WIDE Project Activities

Abstract

WIDE, Widely Integrated Distributed Environment, Project is a research consortium among academia, industry and government, regarding the Internet related technology. WIDE Project has more than four hundreds of active research members from more than 50 private corporations. WIDE Internet was founded to support Internet-related technology researches and developments and to enable the electronical communication among project members.  As it evolved, it played a key role in the growth of the Internet, well before commercial ISPs became popular. WIDE now operates several test projects, such as “WIDE 6bone” to deploy IPv6 technology, “WISH AI3” for satellite internet platform, and “RT-bone” for QoS-aware network. WIDE envisions the future of the “Mobile Computing Environment” and next-generation “native Internet generation” by constructing a “Multi-Gigabit Backbone” such as DWDM.

1. Introduction

  Around the end of 70’s and the beginning of 80’s, the research and development of computer science has stepped into the new stage, that are the interconnection of computers (so called as Internet) and the generalization of computer technology. The computer system has mutated and powered up as followed; main-frame computer, mini-computer, work-station. Also, the system has been toward the distributed architecture.  The UNIX has played the important role for these mutates.  Nowadays, it is generally recognized the role of the computers and the computer networks has to contribute to make rich and effective the human life and society.
  Some members of WIDE project had established and started the JUNET (Japan UNIX NETwork) from 1984. JUNET is aiming to develop and to provide the communication platform among the computers and among the computer science researchers, who were not at the same campus.
  In 80’s, the research and development activities to develop/deploy and to provide the network environment interconnecting computers, that are widely distributed across the WAN, could not obtain good admiring.  However, we has believed and recognized the importance of testbed network to evaluate and proof the developed technology.  The technology can evaluate only when it works in the operating environment, and the operation of these technologies reveal new technocal issues and challenges to be solved.  With the above back ground and motivations, the core member running JUNET has started the WIDE project and has started to developed the WIDE Internet in 1988.
  The JUNET member has researched and evaluated the network protocol architecture proposed by Xerox, that was used for the basic concept of OSI reference model and the system architecture of operating system in the computer.  The discussion among the JUNET members and CSRG (Computer Science Research Group) at UC Berkley has leaded to 4.2BSD founded by DARPA to deploy the TCP/IP network and leaded to the several research activities in the WIDE project.
  The WIDE project has developed the operating testbed, that involves a large number of users, that are not only computer researchers but also non computer researchers.  Involvement of non computer researchers reveal a lot of issues and requirements, that can never occur in the case of involving only computer researchers.  In other words, the purpose of WIDE project is to research and develop the technologies and deploy the network, contributing to the individuals and for society.  To achieve the purpose, the WIDE project involves many events, such as concert using the Internet, Internet Expo, Olympic games.
  The architecture of WIDE project is focusing on the actual operation of WIDE Internet and the global Internet. Without focusing on the actual operation, the WIDE project could not yielded the productive outputs. For examples, IP over X.25, IP over ISDN or UDLR (Uni-Directional Link Routing) using satellite link would not come out without trying the actual operation over these links. The author thinks that the following famous word by Dr. David Clark at INET92, that was held at Kobe, Japan, and was hosted by the WIDE project, represents the mission and tao of the WIDE project.
  “We reject kings, presidents and voting, we believe in rough consensus and running code”
The WIDE project is not a university, laboratory, or individual researcher, but is the researcher’s group. Their (i.e., our) purpose is technological discussions, development and sharing the computer network environment, that uses the research and developed technologies by us or by researchers on the global Internet.

  The WIDE Internet  is ;

(1) Accommodating common traffic
(2) It is research network, is not commercial network
(3) Maintaining scale of network and volume of traffic
(4) Maintained and operated by ourselves
(5) Part of global Internet


The following words represents the structure and operation of the WIDE project.
   “Operations in our right hand, Researches in our left hand”
 
 

2. History of WIDE Project

  Since it’s beginning in 1988, the WIDE project has been advocating that the Internet could deliver a vastly more global world. The WIDE Project is one of the world’s most visionary and ambitious internet projects ever, as it attempts to design the world’s best next-generation computer environment in diverse field such as network education, law/regulation, medical applications, government, the arts and business. The member of the WIDE project includes a broad range of researchers and engineers from universities, corporations and government agencies.  As of January 2000, there are over 100 participating institutions and it’s members is including.
  The followings are the brief mile stones, that the WIDE project experienced.

== begin {before WIDE project} ==

1981 N1 Started
1982 Network installed at Keio University Yagami campus
1984 1985   Rearch Group started (Predecessor of WIDE Project)
1986 1987  University of Tokyo and Tokyo Institute of Technology are connected
== end {before WIDE project} ==

== begin {WIDE project} ==

1988 WIDE project started (August)
1989  1990  1991 1992  1993  1994  1995  1996  1997  1998  1999 2000  INET2000 (Yokohama)
== continue (WIDE project) ==
 
 

3. Structure and Operation of WIDE Project

  The WIDE project has started in 1988. It was a collaboration with ten private companies. In other words, the WIDE project was founded by ten corporations, via technical collaboration with the WIDE project.  This structure has not changed, i.e., each participating company founds the WIDE project research activity via research and development collaboration.  Figure 1 shows the number of corporations, that collaborated with the WIDE project.  Here, the number of institutions in figure 1 does not include the academic institutions.


