Africa Digital transformation 2020-2030 - short term, Long-term implementation Scenarios, an alternative view

 Executive Summary

Sam Atuhamya is a young, energetic casual laborer on a small four acre banana plantation just 25km outside of Kampala, the capital city of Uganda. He earns about USD40 a month. Sam has a small 2G phone he uses to communicate regularly with his family, a wife and 2 school going children, about 400km away in the south west of Uganda. They have two lactating cows and regularly earn a small monthly income, about USD30 to supplement Sam's’ monthly income.

The family therefore earns about USD70 monthly part of which is used to pay school fees for the children.

Sam has no electricity at his house, even though the grid is about 4km away from his workstation. So Sam has to carry his phone, always walking past the overhead electricity lines to a nearby trading centre to charge. It is a similar situation for the Wife, with the Grid several Kilometers away.

Sam’s household represents the broad spectrum of families in Africa, from Madagascar, Nigeria, South Sudan or Kenya, low income generating, largely living in rural areas with scanty, if any, access to utilities like electricity, water and telecommunications.

It represents the focus of the broadband and digital transformation agenda of the World Bank Moonshot for Africa Report 2019, Dead spots or areas outside of broadband coverage – near towns or inland in ultra rural areas, with a population whose revenue potential isn’t attractive to established telecom operators to fast track investment.

Objective

The key objective of this paper is to look at ways of speeding up implementation of broadband roll-outs with the forecast investment expenditure to achieve the World Bank 2021 and 2030 digital transformation milestones.

Some Definitions

 To help us understand the end game towards World Banks’ digitalization agenda, we extra some definitions for context in the table below;

Broadband	Defined as a download rate of 3mbps(2021 target) and 10mbps(2030 target) over a technology neutral connection
Unserved, Unconnected Areas	Areas with no network coverage, with no incentive for traditional telco investment, or areas with traditional 2G and require 3G /4G or equivalent technology upgrade to support broadband connections
Budget (2021 target)	USD9bn investment forecast to Double broadband connections from 2016 numbers, and 220m new users online
Budget (2030 target)	USD100bn investment forecast to connect 1.1billion users online, with 250,000KM of fibre & 400,000  4G/5G and other technology Base stations

A snapshot of Africa Telco Roll-outs

Telecom companies will always follow the money. This means that their business plans, their investments will be tied to areas with proven, demonstrable economic activity and sizeable population density. You will therefore always find fiber or transmission roll-outs following highways, site roll-outs targeting areas with sizeable population numbers fitting into their Average Revenue per user (ARPU) agenda.

It’s therefore inevitable that some areas inland are locked out of Broadband network coverage. For digitization to happen in such areas, in line with WB 2030 agenda, Development Partners, country regulatory bodies, Integration Partners must work together to find innovative, sustainable and cost-effective broadband solutions to fill this Gap.

Community Set up

Communities in Africa will vary in size, but in order to address the Broadband gap, it is important to understand their various set-ups to provide information on population density and possible economic activities in those areas.

In Uganda for example, you have the smallest administrative unit, the Local Council 1(LC1) with a population of about 250 -1000 People

In Madagascar you have the Communes made up of several Fokontany as the smallest administrative unit, with varying population density based on proximity to the Ocean and/or inland near small towns/communities.

For purposes of this paper we will assume a rural population of anywhere between 50-300 people, under the knowledge that Telco players would in any case have economic motivation to invest in areas with populations exceeding these numbers.

The Electricity Conundrum

Most rural communities in Africa lack basic utilities like water and electricity. In order for the Broadband value chain to be complete, we must address the issue of electricity. This is vital in charging phones and powering base Stations, microwave or satellite links that make up the broadband network.

The cost of extending grid distribution lines is prohibitive (cost per km anywhere between 20-30,000USD in Uganda) and in fact in some cases the rural area just provides a transit/Wayleaves access route to the next town, without necessarily being served.

In places like the South of Madagascar, not only is the road network, crucial in speeding up broadband installations, especially impassable in Rainy seasons, you will find that outside of Fort Dauphan (Tolagnaro) towards Amboasary, Ambovombe, Tsihombe, and Beloha, the sign of electricity is only found within the community town, being run on diesel generators with no sign of Gridlines in between, even though along the route are several small rural communities.

Indeed, 2018 data from one operator there shows just over 50% of sites running on diesel generators, which means high Operational expenditure, translating into higher Broadband prices.

Solution?

