mit ilabs: towards a community of internet accessible laboratories phillip d. long, ph.d. mit mark...

MIT iLabs:MIT iLabs:

Towards a Community of Towards a Community of Internet Accessible LaboratoriesInternet Accessible Laboratories

Phillip D. Long, Ph.D.MIT

Mark Schulz, Ph.D.University of Queensland

Open Educational Resources for Network-enabled educationAugust 18-19, 2006, Hotel Taj Ambassador, New DelhiSponsored by: National Knowledge Commission

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

ContentsContents

iCampus - context iCampus - context MotivationMotivation iLabs at MITiLabs at MIT The iLab ArchitectureThe iLab Architecture Unique issues for iLabs in the developing worldUnique issues for iLabs in the developing world iLabs at UQiLabs at UQ Opportunities for partnershipOpportunities for partnership



iCampus is an MIT research program to iCampus is an MIT research program to enhance university education through enhance university education through information technologyinformation technology Established in 1999 with funding from Microsoft Established in 1999 with funding from Microsoft

ResearchResearch 2000 – 2004: Focus on impact at MIT2000 – 2004: Focus on impact at MIT

Sponsored projects involving over 400 MIT faculty, Sponsored projects involving over 400 MIT faculty, staff, and researchersstaff, and researchers

Has affected nearly 100 subjects with a combined Has affected nearly 100 subjects with a combined enrollment of over 5,000 students per yearenrollment of over 5,000 students per year

2004-2006: Focus on global dissemination2004-2006: Focus on global dissemination

What is What is iCampus?iCampus?



Remote access to laboratory equipment (iLabs)

On-line lectures and exercises in computer science (xTutor)

Technology enabled active learning (TEAL)

Web-administered writing examinations (iMOAT)

xMAS - Cross Media

Annotation SystemSketch for Understanding

Spoken Lecture



Motivation to iLabsMotivation to iLabs There is enormous educational value in

hands-on laboratory experiences

But, conventional labs… … are expensive and have complex

logistics … can’t easily be shared

iLabs: real laboratories that are accessed through the Internet from anywhere at any time potentially by large numbers of students



Goals of iLab project at Goals of iLab project at MITMIT

To leverage the pedagogical potential To leverage the pedagogical potential of iLabs in science & engineering of iLabs in science & engineering educationeducation

To develop a scalable framework to:To develop a scalable framework to: Ease the development of new iLabsEase the development of new iLabs Facilitate iLab managementFacilitate iLab management Enable worldwide sharing of iLabsEnable worldwide sharing of iLabs



iLabs at MIT iLabs at MIT

Microelectronics device characterization (EECS, deployed 1998)

Shake table (Civil Eng., deployed 2004)

Dynamic signal analyzer (EECS, deployed 2004)

Polymer crystallization (Chem. E., deployed 2003)

Heat exchanger (Chem. E., deployed 2001)



Microelectronics Device Characterization



W2000 server (software donation of Microsoft)

Semiconductor Parameter Analyzer, Switching Matrix (donation of Agilent Technologies)

Device under test

Device test fixtures (donation of Agilent Technologies)

Microelectronics Device Characterization


are needed to see this picture.Lab CapacityLab Capacity

2PM: 6.012 exercise out

(75 students)

4PM: 6.720J/3.43J exercise out (25 students)

2PM: 6.012 exercise due

4PM: 6.720J/3.43J exercise due

• When do students carry out assignment?

• What is the lab system capacity?

[Oct. 13-20, 2000]


are needed to see this picture.Lab CapacityLab Capacity

2PM: 6.012 exercise out

(75 students)

4PM: 6.720J/3.43J exercise out (25 students)

2PM: 6.012 exercise due

4PM: 6.720J/3.43J exercise due

[Oct. 13-20, 2000]

System capacity: > 2,000 users/week, > 15,000 jobs/week


are needed to see this picture.Typical Assignment: Typical Assignment: Microelectronic Device Microelectronic Device Characterization ProjectCharacterization Project

Four step process:

Using iLab GUI: Measure DC I-V device characteristics Graph results

Download data to student’s computer

Using MATLAB or EXCEL: Extract device parameters Construct model based on theory presented in class Compare with measurements, discuss

Freely explore other modes of operation…



Formal AssessmentFormal Assessment

Assessment performed in 6.012 in:Assessment performed in 6.012 in: Spring 2005 (90 students)Spring 2005 (90 students) Fall 2005 (65 students)Fall 2005 (65 students)

Assessment through:Assessment through: Individual student interviews (~20 Individual student interviews (~20

students)students) Quantitative surveys (~30-40 students)Quantitative surveys (~30-40 students)


are needed to see this picture.What we’ve learned from What we’ve learned from E ilabE ilab

iLab experiences can significantly enhance iLab experiences can significantly enhance learninglearning

For iLab educational experiences to be For iLab educational experiences to be effective:effective: system has to work well, specially under peak system has to work well, specially under peak

load conditions!load conditions! system must allow free exploration and making system must allow free exploration and making

many mistakesmany mistakes clear documentation and tutorials are essentialclear documentation and tutorials are essential

Several small assignments more effective Several small assignments more effective than few large projectsthan few large projects

