179NISHINAGA Nozomu et al.

1 Introduction

With the continuing development of com-puter networks, distance learning has come toplay an increasingly significant role in educa-tion. The advent of the broadband Internet inparticular is making it possible to send/receivetext, voice, and image data anywhere in theworld at lower and lower cost. In this context,to make distance learning more effective, wemust develop educational software specifical-ly for use on IP networks. Since nonverbalcommunication is an integral part of any lan-guage, we cannot underestimate the impor-tance of achieving synchronous (real-time)and symmetric (two-way) communications indistance learning［1］［2］. To date, a variety ofresearch papers have been published in thisfield, including a proposal for a multimediaservice framework in an IP/ATM environment

for distance business operations and distancelearning［3］and a report on demonstrationexperiments at relatively low transmissionrates［4］.

The principal members of this projectinclude Hokkaido University, Waseda Univer-sity, Stanford University, the University ofAlaska, and the National Institute of Informa-tion and Communications Technology(NICT). In this project we connected universi-ties in Japan and the United States by a high-speed international IP network. Using a high-definition videoconferencing system with atransmission rate of tens of megabits per sec-ond (Mbps) and a web-based educationalapplication program, we have demonstratedthe feasibility of information and communica-tions technology (ICT) and the usefulness ofinternational synchronous two-way distancelearning. The objective of this project is to

5-7 International Distance Learning Experi-ment Using JGNⅡ

NISHINAGA Nozomu, NISHIHORI Yuri, NAGAOKA Keizo, COLLIER-SANUKI Yoko, AOKI Miho, YAMAMOTO Yuichi, HARADA Masahiro,and TANAKA Kenji

This paper reports cooperative experiment undertaken by Hokkaido University in Japan andthe University of Alaska Fairbanks (UAF) in the United States. The experiment was conducted inJanuary, 2005. A classroom in the IT Education Hall at Hokkaido University, Japan and the Dis-covery Laboratory of the Arctic Region Supercomputing Center at the University of Alaska, Fair-banks were connected via the Internet and a domestic next-generation Internet research andeducational network infrastructure called JGN2 that was operated by the National Institute ofInformation and Communications Technology (NICT) in Japan. A high definition video confer-encing system, a chat system, and a web-based voting system were used in the experiment topromote the establishment of an effective learning environment, particularly for cross-culturallearning.

KeywordsCollaborative learning, Cross-cultural learning, synchronous, Symmetric, High quality video conferencing

180 Journal of the National Institute of Information and Communications Technology Vol.52 Nos.3/4 2005

develop a system or form of education thatmakes use of multimedia and ICT, and todemonstrate the effectiveness of such a systemor format. Through the development of such asystem, with an emphasis on basic media liter-acy and communication skills in foreign lan-guages, we intend specifically to improve for-eign-language education. To create a so-called“virtual borderless classroom”, the classroomatmosphere must be reproduced with highfidelity. To this end, we developed an Internet-based videoconferencing system featuringimage quality comparable to that seen in tele-vision broadcasting. Such high-quality trans-mission allows for the use of an over-100-inchscreen, enabling the display of life-size imagesbi-directionally and in real time (synchro-nous).

In our advanced information society, Eng-lish-language skills (particularly writingskills) are considered essential as a basis ofcommunication. Conversational skills areneeded for discussions and debates, but it iswidely recognized that it is difficult for Japan-ese students to acquire adequate proficiency inspeaking. To address this challenge, wedevised a system to develop debating skillsthat allows real-time or quasi-real-time writ-ing-based communications over networks. Bycombining this system and high-quality datatransmission, we are working to establish apractical multimedia communications technol-ogy that can convey not only voice and imagedata, but also text information and an overallclassroom atmosphere.

Through three demonstration experimentsin the past four years, our distance learningproject has shown its effectiveness not only inenhancing learning but also in motivating stu-dents to learn foreign languages［5］［6］. In anexperiment carried out in 2003［6］, we con-nected Hokkaido University and Stanford Uni-versity by a high-speed leased line. To imple-ment a videoconferencing system, we devel-oped application software that could send bothconsumer standard digital video images andhigh-definition images (720/30p) using an IPprotocol.

In this paper, we describe a demonstrationexperiment on international distance learningconducted at Hokkaido University and theUniversity of Alaska Fairbanks (UAF) in Jan-uary 2005. We connected a classroom in theIT Education Hall at Hokkaido University andthe Discovery Laboratory located in the UAFArctic Region Supercomputing Center(ARSC), via the Internet and NICT’s JGNⅡresearch and development network. We willdescribe this network and the tools developedfor this experiment in detail below. Due tolimitations of space, we will not discuss theexperiment results in terms of educationaltechnology, which is specifically addressed inanother document［7］.