We can circumvent this challenge by taking advantage of our natural energy endowment by harnessing the power of Solar Energy.

Solar Photovoltaic (PV) intensity or irradiation in specifically Sub-Saharan Africa (SSA) is such that there are enough hours of sunshine, an average of 4-5Hours, to generate enough energy to power broadband sites and community internet access centers. See sample figure below;

Luckily, the cost of solar equipment today is also cheaper, thanks to regulatory and Government initiatives like some tax exemptions e.g. here in Uganda.
Therefore, in understanding precise power requirements of broadband base stations, we can size, cost and implement an effective and practical solar power solution that delivers affordable, quality broadband.

So whats the average power requirement for a remote rural site?

The assumption is that we want to cover remote rural areas, off-grid, sparsely populated and out of reach for traditional telecoms.

We must dimension a solution that balances Cost, Coverage & quality requirements, meets today’s expectation (3mbps baseline) but also provide a smooth transition into future technology upgrades. The latter is achieved by using Software Defined Radio (SDR) Units that can be software upgradable without change of Hardware.

 The table below shows power consumption by BTSes from some leading equipment vendors

Vendor	Technology	Configuration	Typical Power Consumption(W)	Microwave(W)	Site Load(W)
Nokia(flexiModule BTS)	GSM +WCDMA	S6+U_4	790	100	890
ZTE(ZXSDR BS8906)	GSM +WCDMA	G_S4+U_S1, DC	465	100	565
	GSM +WCDMA	G_S444+U_S111	815	100	915

Table 1 shows the average power consumption for a loaded BTS, including microwave or VSAT transmission equipment.

For the remote rural site, consumption will be way much lower than in the above table because the configuration will be smaller say 1 transceiver unit (TRX) and 1 UMTS carrier. For this paper, we will assume a realistic maximum power Consumption of 400W. With the Solution set up and Maintenance free, this adds to reducing the end cost of broadband bundles for the users, ultimately leading to more connections.

Deployment Scenarios

Connectivity of Base stations in rural areas to the super information highways of the Internet can only be possible with a baseline of a solid transmission backbone.  The ultimate end goal is therefore to have fibre connectivity as near to remote rural areas as possible. However for the case of Sub-Saharan Africa, the nearest point of Presence for fibre is realistically far off, in some cases more than the 25km mentioned in the moonlight light report. Therefore it’s important to devise alternative means of backhauling the rural site traffic. Most deployment scenarios will follow a monolithic and/or distributed architecture framework as shown below;

The baseband unit, BBU can be as far as 40km away from the site in the distributed architecture. In the monolithic architecture you have the BBU and Radio units collocated at the remote rural site location. 

The Ideal Power System

A standalone solar system, with a 24Hour autonomy and as seen in Figure 2 below completes the site configuration, with the number/quantity of panels, battery and charge controllers dependent solely on site load.

 

Rural Tower Design and Site Lay-out

Normal telecom towers take a month or more to erect and commission depending on factors such as weight/ tower loading, civil works completion, route/road access etc. For remote rural sites, It is necessary to dimension a tower that’s low cost, can be installed and commissioned within a day or two to facilitate fast turnaround times for broadband coverage and online access by rural population.

In General, an ideal tower for remote rural broadband coverage should be anywhere between 10-20m, closer to the population, with antennae that meet minimum electromagnetic radiation safety standards as shown in the Figure below;

Security for the remote rural sites to avoid vandalism and theft can be handled at a community level, which adds to reduction in overhead costs.

 Other Access Technology Deployment Considerations

2G, 3G, 4G access technologies assume a lion of sight environment to provide adequate broadband coverage. Alongside this is the need to limit interference to acceptable levels. Therefore cell radius or coverage area will in most cases be limited to 100’s of meters to a few kilometers. With this reality in mind, the number of base stations required will be higher which increases the cumulative cost of deployment. This hence presents a case for looking at alternative technology deployments to achieve the same end goal.

TV White Space Technology (TVWS)

This is a new technology that takes advantage of idle spectrum, the guard bands, within the UHF and VHF bands for Television broadcast to deliver broadband connectivity.

 It works very well in non-line of sight conditions and its coverage area can therefore stretch to up to 10km, with distances of up to 50km in Line of sight conditions and therefore fewer BTSes. The Gen3 Rural Connect Base Station from Carlson Wireless provides these and other features; with plug and play installation times meaning broadband can be delivered to schools, health centres, in remote areas quicker.

