Africa's 5G Future

Anyone living in Africa and  interested in the future of affordable access will have spent some time trying to make sense of the complex evolving web of technology, regulatory policy, cultural issues, literacy, affordability etc that make up the ecology of communication infrastructure.  Fortunately the market has sorted out a big chunk of that out for us.  The future is mobile.  Even your mother knows that the future is mobile.  And mobile phones are evolving at a pace that no one dreamt of.

But other wireless technologies are evolving at an equally rapid rate.  In the little over 10 years of its existence, WiFi has gone from 1mb/s to over 300 mb/s in performance and while performance has gone up, price has gone down.  Nowadays, you find WiFi in an astounding array of devices from mobile phones to laptops to music devices to printers and projectors.  Access to WiFi networks has also exploded.  Nowadays a corporate building, public institution, airport, or even a cafe without WiFi is becoming a bit of an anomaly. But WiFi is not a mobile technology.

So how do those important but seemingly divergent technologies fit into the evolving technological landscape on the continent?  Especially in the context of concerns that 3G operators simply will not be able to cope with the exploding demand for broadband access.  In the U.S. the data demands of iPhone users has at times overwhelmed AT&T’s network.  AT&T has experienced a 5000 percent growth in data traffic in the last three years.  They have noticeably struggled to upgrade fast enough to cope with the demand, although things appear to have improved recently.

Lately motile operators have begun to hedge their bets in North America and Europe with the introduction of technologies like Femtocells to off-load network traffic .  Femtocells are consumer devices which establish a micro mobile base station in your home and use your broadband Internet connection to backhaul your mobile voice and data to the operator’s core network.  This takes the load off the mobile network for the operators and, in theory, saves the consumer money.  It is also a good solution for homes in areas with poor 3G coverage.  Unfortunately, this technology is unlikely to spread very far in Africa because it is designed mostly for people with high-speed adsl or cable Internet connectivity.

WiFi, however, is another possibility.  WiFi is nearly ubiquitous on the recent generations of smartphones. WiFi networks can offer complementary access for mobile users.  Technically, this is already true.  You can connect to a WiFi network with your smartphone in cafes and airports although authentication can be a pain and Skype over mobile IP is a pretty variable experience, from blocked to patchy to hey I remember it worked once somewhere.

So what would it take to have a seamless mobile / WiFi experience where you didn’t actually have to pay attention to what kind of wireless network you were on?  Well, as William Gibson says, the future is already here, it’s just unevenly distributed.  In the middle of 2009, Cherry Mobile, a Mobile Virtual Network Operator (MVNO) in Belgium, launched a converged mobile phone service which didn’t care whether your phone was connected via WiFi or GSM.  In fact, the phone would work without a SIM card as long as there was a WiFi network.  (Wouldn’t that just drive the RICA folk mad).  The downside is that this is a custom app that only runs on new generation Symbian phones.  Still pretty amazing though.

Will networks like Cherry make a difference in Africa?  Maybe but not as they currently exist I think as the solution is not generic enough.  Happily, the IEEE have been hard at work developing standards for making devices work over heterogeneous networks.  An excellent article by one of the smartest people thinking about the future of the Internet, Bill St. Arnaud, highlighted two emerging standards:

802.21 – The 802.21 working group is developing standards to enable handover and interoperability between heterogeneous network types including both 802 and non-802 networks.  This means that an 802.21 compliant device would be able to detect all available networks that it supports e.g. GSM + WiFi and would make the transition from one kind of network to another seamless.
802.11u – 802.11u is an emerging standard for internetwork roaming and authentication.  This would enable not previously authorised roaming on networks within a structured authentication and services framework.

If network and handset manufacturers start manufacturing devices compliant with these standards, things could get interesting.  While the obvious impact of these standards could be to reduce bandwidth demand on mobile networks, a secondary but possibly more significant impact would be to increase competition in mobile markets.  It would potentially allow for the development of bottom-up WiFi-based telephone infrastructure that could extend mobile networks or even provide alternatives.

Why SMS in Africa?

Erik Hersman recently tweeted “I’d like to hear more on whether we should build SMS or internet services in Africa?”  This had the serendipitous effect of breaking a bit of a blogger’s block for me.

