A bleary eyed Massachusetts Institute of Technology student roaming the ground floor of Building 1 in the wee hours might hear a rumbling coming from the “shake table” lab, where researchers simulate earthquakes. But a peek into the lab would reveal that nobody is there. More than 5,000 miles away at Obafemi Awolowo University, in Nigeria, morning classes have begun, and a student might have just loaded the Kobe Earthquake simulation, turned on the lights in the Cambridge lab, and is gathering data from MIT’s equipment.
Thanks to software developed and distributed by the institute, students can control MIT’s cutting edge equipment while eating cereal at their laptop in a Cambridge dorm, or from another continent. Students from Singapore to Sweden have logged in to run experiments on MIT’s equipment in the last few years. Now, thanks to a grant from the Carnegie Corporation of New York, MIT is getting African universities, which would not otherwise have access to top-of-the-line equipment, into the “iLabs.”
“It’s very much in the spirit of OpenCourseWare,” said Steven R. Lerman, director of MIT’s Center for Educational Computing Initiatives, referring to the institute program that made all MIT course materials available online.
Jesus del Alamo, an MIT professor of electrical engineering and computer science, led the development of the iLab concept beginning in 1998. He wanted experimentation to be a bigger part of undergraduate science education. “It started out of my frustration with teaching microelectronics,” del Alamo said. “The classes are large, the equipment is expensive, there was no way for students to make measurements.”
A semiconductor experiment in del Alamo’s lab might take 20 minutes to set up, and only 10 seconds to run. If multiple students could set up experiments from their personal computers simultaneously, the setups could be stored in an electronic queue and run in order. Rather than three students using the lab each hour, hundreds of students could get in electronic line.
That’s where Lerman, a professor of civil and environmental engineering, came in. Remote control lab equipment is nothing new in itself, but in the past the software that students use has been specifically tailored for narrow purposes -- for example, one particular lab usable only by students with MIT logins. Lerman created the “service broker,” a piece of software that has opened iLabs to the world.
“[del Alamo] enlisted my group,” Lerman said. “He said something like, from 4 a.m. to 6 a.m. I want to make equipment available to students at another university.”
The service broker, developed with a Microsoft grant, solved a couple of tricky remote lab problems. At first when iLabs went online, every time somebody from outside MIT set up a user account, MIT had to authenticate and manage the account, which includes storing the data. The service broker created a “software backend,” Lerman said. Now outside users just install the service broker, which is available free online. MIT then authorizes the user’s server, and authentication and data storage is “outsourced” to the user’s server, so as not take up MIT’s time or storage space.
“All we do is provide our lab,” Lerman said. “We’ve designed this to separate out things, so it’s easy to be generous.” Easy enough, in fact, that in the past year, del Alamo reached out to three African universities -- Makerere University, in Uganda, the University of Dar es Salaam, in Tanzania, and Obafemi Awolowo University , in Nigeria -- to share MIT’s expensive equipment.
The Africa project is being aided by an $800,000 grant from Carnegie Corporation of New York, and will include exchange of students between MIT and the African institutions in the interest of developing the use of iLabs. In some cases, students have simply taken it upon themselves to spread the iLab gospel. Loia Na’amani, an MIT civil and environmental engineering graduate student who did his undergraduate work in Lebanon, housed a faculty member from Lebanon so he could work with Lerman for a few weeks in the summer.
Every six months, an updated service broker is released in the hopes of making remote experiments even more convenient. Lerman noted that, even though iLabs are essentially free for institutions outside MIT, African institutions might still face prohibitive costs. At Makerere, for example, “if you adjust for per capita GDP, they pay 10,000 times as much per bit per second” for Internet use generally, Lerman said. “They use satellite links. There’s no fiber conductivity to most of the east coast of Africa.”
To minimize the use of expensive bandwidth, the service broker only communicates with the lab when it has to. A student logs on to the service broker, and designs an experiment all on his or her local server. The service broker then connects to MIT briefly to send the experimental setup.
Even so, iLab adoption has been slow at Makerere and Dar Es Salaam. “They face a lot of barriers,” Lerman said. “If something goes wrong here, we can just call Dell up.” Not so for those institutions.
At Obafemi Awolowo, however, students are starting to get into the swing of iLabs. According to L.O. Kehinde, coordinator of the iLab project at Obafemi Awolowo University, about 300 students will do at least one iLab experiment in the next month, mostly studying transistors. "It's obvious that the iLab is allowing more students to do experiments," Kehinde said in an e-mail. He added that "inadequate access to networked computers" has been a major hurdle, but that, with iLabs available, "there is a plan to expand access by over 300 percent for the students within the next three months."
Currently, thousands of students worldwide are using iLabs, and still that is only a fraction of the number iLabs can serve. The next step is to update the service broker to allow not only experiments that are set up and run separately, but those where the user is modifying the experiment while it is being executed. To that end, the Center for Educational Computing Initiatives is programming a new scheduling system for users who can simply load a setup into the queue and go to lunch.
Sitting in his office in Building 9, Lerman logs in to an iLab and loads an experiment that will measure the current across a diode. He spins his chair around and explains that, in time, perhaps institutions can collaborate on their shopping lists, so rather than every campus buying the same piece of equipment, institutions can buy different things and share. Seconds later, when he spins back around, his data is waiting for him. “The long term hope,” said Lerman, “is among peer institutions you can do quid pro quo … you could have a barter economy. You could certainly imagine, in poor countries, universities gathering together.”