Lean or Thin?

All over the web I read that such and such project failed or was resented and sabotaged by ungrateful employees. This seems to be a widespread and dangerous reality/fear.

In IT, the big picture is that total productivity or cost/benefit often results in change. Naturally people avoid change if it means more work/discomfort while adjusting. This should be a temporary phenomenon and with strong leadership and openness everyone in an organization should accept change. Obvious exceptions are those subject to staffing reductions. Growing organizations should not have that problem except in particular segments.

The common wisdom/best practice seems to be to persuade/cajole/sell everyone on the change before making it happen. This seems to be self-defeating. If one spends similar effort to selling the change as making the change, the benefit may well be lost in delay/wasted effort.

In big changes I have made the effort to persuade upstream was minimal, perhaps a meeting and some spreadsheets. Persuading a hundred people downstream would be relatively a huge effort and unnecessary. Any organization attracts employees/members for various reasons, none of which are that underlings tell the bosses what to do. It seems strong leadership needs to support change and the pushback can be handled otherwise.

In my experience, it is sufficient to notify members of an organization that change is happening and why. This should be done at an early stage so that any real problems that higher-ups may not know about can be considered in planning change. Change should be quick. Done gradually, push-back can get organized and dangerous. Mobs are powerful. Individual grumps are not. Change should show immediate benefit. That prevents push-back from getting anywhere. Change should show long-term benefit. Everyone prospers from that.

Two of my favourite changes to IT systems are migration to GNU/Linux and migration to thin clients. Combined, we can have the perfect world that change can be quick and give immediate benefit: logins are faster, opening applications is faster. Such change can be combined with upgraded monitor, keyboard, and mouse to also show immediate benefit. Combining sugar with the medicine is a tried and true technique to fight push-back.

The fastest way to do such migrations is to obtain and configure a GNU/Linux terminal server, probably test it on some folks who invite change and then change every PC to visit the new server. The testing and development stage might take a few weeks but the changeover can be done as rapidly as thick clients can be converted to thin clients. A neat way to do that is to have a server provide LTSP and merely switch BIOS to boot PXE instead of from the hard drive. One can save some power by disconnecting hard drives at the same time. This takes only a few minutes per desktop PC and small organizations can easily do the change on a weekend. To the extent that larger organizations are modular, they could as well although coordination would be more difficult.

Every business and every non-profit organization needs to minimize the cost of operation to maximize benefits. Going lean in IT is one way to do that. Many organizations can cut the annual cost of IT from $300 per PC to $30 per PC with the change from that other OS to GNU/Linux on thin clients. They get a factor of 3 lower cost of hardware and manpower just by going with thin clients and a similar factor for going with GNU/Linux. The immediate benefits are huge. The long-term benefits are larger.

Consider an organization with 100 PCs going from 3-year refreshes and that other OS to 9-year refreshes and thin clients. Server licensing for that other OS is something like $1000+$40 per client just for file/print/authentication and some management. Server configuration for a GNU/Linux terminal server is about half-hour more work than configuring a normal PC, making LTSP work and creating accounts. You do that once with GNU/Linux and three times with that other OS.

Costs over 9 years (3 cycles for thick and 1 cycle for thin)

Item Thick TOS Thin GNU
Server Hardware $3000 $3000
Server licences $15000 $0
Server configurations $3000 $1000
Manpower $270000 $27000
Totals $291000 $31000

One can, of course, argue about particular items. I am assuming one full-time body at $30000 per year to hover over that other OS and one part-time body for the GNU/Linux system. It’s basically a single PC instead of 100. The lower manpower requirement is based on the longer up-time of GNU/Linux, less malware and centralized software management.

Schools of this size typically have GNU/Linux purring quietly or that other OS with a large percentage of machines down at any one time. I have been in schools where itinerant IT came every few weeks and in one day of work could not catch up with the workload with that other OS. They needed full-time support but had 10% coverage and it was not nearly enough. Where I worked this past year, no GNU/Linux system failed in the last half-year with 75 PCs running GNU/Linux. Of course, I did bork a few systems tinkering with them but they would not have gone down if I had not been there… 😉 I had a full-time teaching job and IT was “other duties as assigned”.

About Robert Pogson

I am a retired teacher in Canada. I taught in the subject areas where I have worked for almost forty years: maths, physics, chemistry and computers. I love hunting, fishing, picking berries and mushrooms, too.
This entry was posted in Linux in Education, technology. Bookmark the permalink.

Leave a Reply