 Trials of this technology to deliver broadband have been conducted by Carlson Wireless & Neul, with support by ICASA  successfully in Cape Town South Africa, with Download speeds up to 12Mbps and upload speeds of up to 4Mbps over a 6month trial period attained.

It’s therefore prudent that for the broadband agenda 2030 to be achieved TVWS must be considered for point to point and point to multipoint roll-out in various communities.

Wi-Fi

This is another alternative technology, based on the IEEE802.11 (xx) Standards that can provide broadband in the unlicensed 2.4GHz and 5GHz bands with coverage radius upwards of a kilometer. Advancements in the IEEE802.11xx technology standards mean carrier grade Wi-Fi can provide stable connections for some rural deployments.

A Quantitative analysis of forecast network deployments and what it means for Africa

The World Bank moonlight report forecasts deployment of four hundred thousand (400,000) 4G and 5G base stations as well as two hundred and Fifty thousand (250000) Kilometers of Fiber

Based on the above figures, a linear analysis of what this means in the context of Africa is shown in the table below, where we have assumed a ratio of 5%:95% investment allocation over the 11 year cycle (2019-2030) respectively between Northern Africa and Sub-Saharan Africa. This assumption is based on the facts below;

• a)     Northern Africa broadband coverage is higher compared to that in Sub-Saharan Africa, with only about 2% out of broadband network coverage according to the moonlight report
• b)     There are fewer countries, only 5, in northern Africa as compared to Sub-Saharan Africa

Total Forecast Investment
Fibre(KM)	250000
Base stations	400000
Region	Northern Africa	SubSaharan Africa
Countries	5	49
% Investment allocations	5%	95%
Categorized Investment breakdown(Regions)
Total Fibre(Km)	12500	237500
Total Base station	20000	380000
Per Country Deployment Forecast
Fibre(Km)	2500	4847
Base stations	4000	7755
Annual Deployment Forecast(Per Country)
Fibre(Km)	227	441
Base stations(BST)	364	705

Table 2 shows that to achieve the 2021 and 2030 targets,

·         Each country in Northern Africa needs to deploy, consistently, 227KM of Fiber annually (19KM monthly),and 364 BST annually( 30 BST monthly)

·         Each country in Sub-Saharan Africa needs to deploy, consistently, 441KM of Fiber annually(about 37KM monthly), and 705 BST annually or( about 59 BST monthly)

From this information we can deduce that applying Country specific coefficients will determine how many more (or less) allocations in terms of length of fiber or number and category of base stations per country per Year. Also at a country level, it is necessary to use a targeted approach in expanding broadband by looking to cover first areas with significant economic potential.

This linear analysis clearly shows that the length of fiber seems small compared to the number of users forecast to be online by 2030. To achieve true broadband, we argue that there needs to be more investment in Fibre / transmission network footprint down to community level.

But how long does it take to do a Kilometer of Fiber?

Execution timelines for deploying fiber underground, overhead or in a mixed configuration is a critical determinant in achieving the 2021/2030 broadband Coverage agenda as it provides the baseline for interconnecting base stations to the internet to facilitate broadband access. The major activities for underground fiber deployment, as an example are below;

No.	Field Teams	Team Composition	Fibre Activities	Day 1	Day 2	Day 3
1	Team 1	100	Route Trenching
2	Team 1		Duct Installation
3	Team 2	20	Backfilling & Reinstatements
4	Team 2		Compaction
5	Team 3	10	Manhole Installation
6	Team 4	10	Fibre Hauling in Duct
7	Team 5		Fibre Splicing & Test

An aggressive activity schedule to deploy 1KM of fiber includes mainly 7 activities each following the other, with an average daily output of 10m per rigour in the trenching team. We also assume that prior activities like pre-casting of Manholes, procurement of fiber & ducts happened previously.  Further, that there are fewer reinstatements due to tarmac demolition as the physical routes are towards rural areas and mainly earth/loam soil.

We can conclude from the above that,

• ·         A minimum of three (3) days and 140 rigors in the field are required to complete a 1KM fibre stretch end to end.
• ·         Multiple teams and multiple installers for each activity are needed to fast track implementation and realize the 2021/2030 broadband agenda by doing most of the activities in parallel
• ·         Technology must be used to fast track deployments

Synergies with other Public Infrastructure projects, and the budget saving

Governments normally engage in big infrastructure projects, with funding from development partners like the World Bank. In Africa it’s mainly in areas of;

a)     Rural electrification projects(REP)

b)     Road construction/ Highway projects(RCHP)

We can extend fiber closer to the communities and fast track the 2030 broadband agenda by including overhead fiber runs (in case of REP) and Utility ducts and fiber (in case of RCHP) within the project scope as opposed to doing them independently.

