Sustainability
Finally, all of this must combine into something that you can create and sustain. For me, sustainability in ICT has a very real meaning (though I have yet to meet someone that can explain sustainability to me overall). Here, sustainability is two fold: setting up a complete, “situation,” that actually does, “just work,” while also teaching a counterpart how to maintain it, not just in their head, but also in their heart.
From a software and hardware standpoint, sustainability means creating a combination of the two that works well together, indefinitely. Perfect example: my friend Paul was asked to set up an installation for old PowerPC-era Mac laptops to be used in a solar-powered computer lab. He spent time researching the hardware and software available, and settled on a FOSS-based solution using Debian Linux pre-installed with some customized environment settings to make them easily usable by both teachers and children. The users of the lab, and even future administrators, don’t need to know what is going on under the hood, they just need them all to work. Of course, the project has only been implemented for a few months, so the sustainability has yet to be tested, but even if things go wrong, I still feel Paul acted in the correct fashion when selecting the software appropriate for use in the lab. Thankfully, when working with such old machines, you can perceive them more as the final solution and not necessarily a stepping stone to, “something bigger.” Why do you need something bigger if these machines do exactly what you need?
Also, creating a solution that continually needs to be tinkered with can set a bad impression on those who may take over for you. I love to tinker but I have had to hold myself back because I am learning that the more I tinker the more complicated the lab seems to people. My lab is infrastructure for learning and the goal of infrastructure is to just work. That is why I never have downtime during the week and perform any absolutely necessary upgrades on weekends, but even those I try to figure out which really are absolutely necessary and attempt to minimize them. This isn’t perfect because it means there may also be the perception that computers do, “just work,” but I hope, once I know who will be taking over from me, to teach them what is necessary while having worked my hardest in the meantime to minimize the workload each computer represents.
Which brings up the second fold of tech sustainability. You need to teach this all to someone. This is also where the heart comes in: you need to teach it with a passion. I have experienced very little passion about computers in my time here (with notable exceptions, don’t get me wrong), and this results in a, “whatever works,” mentality. “Whatever works,” is a great short-term mentality, but passion is what keeps the machines running for the long haul. If you can teach the practical skills necessary to a host-country national while at the same time instilling them with a passion for the machines, passion for how you operate them, passion for maintaining them to their tip-top performance, that will result in a sustainable ICT project. Which brings us to the next point.
One Computer at a Time 
Thanks, that was refreshing to see again that you’re thinking about a lot of the stuff I’d been thinking about in Tanzania. I would expand from just UPS too: I think that AVS and other line conditioning can be more cost effective for labs.
http://ttctip.wordpress.com/2009/09/21/umeme-uchafu-and-you/
I like FOSS CloneZilla cloning software but I haven’t tried PING.