Figure 1. Increase of WIDE Organizations

3.1 Project Management

We have two meetings, reporting the activities of the WIDE project, in a year. These are in July and in January.  Basically, we agree that the research output via the WIDE project activities are open to the public to contribute the Internet society.
  Figures 2 shows the traditional research and development related with the industry and society.  On the contrary, figure 3 shows that the WIDE project is aming.  The WIDE project tries to collaborate with the industries, focusing on the direct contributions to the society, through the commercial products.  Also, we want to obtain a direct feedback from the society to the academic research groups. The direct relationship and feedback between the society is very important to achieve useful and practical research and developments.


Figure 2. Traditional R&D style                     Figure 3. R&D style with WIDE project

  There was a report by Dr.Chien (NSF, Harvard University), analyzing the research projects in Japan compared with the projects in USA.  Figures 4 and 5 shows the results of the report.


Figure 4. Research programs in USA and in Japan
 


Figure 5. NGI and Internet2/Abeline in USA

As you can see, the WIDE project can be similar to Internet2 and Abeline project in USA. These project accommodate a lot of advanced research institutions from academic and industrial domains.
The WIDE project has a board, which is consisted by 20 voluntary members.  The WIDE board members have the responsibility to an every activity in the WIDE project.  We have monthly board meeting.
The research and development activities usually run based on the Working Group.  SIG, Special Interests Group, forms an Working Group including researchers from different companies, in order to explore the research and development of the particular technological challenge.  Recently, we may form a special project /task team to expedite the research and development mission.  KAME, TAHI and USAGI project would be typical examples, regarding the IP version 6 technology.  KAME is researching and developing the IPv6 protocol stack for various BSD UNIX system.. TAHI is researching and developing the IPv6 system evaluation framework and evaluation software for IP version 6 systems.  USAGI project is researching and developing the IPv6 protocol stack for LINUX.
Below, the snap-shot of Working Groups in January of 2000 is shown.

 
1. InternetCAR
2. IP over IEEE1394
3. IRC
4. LAST (LAbel Switching Technology)
5. Lifeline
6. Diff-Serv
7. Lifelong
8. MAWI (traffic measurement and analysis)
9. MC (Multicast Backbone)
10. Mobile Security (mobsec)
11. MoCA (Security and Certification Authority)
12. Reliable Multicast
13. RT-Bone (Real-Time Backbone)
14. SOI (School on the Internet)
15. TWO (operation of the WIDE Internet)
16. v6 (IP version 6)
17. W4C (Web Cacheing)
19. WISH (UDLR via Satellite link)
20. WT (Internet Phone/FAX)
21. WWW Architecture

3.2 The WIDE Internet Operation

  The WIDE project has developed the operating testbed, that involves a large number of users, that are not only computer researchers but also non computer researchers.  Involvement of non computer researchers reveal a lot of issues and requirements, that can never occur in the case of involving only computer researchers.  In other words, the purpose of WIDE project is to research and develop the technologies and deploy the network, contributing to the individuals and for society.
  The WIDE Internet  is ; Also, the following words represents the structure and operation of the WIDE project.
   “Operations in our right hand, Researches in our left hand”

The importance of item (1) should be obvious. We need a common traffic, otherwise we loose the real requirement for the actually operating network.  The user realize that the WIDE Internet is a research network, therefore there would be some trouble in some cases. However, the WIDE Internet operators (i.e., researchers) has maintained comfortable quality of service to the ordinary users, in the past.  Nowadays, the WIDE Internet accommodates around 100 institutions, therefore the WIDE Internet should satisfy the requirement (3).  Of course, even the operation requires extra overhead for the researchers, the researchers operates the WIDE Internet, while working as the usual researchers or engineers.  Finally, the WIDE project member are trying to participate in the operation of global Internet.  For example, the WIDE project has the responsibility to the operation of “M” root DNS server.  Or, the WIDE project is leading the APNIC and JPNIC regarding the IPv6 address assignment.  Also, the WIDE project is involving a lot of international network operations (e.g., APAN or 6REN).
 

3.3 WIDE Camp Network

 The research and development group of each institution has their own small scale experimental network.  The WIDE Internet interconnects and accommodates these small scale experimental networks to maintain the large scale and heterogeneous network environment.  Now, we have the third network environment, so called as the “campnet”.  The WIDE project has a four day workshop twice in a year.  It is called as “WIDE Camp”.  The WIDE Camp accommodates more than 200 researchers, who bring their own note book computers.  Some researchers and engineers bring in very hot codes and software into the WIDE Camp. The WIDE Camp is a large scale and ad hoc network.  Also, the network accommodate the Internet researchers and engineers, who may be equivalent to the next generation ordinary users.  This means that the WIDE Camp works as the ad hoc testbed for the very hot research and developing technologies.  Figure 6 shows the rough network configuration at the WIDE Camp in March of 1998.


Figure 6. Network Configuration of WIDE Camp in 1998

3.4 NSPIXP (http://jungle.sfc.wide.ad.jp/NSPIXP/)

 NSPIXP (Network Service Provider Internet eXchange Point) is the project designed to research IX (Internet eXchange) technology, an essential internet constructing technology.  It extensively quality-checks and trouble-shoots over commercial ISP connections, utilizing such advanced testing methodologies as NSPIXP-1, NSPIXP-2, NSPIXP-3 and NSPIXP-6.  NSPIXP project accommodates more than 50 service providoers, that would be almost all ISPes and ASPes operated in Japan, in order to test and evaluate the advanced technologies related with IX and operational policies.  In the late 200, the NSPIXP-2 will be distributed among the Tokyo metropolitan area.  This should be the next generation IX architecture, therefore, the NSPIXP project try to implement and operate this architecture.
  The NSPIXP is layer 2 IX, as shown in figure 7. This architecture is the same model as the MAE-EAST/WEST or STAR TAP uses.  Here, CIX and 6TAP at STAR TAP uses the layer 3 IX model.


