Improving the Ratio

First off, thank you Module 7 for this presentation, it’s got me thinking in some new directions.

I was thinking through some of the issues with moving towards a 1:1 ratio, many of which are being discussed in the forums and comments, and looking for inexpensive ways of deploying more computers for students to use. I have better than a 1:1 ratio in my class, but my school at large is nowhere near. Here are two technologies that I’ve found, and I’m hoping that together we can find more.

Linux Terminal Server Project (http://www.ltsp.org/)

Linux Terminal Server Project (LTSP) is designed to connect multiple low-cost terminals to a powerful server which does all of the processing and storage for all users. The terminals can be old computers or newer thin clients, as long as they have a basic processor, video card, network card, and keyboard and mouse support. They don’t need any drives, as storage is handled by the server. Input by users is sent over the network, the server does the processing, and the results are resent over the network for the terminal to display. Practically, it’s like having your own machine.

The advantage to this is in costs and maintenance. The server is the only machine that needs to be setup and maintained. All of the terminals are interchangeable, and with the drives removed, the most common failures are gone as well. A single quad-core machine acting as a server (about $700) should run about 30 clients at reasonable speed. If the terminals are powerful enough, some processing can be done on them to reduce the load on the server too. Just about any old computer can be a terminal, so the clients can be old inventory or donated machines. Ubuntu includes a LTSP configuration in the install disk, so setting up a basic lab can be done in just a few hours.

The issue with this setup is that it uses Linux. While the OS and applications are free, it may be difficult getting IT to manage the system. If the IT already supports Linux, it should be an easy implementation.

Userful (http://www2.userful.com/)

From what I can tell, Userful behaves similarly to an LTSP environment, but instead of dummy boxes connected by network to the server, a keyboard, mouse, and monitor is connected directly  to the server. It also uses Linux as the OS, so the same implications with software as the LTSP apply.

There are some performance advantages to this setup. Since each monitor/desktop has direct access to a video card, video performance is improved. Also, since a server is limited to about 10 user desktops, each user could have a larger slice of the processing power.  The company claims that it costs about $70 per user in hardware costs in addition to the server. Here’s my own cost estimate for a brand new lab of 30 machines. You would need 30 modules ($2100), 3 servers ($2100), 15 video cards ($900), 30 keyboards, mice, and monitors ($6000), which comes to $11100. I imagine that many schools would be able to scrounge the keyboards, mice and monitors, and that would reduce the cost significantly, to just over $5000.

If you know of other options, I’d love to see more solutions for schools looking to expand their computer offerings.

6 comments


1 davidp { 10.30.09 at 2:06 pm }

Byron, you are clearly on the en/intrapreneurial path to success.

Heck, this could be the start of your pitch for A3.

EVA kicking in here…

I understand the cost savings of the hardware and software configurations you’re pitching. The two schemes look really attractive and very effective ways to provide service to students at reasonable cost.

But… isn’t the weak link in these terminal-server schemes the transport mechanism, the strength and speed of the network on which thin-client, or even thick-client systems run? How will you achieve the kinds of performance needed to satisfy the various media needs of students on a wireless network, or on a school network in general?

d.


2 Liz Hood { 10.31.09 at 5:58 pm }

Great post Byron. I will be the first to admit I do not understand completely how this would work, but the price certainly makes it attractive (and therefore feasible). You have piqued my interest and I am going to ask my IT guys about it. The only concern I have is “will it do what I want it to do”? Speed? Web 2.0 apps? Philosophically, I support OSS; but bottom-line when I have a class of 35 students I need it to work and work well. (Always have a Plan B of course). And … nothing more frustrating to tech-phobic teachers than to have glitches; hard to convince them to attempt again.


3 Barrie Carter { 11.01.09 at 9:43 am }

Hello Byron:

Thank-you for the informative post even though I do not understand half of what you said.

However, I will definitely approach the principal, the district principal of learning services, and the IT department, using your information as a springboard for dialoguing.

After all, my school’s computer lab needs a complete overhaul and I am in great support of the 1:1 student-computer ratio in every class as well.

Cheers,


4 Byron Kask { 11.02.09 at 8:55 am }

David and Liz,

From what I understand it’s not just the network speed, but also the hard drive access time that are real bottlenecks when there are lots of users running different applications. The feature that is currently being implemented is the ability to run applications locally on the client. Once the program is downloaded, it should run at full speed. With so much being done on the cloud now, a locally run Firefox should meet most of the users demands, even in media-rich Web2.0 environments.

Barrie, I don’t know if I’d recommend this direction if there’s a budget for a new lab. When they replace the lab, the old machines can then be used in an LTSP configuration.

I think the ideal way of using LTSP is to have a few client machines run off of a teacher’s workstation. In a high school social studies class, for instance, you would be able to provide a few in-class machines so that students could do some quick group research without having to go to the library or lab. I know that this is getting away from 1:1, but I think it would be very helpful for teachers without placing any burden on the existing infrastructure.

I’ll set up a network and try this out when I have the chance, but it’s a pretty neat concept.


5 davidp { 11.02.09 at 11:59 am }

Still think this would be a great pitch. Think of it, your own service company, Byron!


6 Byron Kask { 11.04.09 at 8:36 pm }

David, thanks for the encouragement in this direction. I think that I will pitch something to do with LTSP.

You must log in to post a comment.