ASUS EeePC pilot

Getting "enough" computers into classrooms is always a challenge. The definition of "enough" keeps changing. Infrastructure for today's computers is often absent in classrooms built years ago. Expectations of what computers should be able to do keeps increasing.

There is an assumption in our district that all the computers we have should be capable of performing all the functions we support. This is due in part to IT's desire for standardization of hardware, and in part to the ethos of equity valued by educators.

We need to think about providing different levels of computing tools for different tasks. This flies in the face of standardization, but could be a way to let us reduce the resource contention for high-power machines, especially in the OHS labs. Deploying thin-clients, mini-laptops ("netbooks" is what HP is calling this niche) or equivalent UMPC devices for low-level use like word processing and web research in classrooms could reduce the demand for more capable computers. It also provides "just in time" technology access in classrooms, supporting the integration of technology into instruction. 

We have launched two pilot programs in which students use ASUS EeePCs. One pilot is at Brooklyn Elementary in grade 4, and began in March 2008. The other is at Oregon High School and just began last week, with the goal of having English classes use the small computers in lieu of a computer lab (obviously they will have to change their classroom model a bit, but this is just a pilot). Each has five ASUS EeePC 2G Surf computers. At Brooklyn, the driving factor was getting power to multiple computers. We considered using older desktops, but they required too much power and counter space. It was cost prohibitive to install more electrical capacity. The EeePCs have low power requirements, and suprisingly good wireless access- better than our HP nx6110 and nx9010 notebooks.

We intentionally did not pursue a typical deployment of wireless laptops on carts, because we already tried that at Oregon Middle School, and it has not worked well. OMS had basic Linksys access points installed when wireless "labs on a cart" were deployed. The Linksys APs were unmanaged, not configured properly, and were eventually over-deployed. Replacing it with a managed Cisco infrastructure helped somewhat (at least it stopped paralyzing the wired network), but still did not make labs on carts feasible. The problem is with 802.11g's available bandwidth. A theoretical 54 Mbps is only 30 Mbps in real-world experience, and it is shared. Get 20 laptops in a classroom, and each now has only 1.5 Mbps. Now have all the students log in at once and try to load their roaming profile. If it normally takes 60 seconds on a 100BaseT wired connection, it will take 40 minutes on a shared 1.5 Mbps connection. Obviously, this isn't acceptable. So, we added more access points. Unfortunately, 802.11g has only three non-overlapping channels, so we can have a max of three access points servicing one physical space. Assuming the clients divide themselves optimally, the access time can be cut by two-thirds, which results in a thirteen minute wait. Still not acceptable. 

Until 802.11n is viable, there is not a technological solution to this.  The problem is the large amount of volume from roaming profiles being loaded on login. We tried shrinking the size of roaming profiles, but there were complications. Enter some creative thinking. 

We decided to skip the whole profile idea. These EeePC notebooks don't have roaming profiles. We aren't planning to put a Novell client on them. They will attach wirelessly to our public network, which gives them filtered web access, but prevents them from touching our private LAN, including servers and workstations. Unmanaged devices on a public-only wireless VLAN could mitigate security risks and minimize management burden, while providing the "cloud-based" tools everyone is starting to use. Students can either save on Google Docs or use the included OpenOffice applications and save on a USB drive. 

The fourth grade teacher in our pilot taught students how to use them, and then made the home page on the computers an iGoogle page which she had customized with content and links specifically for her students. She has students working on them in stations and small groups, but also encourages students to use them during class whenever they need access to information. We have not configured any printers on them yet, but she has worked around this by having students save documents on a USB drive, which she connects to her teacher desktop and sends to the server and a printer. 

This demonstrates what I expect will be model of the near future- we provide the network, students bring their computing devices, and applications are in the cloud. If Ray Kurzweil is right, computers will be tiny embedded objects all around us. We won't be able to keep student-owned computers out of our schools. This is an opportunity and a challenge. The opportunity is to allow students to individualize their learning as they use their own tools. The challenge is to protect the security of our data and infrastructure while providing access. 

Oregon Middle School and Rome Corners Intermediate are very interested in these as well. OMS would like something to replace the current wireless model. They began requiring students to purchase a USB drive as a school supply this year, so they are already accustomed to saving to external devices instead of the network. Some teachers are asking if we can change board policy so that we can encourage parents to purchase these devices for students. 

The price point of these devices ($299 for the ones we purchased) is a sweet spot, because it works out to the same TCO as a desktop. I anticipate a three year life cycle for the EeePC, versus the standard five year life cycle of our desktops. Since we won't be upgrading them, performance should not diminish during their life cycle. Maintenance will consist primarily of automatic updates and routine virus scanning. If one breaks, I doubt we will fix it. Although ASUS has a good reputation for providing replacement parts, these are tiny units. Once the first one breaks, we'll see how hard it is to replace parts. 

Our pilot programs are only a couple weeks old, so we don't have any data on their success. We know there is much excitement from students and staff about them. Students think they look cool, and staff like the idea of being able to afford a class set. (BTW, I consider a class set to be one computer for every two students; I'm not a believer of 1-to-1 computer programs.) 

Once we have more feedback from the staff and students, I'll post the findings. If you have any similar projects or data, please let me know.

---

UPDATE:
Status reports of this project have been posted in separate blog entries. Click here to see all posts related to the EeePC pilot project.