for rural areas) and some that is so under-used (by some ISPs and some government users) that the best thing to do would be to return it to the CA and let the CA license it to someone else to use. [In relation to the Spectrum Fees Topic] Certainly the cost of that spectrum is a challenge for many, and it is very commendable that the CA proposes to Review spectrum fee framework recognising the need for significantly reduced fees for underserved/rural areas. I fully support this and it should be applicable to ALL spectrum license holders to encourage as many as possible to connect more unconnected areas. Fees can help work out who to allocate spectrum to that is serious about using it, but it also adds major costs. There are multiple fees at the moment, not just for spectrum licenses but also per base station using it, and even for using spectrum for backhaul to base stations.

Since CA licensing is unified, from the webinar we had on this topic, they had indicated they would do a review as you have proposed for the entire sector.
The challenges around spectrum are many, for example
a) CA charges an annual registration fee for 5GHz Wifi for Point to Point connections. PtP/PtMP are important for community networks
b) Current equipment used for WiFi has less range and strength because of the low permitted radiated power rating. Increasing this will make the WiFI equipment more reliable especially in rural areas.
c) If the range of frequency in the license-exempt WiFi spectrum is extended, it would reduce the current congestion where communities are complaining that their connections are not reliable.
d) We need to build the technical capacity of operators and businesses to enable them to conform to regulatory requirements. As Barrack has indicated in his reply, \”Spectrum has been a very opaque subject to many.\”
e. As you have pointed out, there is a lot of allocated spectrum that is being hoarded. Spectrum refarming can be employed to govern the repurposing of spectrum bands to more efficient technologies, investments, or new services where the spectrum is needed most

Secondly, it is important to understand that connectivity relies on infrastructure investments, the business case for which need some predictability. One cannot build a network (whether a community network or a “regular” ISP) without some guarantee that the infrastructure will still be usable for 5 years or 10 years etc.

Probably this is where innovation and new ways of doing things should be allowed to thrive? There are many institutions ready to fund innovative connectivity solutions for the common good because it is in innovation, trial and error … where new things come into being. I wonder is community networks operators can tell us what is the minimum requirements to setup a small network for a typical Kenyan village.

Also, it is critical to avoid different providers using the same spectrum too close to each other that it causes interference. These kind of issues are what makes shared spectrum tricky. When does one decide the spectrum is not being used in a location and let someone else use it compared to recognizing the license holder just has not yet got around to building in that location but will do “soon”?
One of the recommendations for the shared spectrum framework is to expand the frequency on license exempt spectrum, and also increase the power of equipment used. Does this sound as a viable solution?

The two proposed solutions above are:
a) Adjustment to PtP and PtMP EIRP levels could dramatically increase the potential of WiFi equipment in 2.4 GHz and 5 GHz to deliver access in rural areas. See Appendix 2 for specific recommendations.
b) extending the range of frequencies available for license-exempt use. This would have the impact of reducing congestion in backhaul connections by increasing the range of license-exempt frequencies;

Leasing spectrum from a license holder could be a viable option (i.e. that license holder agrees to let someone else the their spectrum in a particular location for a fee and for a determined time). [In relation to IMT spectrum topic]

Interesting solution. I am not sure what the rest feel. There is one school of thought that things this leads to hoarding a public utility with the expectation of exiting the investment at a profit.

In relation to this a critical issue with investment in infrastructure and having predictability is the issue of wanting to change a spectrum usage after equipment has been invested in. Let me give 2 example:

1) TV White Space regulations have been available in some countries (e.g. US, UK) for many years but have had very little adoption. Meanwhile some of those frequencies have been used for regular mobile use, e.g. 600 Mhz network in the US and achieved wide scale and are particularly useful for rural coverage. What if equipment is deployed to use TVWS in this frequency but gets little adoption so the CA wants to change to mobile use; what to do with the existing equipment? How to avoid interference with the new equipment? [In relation to dynamic spectrum access topic]. Since there is plenty of spectrum available in Kenya, just some is not well used and could be taken back/re-distributed and some is too expensive, there may not be a huge need for TVWS, but if ISPs can get it to work with the geolocation databases, and if they can get good enough Quality and Speeds, they could try.

+1 on spectrum refarming.

As you have indicated, if anyone wants to experiment in investing in TVWS, we should give them the opportunity.

The Tier1 operators in Kenya are not deficient in spectrum. What the proposal is trying to do is give the small guy in the village access to usable spectrum to serve themselves, since the big boys are not serving them at the moment.

2) Use of equipment in unlicensed spectrum – once it is in use in the market then it cannot be taken out of the market and new equipment brought in. With licensed spectrum it is easier to manage. So for example as countries weight up spectrum for wi-fi vs 5G, if later there is more demand for 5G, it will not be possible to remove wi-fi equipment since there is no record of who owns it.

Third, it is important to recognize the need and benefits from economies of scale. Nationwide providers get this from non-infrastructure based operations (e.g. creating an organization with administration functions, customer service functions, core network, billing systems etc) and they also get this from having common infrastructure. This is a challenge for localized spectrum access [in regards the IMT spectrum topic].