Figure 7. Layer 2 IX model, applied at NSPIXP

  The following picture illustrates the traffic exchanged at NSPIXP-2, one of the major Internet exchange points in Japan. The statistics is obtained by polling ifInOctets mib variable in every 15 min (every 30min before April 21, 1997, and every 5min after Jan 14, 2000). The figures below are not automatically updated, however, they are updated once a month.
Note that the statistics from Oct 17, 1997 to Nov 07, 1997 was not collected due to snmp problem.
 
 


Figure 8. NSPIXP Daily Traffic

The total traffic reduction effective on the early December, 1997 is due to the bootstrap of JPIX, a commercial based internet exchange in Tokyo. Their traffic statistics is shown here.
However, even with JPIX, the traffic grows more than linear growth. One of the reason is that we operate a News server at NSPIXP2. The server receives about 18GB and sends about 260GB news articles in a day, those numbers yeild to 1.6Mbps and 24Mbps in average respectively. So the server is not considered a big traffic source to NSPIXP2.
   According to the fact that the FDDI interfaces for some major ISPs are getting congested, we've launched its extension program using Gigabit Ethernet switch since December, 1999. Now we use a Cabletron's SSR8600 connected to the Digital Gigaswitches with four FastEthernets huntgroup link. While those switches are on the same subnet, ISPs on the Gigabit Ethernet switch are suggested to close their peering on the switch except for those non-profit ISPs or NICs to avoid congestion over the huntgroup link.
The entire aggregated traffic (red) and the aggregated traffic closed to the Gigabit Ethernet Switch (green) isillustrated as below:  While FDDI is a very nice technology offering dual-homeing, it has only 100Mbps (or 100Mbps each in a FFDT case). We are asking ISPs to removing to the gigabit ethernet switches (or an fastethernet port on them). As the result the traffic exchanged within FDDI gigaswitches are slightly decreasing as illustrated below:
 
 


Figure 9. Traffic at Gigabit Ethernet Switch (SSR)

Figure 10. Traffic at FDDI switch (Gigaswitch)

If we plot the same data for last 3 months with x-axis folded as 24hour, we get the following. You can notify that the peak of traffic happens in after-lunch and mid-night. Latter is thanks to the NTT's "tele-hodai" program, in which the contracted subscribers are not charged for dialup calls to specified numbers (a user can specify up to 2 numbers) between 23:00 and 8:00 localtime. Unlike in U.S.A., flat-rate local call service was not available in Japan. NTT-East and NTT-West, which are the local carriers, have started to offer "Flet's ISDN" service for ISDN customers since mid May. 2000. It is a flat-rate program (for the cost of local loop) which costs 4,500 yen per 1 B channel (64kbps). Other non-intermittent connectivities including CableTV based service as well as xDSL are getting popular in 2000. Are you interested in the weekly traffic pattern? It is illustrated as follows. When the Internet was academic-use only, there was very small traffic in a weekend day compared with one in a working day. No such characteristics has remained. The traffic in a weekend day almost equals to one in a working day.
 
 


Figure 11. NSPIXP Traffic in a day


Figure 12. NSPIXP Traffic in a week

3.5 AI3, Asian Internet Interconnection Initiatives (http://www.ai3.net/)

 This large-scale international project aims at establishing a satellite Internet link covering a whole Asian countries.  With the cooperation of the WISH Working Group, it is contributing to the testing environment for this major “Information Superhighway” upgrade.
A lot of efforts have been made to develop the international Internet environment. In order to achieve a better and efficient information infrastructure around the globe, there are very strong urgent demands in the area of Asia and Pacific for a practical working environment where engineers and researchers in related fields work cooperatively. The WIDE and JSAT(Japan Satellite Systems Inc.) which is a satellite communication company jointly propose a single year testbed construction and a series of research activities using the testbed. The testbed and the activities will be called AI3 (AIII) and will contribute, promote, and encourage any related efforts toward achievement of the international information infrastructure such as Internet, AII, and APII.  Figures 13 and 14 show the overview of AIII network configuration. Using the satellite link, we are going to cover many Asian countries, including Beijin (China).


Figure 13. AI3 link configuration


Figure 14. AIII network configuration


 

4. Research Activities of WIDE Project

4.1 Overview of Research Activities

  The concept of WIDE project research activities are ;

(1) Technology for people

A key objective of the WIDE project is to make work, as much as possible, the vast potential for computer and communication technology to the benefit of people and society.  The WIDE project focuses on creating new computer communication structures.  To achieve these goal, there are many technologies that must be integrated. The research projects covers computer networks, security technologies, operating systems, multimedia information processing, decentralized processing, computer education, system monitoring, et al.  To create the ideal next generation computer communication environment, the WIDE aims at facilitating open and productive communication and discussions among wide range of researchers.
(2) Technology for the future of the Internet
It has been ten years since the WIDE project originated.  During this period, Internet environment evolved so dramatically.  We have been making contribution to construct the Internet both domestically and globally, as one of the leading research group. We also contribute to identify and support promising cooperative research and development opportunities for our member organizations.  Therefore, the technology collected and shared by the various WIDE project research activities will continue to play a significant role in the continuing development and deployment of the Internet technologies.
(3) Cooperation with Social Projects
The WIDE project has been expanding the landscape of the Internet worlds since its dawn.  Participated projects show a diver range including ? emergency services by IAA (Survivor Information System, i.e., “I Am Alive”), Internet education in the “One Hundred Schools” project, disabled accessibility, Paralympics, live internet coverage of Atlanta and Nagano Olympics, concerts, Kabuki plays and other live events.  We will continue to challenge the future of the Internet, to make it as useful as humanly possible.


