Aurora Project

Proposal for a campus-wide hypermedia information system to promote the Electronic Learning Environment and capitalize on the University's investment in CWRUnet

by Ronald S. Ryan, Head of Integrated Library Systems
Library Information Technologies Department
December 16, 1993

Au-ro-ra (aw-raw-ra) n. (Myth.) the Roman goddess of the dawn. aurora n. the dawn.

Forward:

I do not profess to be highly versed in the area of Network Architecture. My expertise lies in Information Delivery Systems and Interface Design. As is the majority of Information System's targeted customer base, I am generally a consumer of networked information. As such, the only time I need to concern myself with the underlying technology of the network is when it is malfunctioning. This same ideal holds true for consumers of transmitted and telephonic information every time they turn on their television or answer their telephone. When functioning properly, the delivery mechanics are completely transparent, as they should be. The following proposal promotes the development of an interface design to deliver to consumers, information in such a way as to integrate the many disparate technologies and development efforts that have been the basis of the Electronic Learning Environment and to make use and retrieval as commonplace as other more familiar information delivery systems such as radio, television, telephones and books. Whenever it becomes necessary to provide technical details concerning the network architecture, I have provided the information correctly to the best of my ability, but recommend that some portions be reviewed by staff members of Information Network Services and other Information Services personnel; both for technical verification and the overall project's integration ability into the future campus network plans as they currently exist.

Introduction

Electronic information resources continue to become available at an explosive rate. This fact comes as no surprise to anyone observing current events in this country. With the advent of the proposed Information Superhighway and the coming involvement of multibillion dollar corporations, this growth trend will probably continue at its exponential rate. Yet today, just as in the early days of the Internet, the most common complaint remains the same: "I know the information is out there, but how can I find what I need?"

While searching tools and information interfaces have existed for years, they suffer from a common deficiency. They require a significant investment in time and effort to learn to use effectively. Today's professionals seldom have the time nor the desire to invest in learning the computing technology necessary to make effective use of the rich environment of electronic resources while still maintaining their selected careers. It is thus desirable to produce an interface to this cornucopia of information that can be used with a minimal amount of effort and training. Information professionals throughout the world have wrestled with this dilemma for years, and while not yet solving the problem, have recently achieved a significant breakthrough towards that goal.

What is Mosaic?

An excellent description of Mosaic was published in Perspectives, SHOW INDUSTRY AND TECHNOLOGY NEWS, January 1994, Volume 3, Number 1

"Internet's 'Killer Application'
by Dr. Tony Rutkowski,
Vice President Internet Society;
Director of Technology Assessment, Sprint;
Research Associate, MIT

In the information industry over the past few years, strategists and consultants have used the term 'killer application' to describe the situation where business sectors begin to massively scale in size when a single magic information service is found that is so compelling or captivating that consumers simply cannot do without it. During the past few weeks, the Internet world appears to be witnessing the emergence of a 'killer app' that is so significant that my MIT colleagues refer to it as a new paradigm.

Mosaic and the World Wide Web

On November 12, 1993, a digital cannon was fired that was felt around the world. The National Center for Supercomputing Applications (NCSA), located at the University of Illinois at Urbana-Champaign, is one of the premier high-performance information networking and computing research establishments in the world. On that date, NCSA placed on its Internet public server the software to implement good, stable versions of Mosaic on all commonly used computer platforms; Microsoft Windows, Macintosh and Unix workstations.

Mosaic in technical terms is known as a 'distributed information browsing client.' It is, in fact, the most elegant, powerful, intuitive, aesthetically beautiful knowledge tool ever created. All in one screen, it provides dozens of lovely fonts and formatted text against a light gray background, with bright underlined blue phrases serving as hypertext links to greater subject detail or additional screens. It is fully multimedia with embedded graphics, pictures, audio, and even motion picture clips. It is optimized to present information for a special Internet-based service known as the World Wide Web (WWW), but it also seamlessly and simply links to most other common Internet-based information services - Gopher, WAIS, FTP, Telnet, NetNews, and directory services distributed across the world on more than 2 million connected computers. Mosaic can even be used with files on one's own computer on disks or CD-ROM.