There is a community network in Easter DRC – the Idjwi island, an underserved area with a population of three hundred thousand people on Lake Kivu – called Pamoja Net with IMT spectrum. There was a blog by Nzambi Kakusu on the KICTANet website <>

This is a viable option especially in areas prone to terrorism – community control gives it so many advantages.

[on the license exempt spectrum] It is clear that in Kenya the majority of people use mobile for their access (from a base station that is using either microwave or fiber) or they are using wi-fi for their access (from a mobile router or a home fiber router). Most ISPs that provide wi-fi networks also use fiber as their backhaul. In the future more and more people may be able to use both mobile routers (5G capacity) or home fiber routers (fiber). In fact there is need for more spectrum for mobile, as more and more people will use 4G and 5G mobile routers to give them wi-fi at home especially in rural or less dense places where fiber may be expensive. In these cases there is no need for more spectrum for wi-fi. Wi-fi provides only short-range internet and is easily blocked by walls, and with 7+ Gbps capacity that is more than enough for the small number of users for each access point (It can get 7Gbps because it already has 560 Mhz of spectrum, way more than mobile) whereas each mobile base station will support thousands of users. Mobile base stations cannot work with unlicensed spectrum.

I think the bigger point is – how do we bring the guys doing KCPE and KCSE in year 2021 access to online learning tools? (usecase). How do we help the farmers tackle locusts in June 2021?

What I am trying to say is, the bigger operators will one day take the internet to everybody. But if I am in the village, I also want the same opportunities for my child now, for my farm now, for my economic empowerment now. Life is for the moment.

Do you think those people, no matter how few, may benefit from better WiFi? So that when they pull backhaul from their provider, it is less a hustle serving their communities leggaly through PtP/PtMP.

Also [on license exempt spectrum], it is noted that in the CA document that there is frequent interference for those using it for backhaul. This is an inevitable problem with unlicensed spectrum which causes quality issues, making it unsuitable for large scale deployments of backhaul in dense populations (with other users of wi-fi for personal use as well as use for backhaul intefering).

True. The congestion is a major issue. You find a small area with hundreds of equipment. But then that shows there is a need. Hunger. Increasing the range and power may help those who can afford more powerful equipment serve their communities better.

I know this can get quite technical, but I hope that some of this information may be useful. Note some comments I have made are simplified/generalized to try to be brief and not too technical.

Thanks so much Adam for the healthy debate. I guess their are two side to this coin. And it is true that this topic is part of the deep end because it is very technical.


From: kictanet [>] On Behalf Of Mwendwa Kivuva via kictanet
Sent: Friday, May 28, 2021 12:20 PM
To: Adam Lane <<>>
Cc: Mwendwa Kivuva <<>>
Subject: [kictanet] Shared Spectrum Framework for Community Networks for Kenya online discussion

Dear Listers,

The tread on Licensing framework is on fire and doing very well. Thank you all for the contributions, reading, learning, relearning … Please continue debating on that thread.

We will start a new thread on the proposed \”Shared Spectrum Framework\”. The Licensing and Shared Spectrum Framework for Community Networks for Kenya that was issued by the Communications Authority of Kenya, available for direct download here<>.

License Exempt Spectrum

WiFi has emerged as a powerful technology for both access and backhaul around the world but regulation has not fully kept up with the backhaul developments.

The draft framework recommends:

• Review the Guidelines on the use of Radiofrequency Spectrum by Short Range Devices to amend EIRP limits for 2.4 GHz & 5 GHz Wi-Fi for Point-to-Point and Point-to-Multipoint use.

• Review options for lowering the barrier to use of other license-exempt bands for PtP and PtMP use, including 24 GHz and 60 GHz.

• Expand the range of frequencies available for license-exempt use, especially in the 5 & 6 GHz range

• To strengthen collaborations with service providers to foster standards and regulatory inclusion.

Dynamic Spectrum Access

• To expedite the commercial availability of geolocation database service and implement required mechanisms to make the TVWS spectrum available immediately to operators.

• To establish an incubatory period for TVWS technologies.

• To evaluate with regional regulators the feasibility of a common approach implementation of geolocation databases

IMT Spectrum

• Establish a regulatory sandbox for localised spectrum access for small operators in an unassigned LTE band.

• Conduct a review of international approaches to the creation of more localised access to spectrum to inform the establishment of a more permanent mechanism for local spectrum access that is well adapted to the Kenyan context.

Spectrum Fees

• Review spectrum fee framework recognising the need for significantly reduced fees for underserved/rural areas.

• Consider a spectrum fee reduction scheme for non-profit community networks.


• What are your comments on the proposed recommendations?

• What in your opinion are the most important considerations the proposed shared spectrum framework should address?

• Are there gaps you have identified in the existing licensing framework in respect to spectrum assignment and utilization?

• How would you recommend CA to address the identified gaps?
Mwendwa Kivuva, Nairobi, Kenya

KICTANet mailing list