2 Overview of experimental class

Hokkaido University conducted thisexperiment as an element of computer-assistedEnglish learning in an English writing coursein its second semester of 2004. The class wasattended by 24 freshmen in the Faculty ofEngineering enrolled in this course. The UAF,on the other hand, participated in the experi-ment as part of its regular course entitled “JPN293 Virtual Study Abroad: Language, Culture,and Geography of Japan”. Participants includ-ed 11 students studying Japanese. The majorsignificance of this experiment lies in the reci-procal learning of students’ native languagesand the rich cross-cultural exchange (as cross-cultural informant) that is the inevitable result.This experiment took place from 1:00 p.m. to2:00 p.m. on January 18, 2005 in Japan, corre-sponding to the hours of 7:00 p.m. to 8:00p.m. on January 17, 2005 in Alaska. Althoughthere was a difference of 18 hours in the timeszones of the participants, both universitiescould hold this event within their regular classhours. Further, despite the several thousandkilometers that separated these two locations,students were able to share the atmosphere ofa “virtual borderless classroom” through dis-cussions via the screens. Figure 1 shows theexperimental classroom setup at both universi-ties.

181NISHINAGA Nozomu et al.

This class consisted of (1) the exchange ofopinions with the use of the multi-participant“Chat’n’Debate” chat system (described in thenext section); (2) discussions with the use ofthe “Culture Box” questionnaire system inballot form (also described in the next sec-tion); and (3) an object lesson. This class wasconducted in English, as the UAF studentswere only beginning their study of Japanese.Nevertheless, some students communicated inJapanese using Roman characters. Two over-150-inch screens were set at the front of theclassroom and used one to display the text ofthe discussion conducted through the chat sys-tem. This display played a very important rolein bi-directional communication. Moreover,this mechanism helped the students feel “con-nected”. While it is nearly impossible to makesense of multiple simultaneous student com-ments, Chat’n’Debate allowed participants toview several comments at once.

Culture Box was also helpful in encourag-ing the collaborative exchange of informationbetween students at the two universities. Thissystem allows students to answer question-naires in ballot form. They can also enter com-ments on the questionnaire results. Changes inthe number of votes are instantaneouslyreflected on the screen. This feature helpedstimulate discussion even further.

Finally, students gained firsthand experi-ence of the differences between Japan and the

U.S.—for example in the physical distancebetween speakers in conversation and in dif-ferences in the students’ personal effects. Inthe former case, students noticed a significantdifference between their cultures in terms ofthe physical distance between speakers, andwere surprised at the further cultural differ-ences when they showed each other the per-sonal belongings they had at hand.

3 Network and tools used inexperiment

3.1 High-definition videoconferencingsystem

NICT and Tokyo Electron Ltd. have beenjointly developing videoconferencing systemsthat can send high-quality images and voice,in general anticipation of the needs of thecoming ubiquitous society. These organiza-tions developed a consumer high-definitiondigital video (HDV) system at required band-widths of either 20 Mbps or 30 Mbps. Thiswas used as a videoconferencing system in thedemonstration experiments on distance learn-ing. Figure 2 shows the configuration of thisHDV transmission system.

A main feature of this system is its signifi-cantly low cost relative to conventional HDtransmission systems. Conventional systemsrequire an expensive HD camera and anencoder/decoder for use in compression and

Fig.1 Classroom in the IT Education Hall, Hokkaido University (right) and Discovery Laboratory,UAF (left)

182 Journal of the National Institute of Information and Communications Technology Vol.52 Nos.3/4 2005

decompression of image and voice data. How-ever, the latest consumer HDV cameras fea-ture a built-in encoder and are available at150,000 to 400,000 yen. These cameras canperform decoding through the use of a con-sumer D-VHS tape recorder or softwaredecoder installed on a notebook PC (althoughperformance depends on the computing capac-ity). In this experiment, we used high-resolu-tion images in 1080i format. A high-definitiontransmission system has the following advan-tages when used in videoconferencing: (1)images are of sufficiently high definition to bedisplayed on an over-100-inch screen; (2) thehorizontal angle of view is so wide thatimages of the entire classroom can be trans-mitted without the need to move the camera;and (3) even facial expressions (of many stu-dents) can be conveyed. On the other hand,one disadvantage to the system is seen in thedelay of approximately one second beforeinput images are output as a result of the inter-frame compression used to achieve high com-pression. However, this delay may be reducedby an increase in the bandwidth of future com-munications network. During this experimen-tal class, Japanese and American studentsshowed each other what they had in theirpockets, and noted differences in the types ofdaily necessities and the amount of moneycarried from country to country (i.e., from cul-

ture to culture). They also exchanged views orquestions about these personal effects usingthe chat system described below. High-qualityimages allowed them to share even the small-est items, which fostered both improvedunderstanding and the motivation to learnmore.

