Why Kenyan telcos need to start thinking like web companies


Image <a href="http://www.contrib.andrew.cmu.edu/~aishah/CC.html" target="_blank">via</a>

I am an entrepreneur in the cloud computing field. In preparation for deployment of a new product in the Kenyan market I recently set out to find local cloud hosting. I found nothing up to standard. Not even from Safaricom despite their widely publicized multi-million shilling investment in the ‘safaricom cloud‘. Apparently proper ‘cloud hosting’ did not make the list of priorities for them in that particular round of investment. In fact I could not believe it when I asked about their provisioning options and the quotation sheet had on it virtual machines with fixed capacities (giving us absolutely no variable let alone real-time control over resources such as the amount of RAM, number of processing cores or amount of storage space in gigabytes/terabytes. Needless to say this also meant there was no auto-scalability of such resources to adjust to usage demands on the system. To make things worse, their payment model is Pay-for-everything (whether or not you use it). This in a world where pay-only-for-what-you-use is practically the unspoken standard for cloud hosts everywhere. Don’t get me wrong. They have decent offerings for businesses looking to do the basic stuff. But nothing for someone looking to support a modern cloud based service with sensitive variables -a category that will see more and more entrants over the next few years.

When I asked the very nice lady on the other end of the line about the possibility of proposing an alternative provisioning plan (a man can hope) she said: “These are the only packages we offer right now and we do it under those terms and conditions.” I thanked her and hung up.

Okay. So maybe I’m the one who’s expecting too much here. But it led me to actually look around and see a worrying truth. Not only are we not where we should be, we are actually not focusing on getting there at all. We need to wake up, people. Being lost and aware of it is so much better than being lost and yet proceeding with confidence.

My three options now:

  1. Building my own datacenter which is obviously impractical (for several million reasons).
  2. Buy my own servers, rent a direct data connection along with rack space in one of the local data-centers and hiring someone with domain expertise in that particular area to run and maintain them which is alot more practical but expensive, cumbersome not to mention distracting and thus not ideal.
  3. Get easy and affordable hosting with the trade off being that it’s on foreign soil and not quite as reliable due to latency and may other factors.

For now we have to make do with international hosts such as Amazon, Rackspace, Toggle box and the like because they actually have what we need. Their products are miles ahead in terms of offerings and they are at the bleeding edge of innovation set on making those offerings even better. Their prices are pretty reasonable (believe it or not despite the vastly superior product quality Amazon is still cheaper than local hosting). There’s just one problem. They are continents away! They do not have data centers anywhere near here. When I host my service with them any new data has to travel across oceans to get to my user. The undersea fiber-optic cables suffer cuts every so often and those of us who rely on their data capacity are routinely left waiting for days till things return to normal. This is no life for web-based innovation -which whether or not we like it is something we will need to make everything else thrive. The sooner we acknowledge this as an industry the better off we will all be. We need to invest in support structures for the eco-system we keep talking about. We will need our own version of Amazon complete with elastic-compute-cloud type offerings if we are to pull this off.

The real problem of course has its root in the fact that Safaricom is at it’s core a telecommunications company (and so is its parent company Vodafone, Europe). As a result they quite naturally think and manouver like a telco. This is the identity that dictates the way they build their business, which areas they go into and the way in which they order their prioroties for investment. Unfortunately it also affects the way in which they innovate. You see what I am talking about here is strictly speaking not in the telco field. This is web company territory. Even so there is a growing overlap. As voice revenues drop, data has evolved into the next big earner for telcos and as such they have poured billions of shillings into optimizing their products and business models to this shift. This has resulted in data centers being built, miles of optical fiber cable being laid and a hitherto unseen tsunami of marketing centred around pushing us all online -as long as we use their data connection. Despite all this investment it is as though none of them have realized that the data business is in fact a web business and in the web business you cannot think and move like a Telco.

I would think that the opportunity here is rather obvious for anyone with a data-center filled with cloud-capable servers. The developer community is growing both in size and clout. The quality for products and services is rising. Standards are being raised -slow, yes but steady nevertheless. Even for innovations designed for feature (non-smart) phones the actual service itself needs hosting. New companies are cropping up here and old ones are embracing more and more technology. As someone right in the mix of things I can say with confidence that cloud computing is going to be a serious factor in the shaping of everything from core-systems we are building to the way we do everyday business. It is going to affect everything. We are talking of building technology cities like Konza with business parks occupied by multinational corporations and SMEs alike. How do we intend to convince anyone to move their company into a country that’s not cloud-friendly? At the core of all these things is one thing. The datacenter. It is the lynchpin upon which everything else is built. In the information age a good, secure, reliable datacenter is worth several nuclear submarines combined. Just a few of those on Kenyan soil would change the trajectory of local innovation in an exponential manner and we already have the capacity to begin building the right foundation for it. This isn’t stuff that needs to wait till tomorrow. It can start today if we so choose. Unfortunately no one seems to actually be acting on it.