The researches and developments in the WIDE project are carried on with Working Group, that is a king of SIG (Special Interest Group).  The members of Working Group come from various institutions to collaborate with.  Below, the snap-shot of Working Groups in January of 2000 is shown.

1. InternetCAR
2. IP over IEEE1394
3. IRC
4. LAST (LAbel Switching Technology)
5. Lifeline
6. Diff-Serv
7. Lifelong
8. MAWI (traffic measurement and analysis)
9. MC (Multicast Backbone)
10. Mobile Security (mobsec)
11. MoCA (Security and Certification Authority)
12. Reliable Multicast
13. RT-Bone (Real-Time Backbone)
14. SOI (School on the Internet)
15. TWO (operation of the WIDE Internet)
16. v6 (IP version 6)
17. W4C (Web Cacheing)
19. WISH (UDLR via Satellite link)
20. WT (Internet Phone/FAX)
21. WWW Architecture

4.2 Integration into IP version 6 Platform

4.2.1 Why we need IP version 6

  The Internet, which is based on Internet Protocol (called as IP), had been developed and deployed using the telephone network.  These are the digital leased line and the dial-up links. However, toward the 21st century, the Internet is going to be independent from the telephone network and to be the “native internet environment”.  In the “native internet environment”, the datalinks and physical links are provided for the Internet use, not for the other purpose (e.g., voice communication).  One example would be the use of dark fiber.  The users can define how the (dark) fiber is used and which datalink protocol the (dark) fiber uses.  As the results, the Internet service shall include both telephone service and traditional broadcast service, as just a one of services provided by the Internet. Also, the Internet has to be for everyone, not only for researchers and engineers, to be a reliable and robust information infrastructure supporting human life and industrial and economical activities. IP version 6 (IPv6) is the core protocol for this upcoming next generation information infrastructure.
  The current Internet uses the IP version 4 (IPv4), rather than IP version 6.  IPv4 has 32 bits address length, so that the Internet can accommodate only upto around 4 billion (2**32) equipments. It is less than the total number of humankind over the world. In order to preserve the end-to-end architecture, we need sufficient number of IP addresses. IPv6 has 128 bits of address length, so that really huge amount of IP addresses can be provided.  It is the 2**96 times larger address space than the IPv4 has, and it should be enough address space to preserve the end-to-end architecture model.
 The dramatic and amazing growth and deployment of new services over the Internet is based on the end-to-end architecture model. The end-to-end architecture model can allow flexible and effective introduction and deployment of new technologies. The technological and operational mutations has occurred based on the open network and open software platform.  In the end-to-end architecture model, the network is as simple as possible (sometime called as “stupid network”) and should be transparent. The intelligent and complicated data processing should be performed at the end-station.
 In the mobile internet service using the portable telephone network or in the internet service over the CATV network, it is sometime hard to preserve the end-to-end architecture model, due to the lack of IP addresses assigned by the regional NIC. Actually, in Europe, some potential mobile internet service provider could not obtain sufficient IP address space from the NIC.  Then, the individual users can not enjoy the all applications available over the Internet, e.g., the user can not run the server serving to the global internet. The most serious problem is that such service providers establish the closed (enclosure) service network, that does not use the globally open standard.  This is similar to the traditional business model used in the appliances industry.  The history shows it does not work in the end. With the closed network solution with proprietary technologies, it is very ineffective and needs large amount of cost to introduce and migrate the new services and new technologies, compared with the open network. It becomes very true, when the proprietary system grows.
 The Internet has already experienced two mutations. The first wave is the open global networking from the private open networks.  The second wave is the generalization of Internet technology through the introduction and deployment of WWW (World Wide Web).  The third wave, that we are going to experience, is a broadband and ubiquitous networking.  Every equipment, including the sensors is always connected to the global Internet with high speed connectivity, using wired or wireless links.  Privacy and security is the important issue to be solved.  Anyway, the ubiquitous computing environment with IPv6 technology will create new services and new business models, that we could not realize at this moment. For example, the InternetCar will be a self mobile multi-purpose sensor to sell the monitored information to the certain agent.  IPv6 technology with ubiquitous computing environment expedites the information exchange between each individual and the global Internet, i.e., Internet is for everyone.
 
 

4.2.2 IPv6 at the WIDE Project

 A large portion of researches and developments performed in the WIDE project are based on or assuming the IPv6 platform.  This means that the “WIDE Internet” runs IPv6, as well as IPv4.  Why we need run IPv4 ? This is because we have to “migrate” from the IPv4 network to IPv6 network. During the migration, we have to interconnect IPv6 and IPv4 network, and establish the migration procedure regarding the network operation. In this subsection, the activities regarding (1) core platform technology, (2) applications over IPv6, and (3) Operations for IPv6 are briefly described.