Even in its earlier beta forms for Unix workstations, Mosaic was causing the use of the Internet to scale at enormous rates for the WWW. Annual traffic growth for the WWW service alone this year exceeds 300,000 percent! Nothing in the history of human electronic communication has scaled so massively - not only in terms of traffic, but also in geographical ubiquity. This is a revolution that is unfolding around the world as servers have spread to 26 countries and encompass virtually every institutional and professional field. What happens now is anyone's guess, but we know the answer is going to be measured in exponential orders of magnitude. Every company and institution, government and commercial - including the publishing and telecommunication industries - is certain to be significantly affected as the demand for these services grows.

Assisting the user are a variety of services themselves Mosaic-WWW based. These include server directory services by location, by subject, and by word search tool. There are discussion groups, a newsletter, and a daily "What's New" bulletin that you can establish as your first screen, From there, you just point and click on words or phrases to navigate.

When I first pulled the Mosaic software for Windows onto my PC to use it, my seven-year-old son was at my side. He asked if he could 'navigate'. I watched as he visited places around the world, including a particularly fascinating collection of art works at the first Fractal Art Museum operated by the University of Rennes outside Paris. After viewing one particularly lovely Mandelbrot picture, he simply clicked on it to bring it over to our PC to display full size on the screen and be stored.

It's worth giving significant credit for this development also to Tim Berners-Lee at CERN, the world's largest high-energy physics research facility straddling the Swiss-French border outside Geneva. Tim conceived the World Wide Web in early 1989 and first implemented it in late 1991 on NeXT workstations. As it began to be used on other computer platforms, its use began to scale. Most of the really dramatic growth only began this year, especially in the past few months.

Mosaic, like many of the significant new network-based services, makes use of an important architecture known as 'client-server' technology. Although it may sound complicated, the basics are simple, making use of the substantial intelligence and power that PCs and workstations now possess. A client is just an information process - often including a display - that runs on a computer that is intended to interact automatically with a related service-oriented information process on some other computer known as a server. These are generally "open" architectures in the sense that multiple clients can usually work with a server anywhere across a network that may vary in scale from a Local Area Network to a global Meta-Network like the Internet.
Where can you obtain Mosaic? Simply FTP to ftp.ncsa.uiuc.edu and browse through their directories that are oriented around Unix workstations, MS Windows, and Macs. Go to the platform of your choice and get the Mosaic program as well as the related display or audio drivers. These are typically compressed."

Advantages Over Other Interface Designs

Several initiatives to distribute information electronically have existed for years and their development and number continue to expand. Mosaic however, offers advantages over other designs to date by incorporating many of the desirable features into one package, including

Multiple platform support: Windows, Macintosh, Unix.
Hyperlinked access to documents, pages and more detailed subject clarification: This allows information to `appear' to be in many locations, addressing the fact that users will search and navigate by different means and logical assumptions, yet still be able to reach the information source. This is a great advantage over the `single path' tree structure found in many interfaces.
Familiar `magazine-like' presentation: Mosaic delivers information in a manner more familiar to consumers less experienced in computer use and Internet navigation. Text-only interfaces are lacking to users that have for decades, consumed information enhanced with the color graphics and video of more tradition information delivery systems.
Backwards Compatibility: Mosaic allows integration of yesterday's most popular interface, `gopher,' directly into the Mosaic page layout. This allows organizations to capitalize on investments already made in gopher information systems. This also allows implementation of a Mosaic interface to occur faster, as many universities are simply adding a Mosaic front-end to their gopher systems while a complete Mosaic system is being developed.

These advantages and others have made Mosaic the interface of choice for many developers and publishers, including the Blacksburg Electronic Village prototype in Blacksburg, Virginia, Spry, Inc. and O'Reilly & Associates with their `Internet-in-a-Box' product and Global Network Navigator (GNN) service, and the pending Interchange Online Network service from the Ziff-Davis Publishing Company which is scheduled for release this fall.

But it is not only large-scale initiatives that promote the interface, as growing use among individuals and smaller organizations is being seen. Within our own University, there are already systems and development efforts delivering information in this form in the School of Medicine, Weatherhead School of Management, Library Collection Services Department, Information Network Services, and individual student dorm rooms. It is a growing practice on the Internet to attach to one's electronic mail signature file, the address of the sender's `personal home page' which includes a picture of themselves, their resumé and particularly interesting projects on which they are currently working. These pages have made it possible for me to recognize and be more familiar with Internet users from as far away as Australia than I am with users on my own campus. This project does not promote an effort to introduce Mosaic to Case Western Reserve University, as stopping its introduction would be the greater task. This proposal promotes to direct these growing efforts into an organized information resource and instructional tool.