Students find difficulty in handling real-Students find difficulty in handling real-world dataworld data offline, post-measurement portion of offline, post-measurement portion of

assignment critical to learning experienceassignment critical to learning experience



over 3600 student users over 3600 student users (for credit) since 1998(for credit) since 1998

Lab UseLab Use



Lab Server

The MIT iLab Architecture

Service BrokerClient

Campus network Internet

University Databases



Unique Issues for iLabs in Unique Issues for iLabs in developing countriesdeveloping countries Opportunities:Opportunities:

Paucity of labsPaucity of labs Lots of young enthusiastic peopleLots of young enthusiastic people Great need for engineersGreat need for engineers

Challenges:Challenges: Limited access to networked computersLimited access to networked computers Limited computer literacyLimited computer literacy Severe bandwidth limitationsSevere bandwidth limitations



Bandwidth limitationsBandwidth limitations(example: Makerere University, Kampala)(example: Makerere University, Kampala)

satellite gateway to Internet (total bandwidth of Uganda=25 Mb/s)

campus wide single-mode optical fiber (2 Gb/s)

metropolitan network (total campus bandwidth=2.5 Mb/s)

academic buildings networked at 10/100 Mb/s

Figures for Spring 2004



Consequences for iLabs Consequences for iLabs (and other rich educational resources)(and other rich educational resources)

• Need to deploy educational resources locally

• Solutions engineered in the developed world not necessarily effective in developing countries

• Pedagogy likely to be different in bandwidth starved situations

• Ultimate goal: “native” iLabs. How do we support this?



An example: iLab-Africa An example: iLab-Africa project project

Obafemi Awolowo University

Goals:Goals: To deploy MIT's iLabs To deploy MIT's iLabs

throughout curriculum in throughout curriculum in AfricaAfrica

To support new iLab To support new iLab development in Africadevelopment in Africa

To create opportunities for To create opportunities for internships for MIT and internships for MIT and African students African students

To create a scalable iLab To create a scalable iLab research network in Africaresearch network in Africa


are needed to see this picture.iLabs in Africa: challenges and solutions

Service Broker

Lab Server

Client



• client GUI’s are “fat”• require complex plug-ins• demanding on local resources



Data courtesy of K. Ayodele (OAU)

v. 6.1 graphical applet• 169 kbytes

• <download time> from OAU=79”

v. 6.1 classic applet)• 94 kbytes

• <download time> from OAU=63”

Partial Solution: “thin” clients



Lab Server

iLabs in Africa: challenges




client served from half a world away at MIT


are needed to see this picture.Solution: local Service Solution: local Service BrokerBroker

Instead of here

Instead of here

<download time> at OAU: 22” (graphical), 17” (classic)<download time> at OAU: 22” (graphical), 17” (classic)

student data stored here

applet served from here

Installed SBs at OAU, MUK and UDSM



Lab Server

iLabs in Africa: challenges




professional lab hardware prohibitively expensive



Solution: inexpensive hardwareSolution: inexpensive hardware

Agilent 4155 ~$50K

NI Elvis ~$2K

iLab Mini ~$40



iLabsiLabs in Australia in Australia



Consequences for iLabs Consequences for iLabs (and other rich educational resources)(and other rich educational resources)

• Need to deploy educational resources locally

• Solutions engineered in the developed world not necessarily effective in developing countries

• Pedagogy likely to be different in bandwidth starved situations

• Ultimate goal: “native” iLabs. How do we support this?



iLab: the OpportunitiesiLab: the Opportunities Order of magnitude more laboratories Order of magnitude more laboratories

available to our studentsavailable to our students Unique labs: Unique labs:

Unusual locations, expensive equipment, Unusual locations, expensive equipment, rare materialsrare materials

Rich pedagogical experiences:Rich pedagogical experiences: More lab time to studentsMore lab time to students GUI to lab integrating graphing, GUI to lab integrating graphing,

simulation, collaboration, tutoringsimulation, collaboration, tutoring

Worldwide communities of scholars Worldwide communities of scholars created around labs sharing contentcreated around labs sharing content



iLab: the ChallengesiLab: the Challenges

Developing an iLab from scratch is a Developing an iLab from scratch is a lot of work!lot of work! Great attention needed to user scalabilityGreat attention needed to user scalability Needs to be done by domain specialistNeeds to be done by domain specialist

Managing a broadly shared iLab is also Managing a broadly shared iLab is also a lot of work!a lot of work! Disincentive for owner to share labDisincentive for owner to share lab

Key challenge: Key challenge: iLab ScalabilityiLab Scalability



ConclusionsConclusions

iLabs can enhance science and engineering education in India as well as the rest of the world

iLabs and their educational content can be broadly shared within India and around the world

iLabs provide a path for the developed world to support education in the developing world

iLabs Architecture is a scalable framework to support iLab dissemination around the world

““If You Can’t Come to the Lab… the Lab Will If You Can’t Come to the Lab… the Lab Will Come to You!”Come to You!”

(Earth at 89 GHz; courtesy of J. Grahn, Chalmers U.)

[email protected]

mit ilabs: towards a community of internet accessible laboratories phillip d. long, ph.d. mit mark...

Documents

mit ilabs

ilab gui

ilab project

j exercise

students assessment

mit faculty

exercise due4pm

lab system capacity