[1] IPv6 Core Platform Research and Development
 We have several special project, that would be rather task force team to develop the IPv6 core technologies. The followings are some of these activities.

(1) KAME Project (http://www.kame.net)

KAME stands KarigoME office, where the project members are working together. KAME project has started from 1998.  The eight engineers from various companies, which are IIJ (Internet Initiative Japan), Hitachi, NEC, Fujitsu, Toshiba, Yokogawa, YDC, Sony, are working together.  Before the KAME project established, the WIDE project has been working on the IPv6 technology at v6 Working Group, that is still active along with KAME project.  KAME project has more intensive and clear mission, than the v6 Working Group has. It is the research and development of IPv6 reference software for the IPv6 community. Before KAME, each institution had implemented their own protocol stack, even in the WIDE project. One of primary purpose of KAME project is unifying the various implementations in WIDE project, as well as in the world.  After the unification within the WIDE project, we have worked with INRIA group in France and NRL group in USA to unify the protocol stack. Finally, most of the protocol stack are come from KAME team, and the KAME team maintains the unified protocol stack.  The KAME project is still running, while improving the quality of software and introducing new functions, such as multi-homing or router renumbering.  Apparently, these are the other purpose of KAME project.
  As a result, many UNIX software packages use the KAME protocol stack to let IPv6 work. These are FreeBSD, BSD-OS, OpenBSD, NetBSD and MacOS-X (Server).  Also, some router vendors, such as Hitachi or Yamaha, uses the KAME protocol stack in their commercial router.
  The followings are the abstract of functional snap-shot for KAME IPv6 protocol stack.
- IPv6 basic functions, e.g., NDP, autoconfiguration
- IPv6 basic and advanced API
- IPSec (IP Security function)
- Basic applications, e.g., sendmail, telnet, web
- Routing protocols, i.e., RIPng, OSPFv3, BGP4+, PIM-SM/DM (multicast)
- Integration with ALTQ (packet scheduler) for quality of service
- Integration with Diff-Serve, RSVP, COPS and Bandwidth Broker
- Multi-Homing function
- DNS server (BIND) and client (resolver) software for IPv6
- Integration with MPLS technology
(2) TAHI Project (http://www.tahi.org)
TAHI project is run by Yokogawa Corporation, the University of Tokyo and with WIDE project. The TAHI project, started from 1998, is also run by the mission oriented task force team.  The project goal is to research and develop the evaluation specification and evaluation tools for IPv6 protocol stack. The output of the project is available as the open source.  The test specification and tools includes the IPSec and routing functions, as well as basic IPv6 specification.  TAHI project is closely collaborating with KAME project and USAGI project, in order to improve both of software quality.
  TAHI project has hosted two international interoperability test workshop, in 1999 and 2000. The first workshop was conjunction with IETF IPng-WG interim meeting, participating from 17 institutions.  The second workshop was conjunction with INET2000, also participating from 17 institutions.  The participating institutions include cisco systems, MicroSoft, 3Com, Ericsson or NOKIA. This means that major vendors for host and routers have participated in the workshop using the TAHI output.
  TAHI project will challenge to investigate into the area of Internet Appliances, that is non-computer equipment with Internet connectivity.  These equipments has some hardware and software limitation, while satisfy the appropriate privacy and security function.
(3) USAGI Project (http://www.linux-ipv6.org)
USAGI stands for UniverSAl playGround for IPv6 ), and focusing on the IPv6 protocol stack for LINUX system.  The USAGI project has started from August in 2000, in order to improve the quality of IPv6 functions in LUNUX.  As same as the KAME project, the engineers from various institutions, such as MPT-CRL, Hitachi, Yokogawa, Toshiba, Keio University, Tohoku University, work together.  The first software package by USAGI project has released Nov.1st in 2000.  The USAGI project will work to develop the LINUX system as the same level of IPv6 quality as BSD system has with the KAME protocol stack.
(4) LIN6 Project
LIN6 stands for “Location Independent Networking with IPv6”.  The goal of this project is focusing on the global mobile environment with authentication of the end node.  The issues, that LIN6 architecture try to solve, are;
- Multi-homing
- Scalable end node authentication in global space
- Separation of routing information and end node identification
The primary implementation for the prototype system has completed, to propose the architecture and specification to the IETF.
(5) Small Group Multicast (MDO6)
There are many small group multicast in the Internet, as well as the multicast including large number of receivers.  In order to provide the small group multicast easily, a new architecture and protocol, that uses the IPv6 option, is proposed to the IETF.
(6) IPv4 internetworking
The WIDE project member implements two approaches to internetworking IPv4 system and applications with IPv6 networks.  One is BIS (Bump-In-the-Stack), and the other is IPv6/IPv4 translator.  BIS, developed by Hitachi Co.Ltd, is the software for MicroSoft Windows 95 and Windows98 system, in order to run the existing applications over the IPv6 network environment.  With the BIS software, you do not need to re-install the existing software for IPv4 MS Windows. BIS is implemented between the kernel and the driver.  BIS has the IPv6/IPv4 translator module within the end node. The other one is IPv6/IPv4 translator, that sits between IPv4 network and IPv6 network. KAME project implemented the NAT-based translator and the application gateway based translator.
[2] IPv6 Applications
  The research and developments performed in the WIDE project is basically assuming the IPv6. We have both middleware and user applications. Here, some middleware software is picked up to be described.