I think most would agree that the answer is not either/or but a mix of the two.  That being the case, it it worth unpacking the merits of either option.  I’ve written previously about why IP-based infrastructure in Africa is essential to break down the walled-garden environments that have been established by mobile operators.  And though it’s good to have a vision of what communication infrastructure should look like, it is equally important to recognise what exists. Here are five reasons why I think SMS will remain important for some time:

  1. Familiarity of Experience:  An important reason for continuing to bet on SMS is the comprehensive familiarity of experience that it enjoys.  Everyone understands how SMS works.  It works pretty much the same on every phone.  SMS is a consistent user experience.  Contrast this with designing something even as simple as a USSD app and you find that individual phone design (both physical and OS) interrupts that familiarity of experience.
  2. Always On:  SMS is always on. You never have to worry about whether you’ve signed in or not.  Or whether it is taking up memory in your phone that is slowing everything else down or worse that it has completely taken over your user interface while in operation. You can rely on SMS to work as long as your phone is on.
  3. Already secure:  SMS is directly linked to your phone number which  provides a level of identification and transaction security.  This makes monetising transactions a whole lot easier.  This brings up a side issue for me as to why phone numbers aren’t as personal, universal, and operator-independent as email addresses but I’ll save that for another post.
  4. 160 chars is a killer app:  What we’ve learned from Twitter is that 160/140 characters is enough space to be valuable to everyone no matter how fast their Internet connection is.  SMS is the same.  There is room for loads of innovation with SMS alone, if only it were priced appropriately.   If SMS cost what it should cost i.e. just a tiny bit more than nothing, then ANY mobile can be an affordable, viable bridge to the Internet.
  5. Network Effects:  The biggest reason of all to carry on investing in SMS-based app development is the massive network effects that SMS enjoys.  Everyone uses SMS and is reachable via SMS.  In order for an IP-based service to be meaningful, it has to reach a critical mass of users.  MXit has achieved this in South Africa through a massive take-up among the teen population but I don’t believe this is a generalizable example across the continent, largely because MXit is not a very open platform for developers.  For an app to become popular a significant number of the people close to you have to be using it too.  It may be that Google’s Android environment does that.  Microsoft are also trying to solve that problem with their OneApp solution (Mibli in South Africa).  One the other hand, the soaring popularity of Facebook on mobiles may lead to it evolving as a generic platform.  Certainly, the first platform to establish itself as a popular consistent sign-on and open application development environment will be a game changer.

SMS Interconnect Fees

An interesting offshoot of my investigation in to Fair Mobile statistics was the discovery that some African operators charge an interconnect fee for SMS messages. Now interconnect fees are a topic of hot debate at the moment here in South Africa.  Interconnection Fees are the charges that operators levy to terminate calls from other operators on their network.  Regulators typically intervene on interconnect fees when they appear to be out of step with the actual costs of interconnection.  Of course the “actual” cost of interconnection is a subject of much debate because it represents both a cost and a source of revenue for operators, the details of which are rarely revealed to anyone by the operators.  Thus they are the subject of much speculation.

Less speculative is the cost to the operator of sending an SMS.  I have written previously on the egregiously high cost of sending SMSes in Africa.  However, to add insult to injury, in at least 17 African countries, operators charge an interconnect fee for connecting with other operators nationally.  In many cases they are doubling even tripling the cost of sending an SMS.  The argument for levying an interconnect charge is based on the need of the operator to recover the costs of terminating a call or in this case an SMS on their network.  But let’s face it, the incremental cost of terminating an SMS on an operator’s network is effectively zero or near enough to zero to as to make no difference.  In many countries, including South Africa, there are no internal interconnect fees for text messages.  So here is a list of the most egregious offenders:

Country Dominant Operator SMS Interconnect Markup
Niger Zain 200%
Uganda MTN 160%
Mauritania Mauritel 150%
Nigeria MTN 114%
Benin MTN(Mascom) 100%
Botswana MTN 100%
Congo Zain 100%
Gabon Zain 100%
Guinea Areeba (MTN) 100%
Mali Orange 100%
Rwanda MTN 77%
Senegal Orange 50%
Algeria Djezzy (Orascom) 43%
Kenya Safaricom 43%
Zambia MTN 40%
Ghana MTN 25%
Togo Togocel 25%