This year, the IGF Multistakeholder Advisory Group which provide assistance in the preparations for Global IGF meetings called for Intersessional work (activities that are pursued in the months between annual IGFs with the aim of helping the IGF produce more tangible outputs that can become robust resources). Previously, the IGF has used best Practice Forums and Dynamic coalitions to bring out key issues that affect the world as it relates to the Internet. This year’s Intersessional activity is centred on “Policy Option for connecting the Next Billion”. This activity might not have come at a better time when just about 4.1 billion of the world’s 7.3 billion people are without Internet access¹ and the UN just released a list of 17 SDG for the world to ponder for the next 15 years. The IGF secretariat provided a portal where anybody can give feedback on the policy options for connecting the next billion², and they received feedback until the Brazil meeting to be held from 11-15 November 2015 in Joao Pessoa.

Over the years, a lot has been said on how to provide affordable broadband to the masses, especially in the so called developing world. Arguments have been made around many policy options including; IPv6 uptake, rollout of IXPs, functional ccTLDs, robust local content and hosting, affordable and available power, rollout of fibre optic, infrastructure sharing, Universal Service Funds (USFs), security, ensuring human rights are observed online, literacy, to political goodwill among others.

Internet access is still a mirage to most in the developing world, let alone broadband. A lot of statistics are thrown around on the number of users with access to Internet. For Example Communications Authority of Kenya (CA) pegs the numbers at 14 million Internet subscribers and 22 million internet users against a population of 41 million. This translates to 34% internet subscribers and 54% internet users over an African average of 19% Internet users according to ITU. The Kenyan numbers are pegged on increased use of the internet in accessing basic services such as mobile banking, health services and education, the availability of affordable gadgets such as smart phones and tablets, as well as the increased affordability of internet due to special offers and affordable bundled internet provided by the operators³. Unless Kenya is very special, comparing those statistics alone of the Kenyan and African average will tell you there is something amiss. Either ITU or Kenya’s CA is wrong. First, we need proper context on what access to internet means, and how useful that internet is. For example, how reliable is internet on a smartphone that has only 2G enabled, with unreliable signal, and billing done per data usage costing on average 50% of $1,587, the GDP per capita for Kenya? It therefore goes that proper statistics and triangulation by different independent research organisations is important before we can formulate policy options for bridging the legendary digital divide. Without proper information it’s hard to plan. In this article, I will give two examples on how innovative initiatives enable access to the masses.

The case for last mile infrastructure without breaking the bank.

One initiative changing lives in Gulu, Northern Uganda christened “Zoom Wireless” is a good model for expansion in underserved areas of Africa, where traditional ISPs and mobile operators fear to tread, especially most parts of Sub-Sahara Africa. This project has been running for slightly over 2 years championed by Brian Munyao Longwe, a renowned Internet pioneer in Kenya. Its products range from a symmetrical 512Kbit/s link to 9Mbit/s with the top-end speeds costing about $190/month which Brian says are competitive enough except compared to slower mobile network bundles. The total cost of the Gulu roll-out was $45,000 for three points of presence, including all of the equipment to run the operation. $25,000 of this total went into the solar backup to provide 36 hours of uninterrupted service in event of an outage. Extension for each town of Lira, Soroti and Mbale cost just about $30,000.⁴

In a post in a popular Kenyan advocacy group called KICTANET, Brian says, “we established a social enterprise that is using internet technologies to improve livelihoods for communities in this post-conflict region. This is mainly through making high speed broadband available to rural communities at low cost. NGOs, corporates and individuals alike have been flocking to take up the broadband services after years of poor quality and expensive services from the mobile operators who sell mainly data bundles that have poor performance. We ride on Uganda’s national optical fiber network (owned by the ministry of ICT’s National IT Authority – NITA-U) and from Kampala interconnect with a variety of bulk providers (Seacom, Liquid Telecom, Simbanet) who are connected to submarine networks via Mombasa. Our service approach has greatly challenged the internet services paradigm and scored greatly with our subscribers, many of whom enjoy better services in these rural towns than their colleagues in the capital Kampala. All locations and equipment are solar powered as there is very little infrastructure in Northern Uganda”.⁵

At the East African IGF held in Kampala, Uganda, where Brian presented their work to participants, he note that although the cost is still high, Zoom Wireless has helped bridge the gap and take broadband to nearby businesses which would otherwise be unconnected. This demonstrates that networks can be set up relatively cheaply to cover towns and districts that have existing mobile data but not a solid broadband service for the home or office.