(1) Digital Video Transmission System (DVTS)

(www.sfc.wide.ad.jp/DVTS/,  www.csl.sony.co.jp/person/kjc/software.html)
DVTS is IP-based, high-quality, real-time audio/visual (AV) communications software that uses Digital Video (DV) camcorders and VCR consumer products with IEEE 1394 (Firewire) interfaces. DVTS works with both IPv4 and IPv6 on a FreeBSD system, and is integrated with the ALTQ module to enable fine packet transmission scheduling and DiffServ.
RTP (Real-time Transport Protocol) provides interoperability among systems, and achieves dynamic flow control among hosts. For the highest quality communication, the system consumes over 35 Mbps of network bandwidth. Using RTP, DVTS achieves dynamic flow control and optimizes the DV transmission for the available bandwidth., e.g., DV transmission over 10base-T Ethernet.
DVTS software is already integrated with the following functions; (a) Internet-friendly flow control, (b) Remote control with SSH, (c) Full encrypted (ESP) DV transmission with IPSec, and (d) DV multicast with PIM-SM/DM.
DVTS software is used many events and institutions in global level. Figure 15 and 16 show the abstraction of DVTS architecture.

Figure 15. DVTS architecture


Figure 16. DVTS system configuration

Figure 17 shows the network configuration of the experiment for DV multicast over the WIDE backbone. The IP protocol version is six (i.e., IPv6).  Also, figure 18 shows the equipments used in the experiment.  The WIDE project workshop is multicast to the nation-wide WIDE NOCs, using the IPv6 with PIM-SM (PIM Sparse Mode).
 
 


Figure 17. Network configuration of DV multicast


Figure 18. Equipments for DV multicast











(2) School on the Internet (http://www.sfc.wide.ad.jp/soi/)

The goal of “WIDE University, School on the Internet”, is to establish a new form of Internet-based educational institute and determine the best system for doing this. Many leading Internet experts and professionals are participating in this project with regard to creating the curriculum and specific learning objectives. By November 1999, the archives equivalent of 320 hours of class work were already uploaded publicly to the internet and were utilized by 300 registered students.  The “WIDE University, School on the Internet” offers real time lectures to other universities and institutions internationally over high speed Internet connections.  “WIDE University” envisions new forms of Internet-based learning institutions where people can lean “anywhere, anytime”.
In 1999 and 2000, the real time distance learning among Wisconsin University, Keio University and NAIST has been carried out for a whole of one semester. Figures 19  shows the abstract of this class.


Figure 19. Network configuration of School on the Internet experiment

The SOI system after the Wisconsin University remote class over the Internet accommodates larger participating institutions and links.  The features of the on-going SOI system are as followed.


[3] IPv6 Operations
  As described above, the WIDE project has researched and developed the IPv6 reference software, that include the software tools (e.g., translator or BIS software) to proceed the IPv6 migration from the existing IPv4 environment. Using these software and IPv6-enabled commercial products, some ISPes and research networks are operating the IPv6 technology.  The WIDE project leads the IPv6 network operation with WIDE Internet and NSPIXP.  The WIDE Internet is achieving the migration to the IPv6 network from IPv4 traditional network, while having the co-existing and co-operation of IPv4 and IPv6 networks. It should include the IPv6 address allocation and registry, as well as the IPv6 routing registry.
The IPv6 readiness in the NSPIXP2 has a significant influence to the industry, since many commercial ISPes (almost all major ISPes) use the NSPIXP2 and NSPIXP3 as the primary IX point in Japan.  Also, the WIDE project hosts the NSPIXP6, that is for rather challenging IX targeting the IPv6 technology.
The WIDE project is going to initiate the IPv6 wide area testbed, including many commercial ISPes, contents providers and internet appliances venders, soon.  The primary purpose of this testbed is proofing that the IPv6 technology and network has been already production quality to be ready for commercial operation. Also, the testbed, that includes the residential as well as campus network and enterprise network, tries to explore the possibility of ubiquitous computing environment with IPv6 technology.
 
 


Figure 20. Root DNS in the global Internet

The WIDE project operates one of 13 global root DNS server, which is called as M-root server. The WIDE project has the leadership for the IPv6 DNS deployment in the root DNS servers. In order to deploy the IPv6 network, we have to deploy the DNS system, that is IPv6 ready.  The activities by KAME and USAGI project will deploy the IPv6 capability in the end host.  Therefore, the deployment of DNS system with IPv6 is important role for the global IPv6 network operation.
The address allocation by the regional NIC (e.g., APNIC or RIPE) is also the important deployment for IPv6.  The WIDE project is closely working with APNIC and JPNIC, regarding the IPv6 address allocation.
Finally, some WIDE project members are working with ICANN, that is in charge of IP address allocation and DNS deployment.  This means that the WIDE project is deeply involving into the global IPv6 system deployment activity.
 

4.3 Working Group Activities

 As described above, the research and development activities in the WIDE project is based on Working Group.  Therefore, this subsection gives brief description for each working group in the WIDE project.

1. InternetCAR (http://www.sfc.wide.ad.jp/InternetCar)

Research and development of system which support automobiles to utilize information on the Internet.  Studies in basic technologies from installation and operating system to practical application.
Figure 21 and 22 show the abstract of the Internet car system.
 
 


Figure 21. Internet Car system


Figure 22. Data flow in the InternetCar system

The position of car make extremely precise using the differential GSP technology.  When we have a referenced base station, that can provide the error correcting infomration from the satellite, we can have a single digit cm (centi-meter) precision. Figure 23 shows the experimental result with the prototype InternetCar.  Also, figure 24 shows the plan, that the WIDE project will be going to build to provide the differential GPS service around the WIDE project key NOCs.
 