Rural Electrification Agency(REA)	Medium Voltage Lines (Km)	Low Voltage Lines(KM)	Total(KM)	Source
	10,000	7,000	17000	REA Website
Uganda National Roads Authority(UNRA)	Paved Road(KM)	Gravel Road(KM)
	5,000	20,000	25000	Unverified
		Total	42,000.00

Above is a table that shows the mileage in terms of rural electricity and roads in Uganda, representing a missed and also available opportunity to extend Broadband to nearby communities. The REA electricity network has particularly reached 109 of 122 district headquarters, meaning all these can easily become the fibre Points of Presence from community networks.

Network Opex Vs High Availability 

Item	Budget  2021($bn)	% of 2021 Allocation	Budget 2030($bn)	% of 2030 Allocation
Target Broadband Users(millions)	220		1100
Focus Area
ICT skills & Dev't	1.7	4.05	18	17.5
Policy & Regulation	0.5	1.19	2.4	2.33
Network Operations & Maintenance	2	4.76	53	51.5
Infrastructure Capex	5	11.90	29.5	28.67
Total	9.2		102.9
Cost Per Online User connection	42		93.5

The above table shows budget allocations as forecast from the World Bank moonlight report to achieve the broadband milestones. What seems to stand out is the high operational cost (opex) allocation at 51.5%, perhaps out of the need to run off diesel generators to power broadband sites, staff salaries and the like.

To achieve high availability and significant reduction in maintenance costs from site outages and the like, we argue that an increased investment in the site Solar power systems and transmission/fiber networks redundancy will go a long way in achieving this milestone.

By increasing days of autonomy and number of storage batteries per remote rural site, and an increased investment in self healing/redundant fibre and/or transmission network rings at targeted aggregation and access network levels, we automatically increase site reliability and availability. Several scenarios on how to achieve both physical/route and logical redundancy are available but outside scope of this paper.

As well, as noted previously above, we need to increase fiber footprint closer to communities, in a range that accommodates high speed, small aperture, short-haul last mile microwave connections to backhaul traffic to nearest fibre point of presence and guarantee user end to end broadband speeds as defined by the moonlight report.

The budget for this investment can be a reallocation from the current opex budget, a reduction by say 10% and this reassigned into the Infrastructure capex budget for the same period. The outlook changes as displayed in the Table Below;

Item	Budget  2021($bn)	% of 2021	Budget 2030($bn)	% of 2030
Target Broadband Users(millions)	220		1100
Focus Area
ICT skills & Dev't	1.7	18.48	18	17.49
Policy & Regulation	0.5	5.43	2.4	2.33
Network Operations & Maintenance	2	21.74	42.72	41.51
Infrastructure Capex	5	54.35	39.78	38.67
Total Budget Forecasts	9.2		102.9

Affordability of Broadband today

A monthly bundle of 125 voice minutes of use and 1GB of internet data, including Over the Top (OTT) Tax(in the case of Uganda), costs about Uganda Shillings 10,000 on the MTN Uganda network. This represents about 6.8% of Sams’ monthly salary. This is still higher than the recommended 2% or less by the Broadband commission in the moonlight report.

In Madagascar, a monthly voice bundle, with 3hours 45minutes costs 10,000Ariary and a 2Giga bundle costs 30000 Ariary, a monthly total of 40,000Ariary on the Orange network. There is no equivalent 1G bundle to compare. These price plans are even more expensive than in Uganda.

Therefore, in this digitization agenda, Regulatory bodies need to enforce and industry players implement special, innovative rural-based broadband price plans with In-bundle rates that attract, and not deter users in this space. For the same user experience, Sams’ Wife should be able to pay far less than a normal broadband subscriber in an uptown neighborhood.

Industry players can implement these specific price plans by using location based services feature, that tags service and charging plans to a location and is now available in most core network deployments for 3G and 4G networks of leading vendors. Also, deployment of such would be similar to say night shift bundles offered at a cheap in-bundle rate to keep the network active with users online say after midnight.

But more importantly, prohibitive regulatory policies, like introduction of Tax on over the Top Services, like is the case in Uganda, and the recent passing of new licensing regimes that ideally constrain smaller players in investing regionally(via exorbitant licence fees etc) makes rural business models commercially unsustainable in the face of dominant industry market players.