Monday, April 24, 2006

Case Study: Linux Thin-client Laptop Cart

From Daniel Howard (dhhoward@comcast.net):

Folks, a while back I built a rolling laptop cart for old laptops we had at the school that were pretty useless trying to run Windows, but work wonderfully as thin clients. With 12 laptop stations adding to the existing 5-9 thin clients in each classroom, we can provide a class with an instant 1:1 ratio for activities like creative writing, research, etc. Unlike most laptop carts, this one has removable shelves to put the laptops on so kids need only pull up a chair and start working. Since the laptops are thin clients, they stay on all the time, so transition from one group to another merely requires logging off one account and logging on another. It used to take 10-15 minutes to roll the cart into a room, distribute the laptops to students, boot them, and get connected to the wireless network, now when a class signs up for access to the cart for an hour, the students are on the cart working for at least 58 minutes of that hour. The laptops are all hard wired to a K12LTSP server on the cart with UPS battery backup, and there is a printer on the cart.

The cart has been a huge success. It's now placed in a wide section of a hallway with ample room for it, and classes on that floor share the cart; the principal reports that kids are on the cart every time she's looked at it.

Based on that success, we're building two more carts for similar floors, and what has dawned on us is that this is a fantastic way to transition cost effectively to a 1:1 ratio in the school. With 5-9 clients and a 12 station laptop cart outside in the hallway nearby, shared by only a handful of classes (as opposed to the entire school), the teachers get a 1:1 ratio available to them every day.

I'm posting this info not only because I want to share our experiences, but because it's a new (or return to an old) paradigm: the distributed computer clusters model. By distributing clusters of servers and thin clients, we move quickly to a daily 1:1 ratio for the entire school. We hear every day about the two extremes (computer lab and handful of PCs in each class vs. 1:1 ratio in entire grades or entire school). Our model is in between these two cases, and is much more cost and space effective than trying to put a 1:1 ratio in each class overnight, and faster than merely gradually increasing the number of PCs in each class. Until we move to all LCD monitors and diskless mini ITX thin clients, we simply could not put all those PCs in a classroom, nor would we have sufficient electrical power for that many CRTs and conventional PCs.

On another related note, we have found that the difference between having 1-2 working PCs per class and at least 5 PCs per class is huge: having 1-2 working PCs per class is almost like having none, the teachers didn't regularly incorporate the technology into their instruction. But with at least 5, the teachers are now all creating 'centers' based on the PCs, where at least a third of their students get to work on the PCs for up to two hours every day without having to go to the computer lab. It's been a seachange in how our teachers are using technology, and some teachers are already reporting performance and motivational benefits of K12LTSP and more PCs. The youngest grade teachers are raving about GCompris and Child's Play.

I got interested in this whole 1:1 ratio thing over the weekend, and found the following links for those also interested, first is a Massachusetts school that put laptops in front of every higher grade student:

http://www.portical.org/Presentations/bebell/


Summary slide:
What did we learn from exploring ?
• Technology Used More Frequently in 1:1 settings
• Higher motivation and engagement in 1:1 settings
• Computers became students’ primary writing tool in 1:1 settings
• Differences in classroom structure
• Differences in home computer use for 1:1 students

We've seen all of these effects at Brandon except that we haven't explored how home PC use may be changing, and have also seen improved student performance directly linked to more PCs.

Also, here's a consortium on 1:1 evaluations:

http://ubiqcomputing.org/eval_mat_framework.html

Good templates for surveys, evaluations, etc. on impact of technology on instruction.

Regards,
Daniel
Post a Comment