 


Figure 23. Result of differential GPS position tracing
 


Figure 24. Differential GPS infrastructure plan

2. IP over IEEE1394
The architecture model and protocol for IP service over IEEE1394 is discussed and evaluated.  Now, we have the two implementation around IEEE1394.  One is IP over IEEE1394, the other is IEEE1394 over IP.  The later one is related with DVTS application software described above, and is used in the School on the Internet system to distribute the DV (Digital Video) data with IEEE1394 interface from the digital camera.
3. IRC (Internet Relay Chat)
IRC working group is working on the Internet Relay Chat system.
4. LAST (LAbel Switching Technology)
The WIDE project member has proposed the CSR architecture, that is the root of IP Switch by Ipsilon and of Label Switch Technology, to the IETF and ATM Forum in 1994. The MPLS (MultiProtocol Label Switching) working group has been established at the IETF in 1997.  The WIDE project has started the operation of CSR prototype system around 1996 using the WIDE Internet, and has provided a lot of technical contributions to the IETF. Also, the LAST working group has achieved the integration of IPv6, Diff-Serv, PIM-SM/DM.  Figure 25 gives the typical NOC configuration for the testbed operation.


Figure 25. Typical NOC configuration for CSR test operation

5. Lifeline
Life line working group researches and develops the internet system, that is effectively useful in the case of disaster, such as earthquake. Actually, Japan has experienced serious earthquake and several eruptions, and realized the importance of lifeline capability of Internet system.  Actually, at Kobe earthquake in 1995, people did realize the importance and effectiveness of the Internet system, in order to the information who are alive.  So, the lifeline working group has developed the IAA system, that stands for “I Am Alive”.  IAA system provide the information, who are alive, to the world using the Internet technology.  The lifeline working group has started the disaster drill from 1996. We have disaster drill twice in a year, that is working with many institutions including the local governments.  Also, the developed system will be introduced to the case of disaster, such as eruptions in Japan.
6. Diff-Serv
Diff-Serv working group has established in 2000, in order to research/develop/deploy the differentiated service technologies and network. The working group will establish the experimental network, that is Diff-serv aware.
7. Lifelong
The internet system and computer system should support the people’s activities during his whole life.  For the existing computer system, we have to change many information (e.g., email addresses or account), during our life.  Lifelong working group is researching the computer system, that is effectively useful during our whole life.
8. MAWI (traffic measurement and analysis)
The Internet system runs based on the cooperation of networks and end-nodes. Also, it is very hard to control and manage the network effectively.  In order to control and manage the Internet system well, it is realized that the traffic measurement and analysis is an important issue and research topics.  For example, in the USA, there are CAIDA project, that researches and develops the traffic measurement tools for the Internet.  The goal of the MAWI working group would be similar to the CAIDA.  The MAWI working group works with the WIDE Internet.
9. MC (Multicast Backbone)
The WIDE Internet is a part of global M-Bone (Multicast Backbone).  The WIDE Internet includes the satellite link with UDRL, as well as usual wired links.
10. Mobile Security (mobsec)
The mobile system needs the security function, especially for mobile user belonging to certain private company.  The mobile user, accessing the private network from the Internet space, has to work with firewall router, that is at the boarder between the Internet and private network segment.  Mobsec working group researches and develops the security functions and protocols for mobile host.
11. MoCA (Security and Certification Authority)
MoCA working group researches and develops the security related technologies, such as authentication. Also, this working group establishes and deploys the security related infrastructure, such as Certificate Authority.  The WIDE project operates its own certification authority, called MoCA.
12. Reliable Multicast
The multicast service is going to step forward to provide the error free packet delivery  to every receivers.  This is reliable multicast, where the conventional multicast provides just a best effort packet delivery to the receivers.  Working group is clarifying the system and technological requirements and approaches to achieve the reliable multicast service. As discussed at IRTF and IETF, it will be a building block based architecture. Some engineers have worked to develop the FEC (Forward Error Correction) function, that should be recognized as one of the key technological element in the reliable multicast framework..
13. RT-Bone (Real Time Backbone)
RT-Bone working group researches, develops and deploy the network, that supporting the realtime data transport over the Internet. The working group activity is going to be merged with the Diff-Serv working group.
14. SOI (School on the Internet)
As described in the section 4.2, the WIDE project has been developing and deploying the SOI infrastructure and operation, that is open to everyone connected to the Internet. The SOI activity is global, so that the lectures are multicast from oversea and to oversea.
15. TWO (operation of the WIDE Internet)
Two working group is in charge of the operation and management of the WIDE Internet. Since the WIDE Internet serves not only to the WIDE members but also to the other ordinary people who are connected to the WIDE Internet, the operational quality has to be a production quality.  The uniqueness of the WIDE Internet operation by the two working group is the introduce of the new technologies developed by the researchers, while keeping the production quality operation.
16. v6 (IP version 6)
v6 working group would be a kind of mother working group for the several task force oriented projects, such as KAME ands USAGI projects. The v6 working group works for the wide range of researches, developments and deployment of the IP version 6 technologies.
17. W4C (Web Cacheing)
The growth of the traffic by the WWW system became the dominant, regarding the traffic volume in the Internet.  In order to have efficient and comfortable Web access, a cache technologies should be established.  W4C working group researches, develops and deploy the Web cache system, in order to improve the Web access performance, such as content delivery network technologies.
19. WISH (UDLR via Satellite link)
WISH working group works on the research and development of the Internet communication platform using the satellite system. This working group is a technological core for the AI3 project.  Also, one of the significant research outputs is the UDLR (UniDirectional Link Routing) system architecture and protocol.  The UDLR has been proposed to the IETF to established the working group.  The working group has been managed with INRIA (France) to have the global standard for UDLR system.  Also, this working group is working with INRIA, in order to make sure the interoperability between the two UDLR system by the WIDE and INRIA.
The WISH working group has deployed the network shown in figure 26 and figure 27.


