Uncategorized

Why we urgently need to build out the local cloud.

What do I mean when I say “the cloud” or “clouds”? Well, basically, the term “cloud” is derived from the term “cloud computing” which is basically Internet-based computing. Many of us already use cloud computing on a regular basis for accessing services like web-based email (e.g. Gmail or Yahoo Mail) and Business Applications (e.g. Google Apps). The term “cloud” is therefore normally used as a metaphor for the Internet.

The beauty about the cloud is that services are accessed directly via a web browser and there is hardly ever any need to install software on your computer or server(s). In addition, it means that an organization does not need to invest in its own servers and applications so that instead it may use free or subscription based cloud services – thereby saving lots of time and money in the process. Another upside for using the cloud is its highly scalable and low cost which makes it highly attractive for businesses of all sizes. Which brings me to the reason why I wrote this post – that we urgently need to build out the local cloud.

Just a couple weeks ago East Africa experienced a major Internet outage due to a fault on the Mediterranean section of the SEA-ME-WE 4 cable. This cable connects cables like SEACOM and TEAMS onwards to Europe which is why we we’re affected. This came as something of a surprise for many in the region as we have more or less enjoyed uninterrupted broadband for the better part of a year. It was shocking to experience speeds that we’re even slower than what we had a year ago before TEAMS and SEACOM went live. It was Internet as we knew it, circa 2000. It was like being on a dial-up Internet connection once again.

The outage required Internet Service Providers (ISPs) and Telcos to re-route their bandwidth via expensive satellite connections so as to maintain some semblance of connectivity – this tended from very bad to quite slow depending on your choice of service provider. It is for this very reason why we urgently need to have local clouds that function even when international bandwidth and cloud are not available. This way, essential cloud-based services will continue to function locally.

The consequences, which we are already aware of is what happens when most of your business applications sit in the international cloud? What happens when you cannot get email or access your accounting service in the international cloud? The outcome is that you can’t work when this happens and the worst part is that you absolutely can’t do anything about it until international bandwidth and cloud are restored.

Given that its a well-known fact that undersea cables regularly do get cut or damaged, its likely that we will see more of the sort of outage we had two weeks ago. It also means that ISPs and Telcos will retain their costly satellite-based Internet connections for redundancy – just in case. Therefore, even if SEACOM and TEAMS have made broadband Internet a reality in East Africa we still really need reliable, high quality and world-class local cloud-based services.

The biggest caveat to setting up local cloud-based services are the massive financial investments required to set-up and operate data centers and applications. Data centers are also notorious for consuming large amounts of electricity (which is often unreliable in this part of the world) and require continuous technical monitoring and management on a 24 X 7 X 365 basis.  There are companies in the region that have set-up data centers to co-locate customer servers and provide bandwidth to the Internet. However, pricing tends to be rather prohibitive for most businesses.

Going forward, in terms of local cloud-based services as well as the applications and services offered, this could be best addressed via a three way strategic partnership model. In the first instance Government would provide the required incentives for businesses to set-up local data centers and cloud-based services. In the second instance local private sector players would provide the needed investment to make the local cloud take off. In the third instance international cloud players such as Google and Microsoft could provide the best practices and co-invest in building local cloud-based services. Ultimately, the local cloud build out would ensure service continuity even when International bandwidth or cloud are “offline”.

Previous post

Safaricom VS CCK.

Next post

LG's Android Optimus coming to Kenya in June 2010.

4 Comments

  1. May 9, 2010 at 5:53 am — Reply

    Moses, I think you should take this idea on step further out – cloud computing server gardens to backstop server farms:

    http://www.ictworks.org/network/ictworks-network/384?

    .
    .-= Wayan´s last blog ..African Open Source Technology and ICT4D Differences =-.

  2. May 9, 2010 at 7:32 am — Reply

    Moses,

    Your contention that backup continues to be a necessity despite the arrival of Seacom and TEAMS is absolutely true – and couldn’t have been illustrated more graphically than the recent outages. However, this will undoubtedly NOT result in a stampede to invest in satellite capacity… sou why not?

    The reason this outage occured was that all of Seacom’s traffic to Europe (and the Internet), and much of that on the TEAMS cable was (and as far as I know still is) being routed over a single cable – Sea-Me-We 4. Seacom takes Europe-bound traffic to India and TEAMS to the Middle East, before transferring it onto SMW4 to bring it back towards Europe.

    When the lifeline between East Africa and Europe was cut due to the damage to SMW4 in the Mediterranean Sea, the result was the problems we’ve all seen.

    Both Seacom and TEAMs are unprotected out of Africa – each just a single fibre pair running across the ocean floor – so any ISP with capacity on only one of them is at risk from ANY cable break north of Mombasa. Even those with capacity on both suffered during the latest outage because of a break along the 8,000 – 8,500km single point of failure between Fujariah (UAE) and Palermo, Italy or Marseilles, France.

    As you’ve highlighted in the article, satellite connectivity is currently the only option for diversity, but is extremely costly (and suffers its own challenges). However, a much more affordable alternative will be available within just a couple of months.

    WIOCC’s EASSy cable, which will launch in June/July, runs along the eastern seaboard of Africa from South Africa to Sudan, with landings in 8 coastal countries in between. It interconnects with a variety of international cables – including SMW3, FLAG and, later this year, EIG), taking African traffic directly to key internet exchanges and commercial centres in Europe, N America, the Middle East India, and Asia. It is designed to reroute in the event of a failure anywhere on the cable, either bypassing the affected part via a submarine route or via the terrestrial optical fibre networks of its shareholders and consortium members. EASSy offers a degree of diversity that simply doesn’t exist today.

    We are already talking to many of the ISPs affected by the recent outages about using WIOCC’s EASSy system as the basis for improved network resilience.

    Whilst satellite may have seen a blip in popularity for a week, I expect to see a very rapid migration from satellite to submarine systems that offer built-in resilience – such as EASSy.

    Mike Last, author of the WIOCC Blog at http://wiocc.wordpress.com

  3. May 13, 2010 at 11:51 am — Reply

    Yes Moses we do need to build our own local clouds as they offer us better redundancy should any problems arise.

    We at http://www.whive.com are opening our data center and will be offering cloud computing services to Kenyans.

    The main problem is the quite simply dismal performance by KPLC and Telco’s. They haven’t got customer service right. This i suppose is because they are a monopoly.

    Govt needs to liberalize completely all this sectors and ensure no one has more than 30% of market share at any given time.

    This will reduce costs and spur innovation.

    Kenya can offer cloud computing services to everyone if they get their act right.

  4. daniel
    May 27, 2010 at 9:18 am — Reply

    i think what we really need is a well designed cache… Too much stuff is being re-downloaded locally.

Leave a reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.