3.2 Chat’n’DebateChat’n’Debate is a web-based chat system

implemented as a Java applet. This tool wasdeveloped at Hokkaido University. Figure 3shows a screen shot of this chat system.

In our advanced information society, Eng-lish-language skills (particularly writingskills) are considered essential as a basis oftoday’s communication. We developedChat’n’Debate as a debating tool enabling stu-dents to exchange opinions in real time or inquasi-real time. In distance learning sessions,this chat system provides each student withthe opportunity to join the discussion on anequal basis.

3.3 Culture BoxCulture Box is a Web-based voting system

written in PHP that features a comment entryfunction. This tool was also developed atHokkaido University. Figure 4 shows a screenshot of the Culture Box. Participants castvotes for answer options, and can also add

Fig.2 IP transmission system using consumer HDV camera (1080i)

183NISHINAGA Nozomu et al.

comments to their answers. During this exper-imental class, for example, one student asked,“How do Japanese university students addresstheir fathers?” After voting, students com-pared answers (by Japanese students) withguesses (by American students). There wasalso a question directed to American students:“How do American university studentsaddress their mothers?” They then similarlycompared their answers and guesses. The stu-dents then had a lively discussion by enteringcomments and responses on the results.

3.4 Network configurationFigure 5 shows the network configuration

used in this experiment. This network wasdesigned with the following three points inmind: • Use of the Internet in a trans-Pacific connec-

tion • Use of JGNⅡ for local connection in Japan• Enabling communication of data to multiple

points, in addition to direct point-to-pointconnection between the two countries

We connected the domestic segments byJGNⅡ and the remaining segments by theInternet. Since this experiment coincided with

Fig.3 Chat’n’Debate screen shot

Fig.4 Culture Box screen shot

184 Journal of the National Institute of Information and Communications Technology Vol.52 Nos.3/4 2005

the JGNⅡ Symposium held in Osaka, weshowed the class at the symposium sites inOsaka and Tokyo (NICT Koganei) as ademonstration of the technology involved.Dozens of visitors viewed the class at the twosites.

Figure 6 shows the network configurationin detail. All sites in Japan were connected byvirtual LAN within the same subnet. This sub-net was connected to the Internet via a routerinstalled at JGNⅡ. We used a shared 100-Mbps network to connect this router andHokkaido University.

We used a U.S. Internet service to connectHokkaido University and the Discovery Labo-ratory in the ARSC of the UAF. According toa preliminary experiment on the networkusing UDP, we were able to send data fromthe JGN segment at NICT Koganei to ARSCat 70 Mbps or higher with the loss of only afew packets. This rate was possible principallybecause we conducted this experiment justbefore the end of the winter vacation, andtherefore traffic was relatively low within thecampus networks. The Discovery Laboratoryhas an immersive display system that consistsof several over-150-inch rear-projection dis-plays. They used this system to display imagesand text sent from Hokkaido University,

enabling the students to view life-size imagesin the course of the experiment.

There are several methods of distributingimages of a distance learning session to multi-ple locations. We selected IP unicast in thisexperiment because the experimental networkfeatured an Internet segment in which multi-casting (usually more efficient for such pur-poses) was difficult. Our high-definitionvideoconferencing system consumes a band-width of about 30 Mbps on the IP layer. Tosend image data from Hokkaido University tothree locations using unicast, the system willrequire a bandwidth of approximately 90Mbps, which is comparable to the networkcapacity between Hokkaido University and theJGNⅡ router. To use bandwidth most effi-ciently, we adopted a star topology consistingof distribution servers located in Japan. Figure6 shows the network configuration in detail.

During the actual experiment, data trans-mission was affected by traffic associated withthe JGNⅡ Symposium. It became difficult tosend stable images of the U.S. side from thedistribution server in Tokyo to Hokkaido Uni-versity. Therefore, we sent this data directlyfrom UAF to Hokkaido University. At thattime, the data stream was approximately 60Mbps from UAF to Japan. We can conclude

Fig.5 Network configuration

185NISHINAGA Nozomu et al.

from these results that it is possible for otheruniversities to hold a high-definition video-conference using their ordinary campusLANs, provided the network connection con-ditions are good.

