The Real Problem With Windows AND Linux In Emerging Market Education

  • Article

image 
The Ecosystem Impact of Affordable Computing

This is a post I've been meaning to write for a while.

This past spring Microsoft hired Vital Wave Consulting to create a five year Total Cost of Ownership (TCO) model to help us and our customers better understand the true cost structure for deploying large numbers of PCs into schools serving under-served student populations around the world. This is part of our goal to help transform education and is a hot topic these days in government circles.

You can find a copy of the Vital Wave paper here.

Among other things, we wanted to understand if Linux has a cost advantage over Windows when it comes to deploying large numbers of PCs into schools in emerging market countries. The study indicates that both operating systems have about the same TCO for these types of scenarios. Windows systems have a slightly higher up front purchase price, but this is offset by the hirer salaries required for Linux-skilled systems administrators in places like China and South America. So over a five year period, the total costs for a school system to deploy and maintain a large number of Windows PCs and Linux PCs are about the same.

Now before some readers of this run off and complain that this study is simply another example of Microsoft tech industry propaganda, please make sure that you read through the white paper that describes the model and and understand what it means. Vital Wave is a good company with smart people who have relevant experience in emerging market technology adoption, and they have done a thoughtful job in assembling their analysis.

For me, the huge, eye-opening takeaway from this work isn't that Windows and Linux cost about the same to put into school labs in poor countries, it's that the 5 year cost of ownership for doing so is about $2,700.

That's right, $2,700. At a time when the press likes to write about whether the $100 laptop costs $200 or $300, economists who live in the countries where these systems are being deployed went out, assessed actual computer implementations, and came back with an estimate that the actual 5 year ownership cost is about 10 times as much.

Kentaro Tamoya, who runs Microsoft's Technology for Emerging Markets lab in India, has observed situations where the cost of maintaining a PC in a rural village in India can run $100 a month.

Why so much? Well, machines break and need to be fixed or replaced (especially when they are used by kids). Teachers need to be trained. Software needs to be upgraded. Electricity can be expensive. These are the "laws of physics" involved in the deployment of large numbers of PCs and shouldn't come as a big surprise for anyone who has deployed computers for big enterprises. Simply because we are now deploying computers to a large number of rural locations doesn't make these laws of physics go away, in fact it can make them worse because in addition to the traditional fixed costs of computer deployments you now need to deal with environmental problems (heat, dust, rodents) and infrastructure problems (things like occasional 1,000 volt surges in power grids).

Don't despair though, because there is hope. Because the same techniques that enterprises developed in the last decade to drive down computer ownership costs to under $1000 over 5 years can be applied by school districts for their PC deployments. No one is disputing the power of computers as learning tools in the hands of children, the challenge is to drive down their costs, especially after the initial acquisition.

Erika Twani, who leads Microsoft's Unlimited Potential efforts targeting poor schools in Latin America, recently co-authored an academic paper that explains how to do this. Their approach is to take the Gartner Group's infrastructure maturity model -- a technology management framework with four levels (Basic, Standardized, Rationalized, Dynamic) used by many enterprises to manage technology costs -- and apply it to schools. The authors even added a fifth level, the "Chaos" level, where

"there is no network infrastructure, management policies do not exist, and there is basic or very limited dial-up access to the Internet. This is a scenario where the dynamics of teaching and learning are reduced to the level of the individual in a disconnected school. "

My assumption is that most of the schools surveyed in the Vital Wave analysis are "Chaos level" schools in terms of the sophistication of their IT infrastructure and ability to drive down deployment and maintenance costs. The schools bought PCs, put them in a classroom, and hoped for the best.

Erika and her co-authors go on to provide guidance on how schools can get out of this cost chaos:

How do you identify your school’s maturity level? What
are the milestones for each level? There are two simple
aspects to consider: the presence of a server and the level of
automation.

  • Server – the existence of a server is the milestone
    between the Chaos and Basic levels. Without a server, it
    is impossible to implement any kind of service
    automation, security or management. A simple software
    upgrade would require one workday for a small lab of
    20 desktops.
  • Automation – the level of automation (need of human
    intervention on a daily basis) defines the transition from
    Basic to Standardized levels.
  • A server with an ordinary operating system and no
    automation services requires approximately the same
    work as needed at the Chaos level. However, the
    simplest server currently in place is an advantage.
  • An effective operating system with resources of
    recovery policies, desktop backup and security tools,
    upgrades the IT to the Standardized level. This
    requires only a few hours of maintenance per week.
  • Adding the functions of client management (software
    distribution, asset management, desktop backups,
    desktop management and configuration), network
    anti-virus, and Internet firewall and filtering, upgrades
    the school’s infrastructure from the Standardized to
    Rationalized level. The need for human intervention
    is reduced to a few hours per month.
  • And finally, by implementing an external data
    warehouse or datacenter, the ICT infrastructure
    reaches its highest level of maturity, the Dynamic
    level. Services include disaster and recovery, remote
    management, remote software distribution and remote
    support.

This is the basic approach Microsoft is taking in our Unlimited Potential school deployments, teaching school districts and Ministries of Education how to take lessons learned from the enterprise and apply them to school labs, especially school labs in very remote and rural locations. Because these deployments won't work if we can't figure out a way to get ongoing ownership costs down to manageable levels.