Why is no one in these companies thinking in this direction? Telco company or not I understand the importance of hanging onto every cent the traditional business models yield even as you try to expand into or grow relatively new markets but if you’re going to try your hand at something you might as well go for the win. If you have a data-center on Kenyan soil (no matter how small) you have billion dollar companies like Amazon beat as far this market is concerned. But only if you can adjust your operational models and offerings to become a worthy rival. Safaricom and other Telcos going into data need to realize one thing. They are a huge part of the data-driven tech eco-system in the country and if they go astray or lag behind so does everything else.

That said, if they do not wake up, smell the coffee and do it someone else eventually will -and in the process create a colossal business around it. As one of the people who dream of this day I really don’t care which happens first. I only hope that it happens soon and we get a fighting chance at keeping up with the global leaders because otherwise anyone innovating in the web-based arena locally will stay frustrated and the few that do make it will go the Craft Silicon way and simply move to other countries.

  1. While I agree with the general premise about the lack of viable local cloud computing offerings I don’t think it’ll be a simple business for Safaricom or other telcos to get into. For example, if Amazon had an African / Kenyan availability zone I doubt any of these telcos would be able to compete.

    1. True. But at the moment it is easier for them to get into it than anyone else which is why they are the subject of the post. Yes, if the likes of Amazon set their sights on Kenya then the competition would be ridiculous but seeing as they aren’t that still leaves innovators out in the cold. What is being suggested here is providing a viable alternative with the one chief advantage of proximity (where the cloud is concerned geography controls everything from legal jurisdiction over data to extrapolations such as taxation implications).

      The reason the telcos are the ones in the best position to do this is because they do not need to start from scratch. They have 85% of the job already done so the CAPEX and overall friction to market entry is greatly reduced. This is simply a use for the infrastructure they already have. Capex is roughly the same regardless what you use the data center for. All they need to add here is an international standard type offering. Secondly, mobile is the big thing here and as far as developers are concerned the telcos are the big fish in this arena. With smartphone prices dropping and web infrastructure improving it would be prudent to invest in a market presence on the back-end side of these innovations as opposed to just trying to maximize revenue on data. It’s actually more profitable for them to do this because their revenue will increase long before the entire process of making internet access ubiquitous is complete.

      All these factors come together to salvage local web-innovation (in a timely fashion) while at the same time creating a better-adjusted revenue model for their data center(s), not to mention the significant positioning as the go-to local cloud host for all the new stuff coming up now and in the near future. They immediately become globally competitive even if they focus on the neglected East African market only. Looking at it objectively it is not only the simplest long term solution (occam’s razor), it is actually the best under the circumstances.

      If and when Amazon et al ever decide to show up here they will be fighting from behind as opposed to simply swooping in to dominate an ignored but growing market segment. Still good for both telco and innovator.

      1. There’s a lot of truth in your response. However, worldwide, the infrastructure part of cloud computing is fast becoming a utility, like electricity. The real innovation is now on the value added services, of which I categorize into two. First are those targeted directly to consumers e.g. Dropbox – which used to actually use EC2. Not sure if they still do. The second is targeted to Developers / Integrators e.g. Heroku. This VAS layer is what powers the goodies you seek such as flexible pricing, high availability e.t.c. I wonder how much of the 2 billion Safaricom ‘spent’ on their cloud was on dedicated to software? I speculate that it was mostly on buying boxes and virtualization licences.

        For now I recommend anyone getting to the cloud computing space to check out Eucalyptus (http://www.eucalyptus.com). It’s like having your own EC2 availability zone. It’s actually ‘simple’ to convert you local server into a private cloud. I’ve used this for an instance where a client was mandated by law to have some data physically located within the country.

        Also I wouldn’t rule out Amazon setting up locally. They already have a development centre in South Africa http://www.amazon.co.za/aboutsa.htm and they are expressly targeting EC2. If I were in Safaricom, I’d be worried / busy innovating.

  2. It just so happens that I’m starting this very service and it will be co-located in Nairobi and Cape Town with one of the sites going live this summer. It will be API compatibile with AWS and will support the equivalent of ec2, S3, and ebs. The underlying software will be OpenStack and I’d love to talk to anybody who’s interested in the project.

    This way, people get all the benefits of local hosting without the hassles.

    1. This summer you say? Sounds interesting. That would make a huge difference. I’d definitely love to talk about it sometime. Ping me on twitter? (@the_kageni_mind)

Comments are closed.

More Stories
Carol Musyoka has been appointed as an EABL Director