4 Conclusions

In this paper we described a demonstrationexperiment on international distance learningconducted at Hokkaido University and theUniversity of Alaska Fairbanks in January2005. We connected a classroom in the ITEducation Hall at Hokkaido University cam-pus and the Discovery Laboratory located inthe UAF ARSC using the Internet and NICT’sJGNⅡ research and development network. We

presented the objectives and outline of thisexperiment as well as the tools we used. As anext step, we intend to conduct experimentalclasses in which participants at three or morelocations (or, in three or more different coun-tries) are simultaneously connected, learningeach other’s languages and sharing theirrespective cultures.

Acknowledgements

We would like to express our gratitude tothe staff of ARSC for their support and coop-eration in this project. Special thanks to Dr.Frank Williams, Ms. Barbara Horner-Miller,Ms. Virginia Bedford, Mr. Bob Huebert, Mr.Paul Mercer, and Mr. Nathan Bills.

Fig.6 Details of network configuration

186 Journal of the National Institute of Information and Communications Technology Vol.52 Nos.3/4 2005

References01 K.Nagaoka, H.Kato, H.Fukuda, and Y.Nishihori, “Collaborative Distance Learning Using International

Broadband Network between NIME and Stanford”, Proceedings of the 18th Conference of Japan Soci-

ety for Educational Technology 2002, E1-208-3, 855-856.

02 Y.Nishihori, K.Nagaoka, and D.Harris, “Evaluating the Impact of the Gigabit Network on Learning”, Pro-

ceedings of the 19th Conference of Japan Society for Educational Technology 2003, E2a-209-9, 949-

950.

03 D.Fernandez, A.B.Garcia, D.Larrabeiti, A.Azcorra, P.Pacyna, and Z.Papir, “Multimedia Services for Dis-

tant work and education in an IP/ATM Environment”, IEEE Multimedia, 8(3), 68-77.

04 T.Kishida, K.Maeda, and E.Kohno, “Realization of Active Collaboration in Distance-Learning on the

Internet”, Information and Systems in Education, 2, 77-84.

05 N.Nishinaga, Y.Nishihori, K.Nagaoka, K.Tanaka, S.Okabe, Y.Yamamoto, Y.Ichioka, L.Leifer, and D.Har-

ris, “Cross-Cultural Learning Experiments through the Utilization of the Transpacific IP Network”, Pro-

ceedings of PTC2004.

00 (http://www.ptc.org/PTC2004/program/public/ tuesday/t23/index.html)

06 N.Nishinaga, Y.Nishihori, K.Nagaoka, D.Harris, S.Okabe, Y.Yamamoto, and K.Tanaka, “Enabling a

Cross-Cultural Collaborative Community - Networking Technologies to Form Meaningful Environments

for Higher Education”, Proceedings of ITHET2004, IEEE Catalog Number:04EX898C (ISBN: 0-7803-

8597-7).

07 N.Nishinaga, Y.Nishihori, Y.Collier-Sanuki, K.Nagaoka, M.Aoki, Y.Yamamoto, M.Harada, and K.Tanaka,

“Cross-cultural Learning Experiments through the Internet”, Proceedings of ITHET2005, (ISBN: 0-7803-

9141-1/05).

187NISHINAGA Nozomu et al.

NISHINAGA Nozomu, Ph.D.

Senior Researcher, Smart SatelliteTechnology Group, Wireless Communi-cations Department

Satellite Communications

NAGAOKA Keizo, Dr. Eng.

Professor, Graduate School of HumanSciences

Educational Technology

NISHIHORI Yuri

Vice Director, Professor, InformationInitiative Center, Hokkaido Universtiy /Professor, Graduate School of Interna-tional Media & Communication,Hokkaido University

Applied Linguistics, Media LanguageEducation, English Teaching Method-ology, Collaborative Learning

COLLIER-SANUKI, Yoko, Ph.D.

Director, Japanese Studies, Universityof Alaska Fairbanks

Linguistics, Language Education

YAMAMOTO Yuichi, Dr.Sci.

Assistant, Hokkaido University Infor-mation Initiative Center

Information and Multimedia Education

AOKI Miho

Assistant Professor, Joint Faculty, Arc-tic Region Supercomputing Center,University of Alaska Fairbanks

3D CG, Digital Art

TANAKA Kenji, Ph.D.

Senior Researcher, InternationalAlliance Division, Strategic PlanningDepartment

Image Coding

HARADA Masahiro

Tokyo Electron Ltd

High-quality Video Image Transmis-sion for Internet