Figure 26. Network configuration of WISH backbone
 


Figure 27. System configuration of NOC for the WISH system

20. WT (Internet Phone/FAX)
Conventional services using the POTS infrastructure should be important to integrate all services over the Internet, as well as internetworking with the POTS network.  Voice over IP and Facsimili service over the Internet are the target of this working group.  The voice over IP system has been developed and the experimental system across the Pacific ocean had been deployed.
The facsimili service over the Internet is called as the “InternetFAX”.  This working group has researched and developed the architecture and protocols for the InternetFAX system. The output of the activities are proposed to the IETF and formed the working group, named as “ifax” working group.  The member of this working group contribute to the IETF, and publish the standard track RFC.  Also, the ifax working group of IETF has worked with the ITU-T, that standardizes the technological standard of telecommunication system, and the ITU-T has adopted the specification defined by the IETF.  Figure 28 shows the system diagram of ifax system.
 
 


Figure 28. System diagram of ifax system


 

5. Operation of WIDE Research Network


The figure 29 shows the network topology of the WIDE Internet backbone.
 
 


Figure 29. The network topology of the WIDE Internet backbone









[Domestic Wide NOC]


[Internatoinal Wide NOC]


[Cooperating domestic network]

The WIDE Internet uses wide variety of datalinks, which are satellite link, dark fiber, ATM, SDH, Ethernet, FDDI et.al.,  Also, as described in the previous sectoin(3.4), the WIDE Internet includes three IX points. These are NSPIXP2, NSPIXP3 and NSPIXP6.
  The "two" working group has the responsibility to operate the WIDE Internet.  The WIDE Internet is a research and educational network, that introduces the latest technologies and services developed by the WIDE project members or by the other Internet reseacheres in the world.  In the WIDE Internet, all technologies are operated and evaluated with the integrated environment.  Also, the WIDE Internet carries the so called common traffic, that is the IP packet service for ordinary people connected to the WIDE Internet.  For example, Keio University uses the WIDE Internet for ordinary Internet access.
  The operation of the WIDE Internet is carried on by the WIDE project members, in order to realize the real Internet operation and to realize the technological and operational issues from the actual network operation.  Also, many technological issues has recognized through the WIDE Internet operation.   This represents the "research in out right hand, and the operation in our left hand".  We believe that this is the key why the WIDE project can deliver a lot of productive research outputs.
  The latest WIDE Internet uses both IP version 4 and IP version 6, simultaneously.  This means that the WIDE Internet evaluates the IP version 6 technologies, as well as the transition and migration mechanism from the IP version 4 to IP version 6 environment.  There are IPv6 and IPv4 translation routers, in order to internetwork IP version 4 network and IP version 6 network smoothly.
  The WIDE Internet is, of course, internetworked with the internation research and education networks, such as APAN or Abeline.  The networking with the international research network is described in the next section.
 

6. International Relationship

6.1 Standardization

  The WIDE project has been contributed to the IETF, in the various fields and aspects.  The following activities would be some examples.
 

6.2 International Research Network Operation

   The WIDE project has cooprrated with many international network and projects.  The followings are some recent examples, that the WIDE project are leading or involving.
 

6.3 International Internet Governance and Operation

  The WIDE project has been involving a lot of activities regarding the Internet governance.  The followings are the examples of the activities, related with the WIDE project.
  Some WIDE project members have deeply involved in the international internet governance and operation.  Also, many WIDE project member have involved in the actual and daily operation regarding the international (as well as domestic) internet governence.  The WIDE members think that it is a responsibility of them for the people enjoying the Internet.
 
 

7. Conclusion

  This memo describes the brief history, structure and operation of the WIDE project.  The WIDE Project is a research consortium among academia, industry and government, regarding the Internet related technology.   WIDE Internet was founded to support Internet-related technology researches and developments and has continued to forward to the next generation internet and to controibute to the future information society.    WIDE now operates several test projects, such as “WIDE 6bone” and "KAME/TAHI/USAGI projects" to deploy IPv6 technology,   “WISH AI3” for satellite internet platform, and “RT-bone” for QoS-aware network. WIDE envisions the future of the “Mobile Computing Environment” and next-generation “native Internet generation” by constructing a “Multi-Gigabit Backbone” such as DWDM.
 
 

Appendices

Appdendix 1.  Sponsor's List


                                                                          Total 90 organizations
                                                                As of January 2000 alphabetical order
 

Appendix 2. WIDE Internet Operational Partners


                                                                          Total 35 organizations
                                                                As of January 2000 alphabetical order
 

Author ;

 Hiroshi Esaki, Ph.D,
   WIDE Project Board Member,
   The University of Tokyo,
   TEL: +81-3-5684-7303,  FAX: +81-3-5684-7775
   Email: hiroshi@wide.ad.jp