Benoit and Revolution Linux:

Two school boards in Canada. Infrastructure: for up to 35,000 users. File/mail servers, LDAP/ authentication, printing servers. Firewalls, proxy, etc. all being Open Source. Getting this automated and easy to manage. Automated doesn't mean simple, but have the tools for configuration. This is scalable, so they don't spend days and days on tweaking one server. Cluster functions. Virtualized or not. Mainly address all those functions. Main key: automating account creation. Centralization --> Decentralization. Each school given a certain amount of space, but they can decide how to use it. IT department doing less and less, transfering management locally.

Michigan City

1100 machines with 1900 students in one building. How to manage this many machines. How to image. Maintaining an understanding of what teachers want when they turn on the machines. 29 classrooms x 30 = 900 for Indiana Access. Laptops in the science classrooms--one science class. Made the decision that desktops are the way to go: 6 years, but laptop won't last that long. Mini laptops are wonderful, but don't meet the testing specs, most of which spec a 15" screen. The problem of managing laptops. In science, physically don't have space for full desktops. Curricular side: within 6 months of having technology, teacher says she is the "information facilitator, no longer the "talking head." Moodle is the cornerstone of the entire project, putting all facets of learning styles into environment. Without a certain number of teachers in a building immersed in the technology, hard to get acceptance. Teachers helped to build program. Have to have 5 classrooms in this environment. Otherwise they feel stranded. Helps to have same grade level / subject. Really helps to have "champions."

Forest Gaston

Different environments. 2) One school the network guy is a deputy sheriff. Teachers come out in fear--wears uniform at school. Everything is security. Openness of tech director and relation with teachers is critical. Relationship. 2) Struggle with staff / budget. Went from two tech to one, overwhelmed. This project will probably never take off. Very limited Internet. 3) Greensburg. Ncomputing--still have beta version, one year later don't have stable version. Technology that they chose wasn't ready. One of the large-scale LTS environments, good but political environment is not--problem of relationship with tech director and others. It's not just about having perfectly-working computers. You have to have: good working computing, good educational use, and good training and relationships.


1) Thick-client Extremadura. Total control or no control. Control every students PC with the mouse on their screen. 2) Low-fat, diskless solution. Half the cost of operation. Same facilities, same video, software is loaded from server and runs on PC. Different from LTSP. 3) Large city with multiple clients: Mac, Windows, or Linux. Take free software and connect to MS infrastructure. 4) Open service infrastructure on free software and you can connect. Have choices of architectures. They think they have 7% of schools now. Expect to double in the next year. But are competing with too much money, so no pressure to solve, not cost-conscious.

René Marquis

Thin client implementation. Goal to go to 80% network-centric PCs. Small footprint, 14 watt, fanless. Everything they do is web-based. Virtual desk. Not enough money to do 1:1 or stand-alone PCs. Universal access to the virtual desk, so went to thin client. 10,000 PC to upgrade to new version of OpenOffice. But with thin client can do in 30 minutes. Replicates automatically. This is where they save a lot of money. Scale down to 6 people. Phase 1 done now. Went 1:1 with teachers. Each teacher has their own PC (thin client), and can access from home. They are a "professional" and they should have their own machine. The younger teachers are not necessarily computer literate. Still looking for the killer app, that will save the teacher time. Thin client--not easy to do, but a great time-saver and less expensive. They also have laptops for students (2k or so). Phase 2: go to administrative personnel and trade machine for thin-client. Initially will be windows-based, concept will be virtual machine. Last May: 2600 computers in five weeks. This is the only way that you can do this. 500 a week without increasing staff. Doing the old way would have taken many months. Has 1.5 people running the thin-client from central.

Jim Hare

How do you plan for large-scale environments when there is no money? Detroit. School district has a $500 million deficit. How do you pay for a large-scale deployment if there is no money? Looked at grant opportunities that are available to specific low-income areas. Title I: how many children are enrolled in the free lunch program. E-rate: universal services tax on phone bill. $2 billion each year. Every school in US will get their Internet and telephone service taken care of by this fund. Very low income schools can qualify for rewiring. Some schools being rewired every to years. Sometimes just changed labels, but bills were sent in and paid by funds. What if you could do something legitimate with this money? LTSP: category 2 allows you to do terminal services. A school district can build out their infrastructure with all of it being paid by E-rate. Zero-cost desktop. Focus on reading and math. Thinks they can get 100,000 users on a single image. Big school districts own the "Z" servers. The money is there.

Wanted to take scalability off the table. 110 --> 200 volt. Worked with Sun. Working hard to make sure that IT and curricular are joined at the hip. Community piece: getting everybody talking about it at the same time. Having champions in large scale is critical. And multiple champions are even better. Using E-rate: have access to Internet and Moodle and other tools. As an ISP.

John Gosney
Similar issues. 100k users. Founding member of SAKAI. They moved to quicky with Sakai implementation. A lot of very negative kickback. Perception that they weren't delivering as much. So ran systems concurrently. Worst possible situation: terrible performance, etc. (Last fall.) This semester was without incident, retooling: support, application, and infrastructure. Still an uphill battle: to show how to do what they were used to doing, and what they can do that's new. Faculty member purposely put in role to serve as intermediary to builders/supporters. Now that system is stable, pedagogical community is beginning to grow. Starting to be fun. Starting their next strategic plan: user expectations, ubiquitous availability, delivery to mobile devices. Dealing with same issues, on large scale. Martin: don't say, we saved a bundle of money that can now go out of IT.

Marcos Castilho
Computer labs in each public school in Parana State. 1.5 millions students, 57k teachers. Very few Linux managers. Secty of education wants labs to work. How to manage the structure with very few people. Bring management to central point. Two years of preparation, now can manage 44,000 terminals with multi-seat. 20 heads from multi-seat server in each school, using one or more modern server(s). Server works as contol point in the school. Local manager is a teacher who does not necessarily know computer steps but he can through a simple interface create user accounts, allow machines to be put in network, or printers, and local structure. He can contact the center for more help. Didn't have to worry about power or Internet access. Power company, owned half by state, put in Internet; bank financed the computers. All teachers trained, and educational portal. Being trained to use Moodle also. 32 educational regions. Train people there who train the rest of the teachers. Their governor loves the free software concept. Microsoft has tried to give software for free, but they turned it down. 70% of students did not have access to any computers, so if first computer is Linux, it's OK. Federal government likes Linux as well, and made mandatory the use of Linux in schools. Private schools use Windows.