The case for local hosting, and peering.

David Weekly, a Google product manager, did some tests while in Kenya to determine what really affects the local internet speeds in most African countries. David noted that the quality of the internet link was good but many sites were still loading slowly. On keen investigation, he found that Facebook and Google Kenya were reaching him from London; Apple from Paris; and Twitter from Atlanta.⁶ You can imagine that is the story for nearly every other service may it be Netflix, Amazon, Wikipedia or BBC. Angani, a leading Kenyan cloud service promoting local hosting did some good visualization on latency speeds across the world from Nairobi as their selling proposition on why local hosting rocks. Their results showed content from Europe takes 200 milliseconds while that from US takes 300 milliseconds to reach the end user in Nairobi while local hosted content would take just below 5 milliseconds.⁷ So where is the problem? David notes that a connection to the Kenya Internet Exchange (KIXP) for one gigabit per second costs internet service providers (ISPs) $350 per month, while the same amount of data from an international submarine cable costs a staggering $200,000. So even if a user pays an ISP for a fast internet connection in Kenya, data may flow much more slowly from overseas because of latency, and cost a hundred times more.

This compelling evidence points us to two things, local hosting, and local traffic—usually achieved through Peering at Internet Exchange Points. And that is exactly what Internet society and African Union has been trying to do with the AXIS project; taking IXPs where they are needed most.⁸ The AXIS project aims at keeping Africa’s Internet traffic local to the continent by providing capacity building and technical assistance to facilitate the establishment of National Internet Exchange Points and Regional Internet Exchange Points in Africa. The AXIS project aims to have 80% of Internet traffic exchanged in Africa by 2020, keeping local traffic local.

There are great sessions lined up for the IGF on IXPs. Please follow remotely or come to the meeting halls:

1. IXP Best Practice Forum (BPF) on Tuesday 12 November 2015.
2. Workshop 171 IXPs: driving connectivity and local economy on Wednesday 11th November 2015.
3. Workshop 201: Ensuring sustainability for IXPs in the developing world on Friday 13th November 2015.

Studies in several African countries have shown that poor people would rather walk to work, or go hungry, than have no money to buy airtime for their phones. So even if people are poor, they will spend the little they have to buy a cheap mobile phone and airtime. And when they become better off, they spend their new fortunes on better phones. Even without electricity, they power electronic devices in the most innovative ways. Solar chargers, bicycle dynamos, and old car batteries come in handy. But cellular companies, which often the case are profit oriented would rather concentrate their capital in larger cities than invest in poor rural neighbourhoods. And this is the case for USF. Ensuring availability of internet service and other communication oriented services even in remote areas. The USF should be used to build cellular masts in those remote places where there is no coverage, and develop policies to enable different network providers to share the infrastructure. Most governments in developing countries are in control of the USFs, but they are not doing much with the funds. One argument I have heard over and over again is that the funds are not enough to do anything tangible. But that argument alone shows lack of vision and commitment. The end user will appreciate even the basic services. Proper policies and implementation are key in the success of USFs.

¹ ICT Facts and Figures—The world in 2015
² Policy options for connecting the next billion
³ Communications Authority of Kenya Annual reports 2013-2014
⁴ In Africa, ISPs fill the broadband gap
⁵ [kictanet] Here’s the Real Way to Get Internet to the Next 4 Billion People
⁶ Why your internet connection is slow wherever you are in Africa
⁷ Average round-trip times (latency) for server requests from…
⁸ AXIS Project and AXIS Workshops

Mwendwa Kivuva is currently a Google Policy fellow based at Strathmore University’s Centre for Intellectual Property and IT Law, and a felow at Kenya ICT Action Network (KICTANET), Nairobi, Kenya, and is also an Internet Society Ambassador to the 2015 IGF held in Joao Pessoa, Brazil. Views expressed here are his own.