Aurora Project Design Suppositions

The following are suppositions regarding the design of the Aurora interface. Any information item should adhere to the following guidelines before becoming publicly available on the system.

The system's target audience will have a variety of associations, both within and outside of the University. The interface will be a highly visible representation of CWRU. As such, it must promote the University, its mission, people, organizations, and programs in a positive and professional manner.
The system's target audience will have a variety of computer literacy skills. The interface will allow for ease-of-use for the beginner as well as shortcuts for the "power user". The interface will never be condescending in its responses.
The system's target audience will work with a variety of computer platforms and operating systems. The system will initially be designed around a standard 14" monitor, even though larger screens are becoming more prevalent. The automatic word wrapping capability of current World Wide Web clients will allow proper presentation on smaller and larger monitors. The system will initially be available for Macintosh, Windows and Unix clients.
The system's target audience will work from a variety of processor and network speeds. Whenever possible, processor power and bandwidth intensive objects such as graphics will be presented as "thumbnail" images, allowing the user to choose if they wish download a higher resolution depiction.
The system will display a "look and feel" design continuity across its spectrum. Local information sources will adhere to presentation standards deemed acceptable by the University. LIT will freely distribute Graphics Toolboxes to interested local developers.
The system will initially enhance and subsequently replace the reliance on Free-Net as a campus-wide information system, and as such, will have to deliver the rich variety of services currently available in addition to new ones.
The system and its maintainers will promote Computer and Network Ethics, as ascribed to by policies of the University and the National Science Foundation.

Implementation at CWRU

It should not be believed that I perceive the implementation of the Aurora interface at Case Western Reserve University to be an easy task. It will require many staff-hours of effort, as well as hardware purchases plus maintenance and support costs. But it will also capitalize on the efforts already underway, as well as allow the use of owned resources currently supporting information delivery mechanism for which the interface is the logical progression. For CWRU to stay at the forefront of electronic information delivery for educational use, as it has with CWRUnet, continued effort and advancement will always be necessary. Some of the major requirements are listed below.

Hardware: A full implementation of Mosaic at CWRU will require at least two midrange UNIX systems; one development system and one production WWW Server for campus delivery. I would highly recommend not combining the two on a single system for single point of failure, security and testing/service level issues, but this is a lower-cost option which could be employed during the initial phases of the project. Multiple developers may access the development system through the use of X-Windows clients which would allow work to progress from anywhere on CWRUnet. I would recommend a Sun SPARCstation 10 Model 51 for this platform. Since we expect use of the system to surpass that of Free-Net, growth paths must be planned from the outset. Sun's SPARC technology with the Solaris 2.x operating system allows processors to be added by purchasing `daughter cards' instead of entire new systems, thereby saving significant funding for future growth.
It has been proposed to use one of the existing Sun 690MP platforms for the delivery system. The 690MP offers the same I/O capability as the SPARCstation 10 platform since they share the same motherboard architecture, but has a slower processor and slower peripheral bus than the SPARCstation's FastSCSI interface. The 690MP costs 3 to 4 times what a SPARCstation 10 does and there would be costs, both financial and in service degradation, associated with moving services off the existing machine. It should be noted however, that while not offering quite the speed of a SPARCstation 10, the 690MP is still a quite capable machine. The added incentive that the 690MP is already owned by the University and would significantly reduce the start-up costs of the project would be fiscally irresponsible to ignore and should be weighed heavily in the final hardware design.

An additional advantage of the Aurora design as it pertains to future growth planning is the ability to perform load balancing without impact to the user. Load balancing with this system is accomplished by splitting the most requested pages among delivery systems. This is quite unlike the current Free-Net system where growth has necessitated users to attempt connections to freenet-a, freenet-b, etc., until an available machine is located. With Mosaic, the system appears to be one system, regardless of whether the information resides on 2, 10 or 1000 platforms. This in actuality, is the design of the entire World Wide Web.
Software: Mosaic clients and WWW httpd server software are public domain and as such, would incur no cost to the University. Macintosh support requires MacTCP 2.x. This software is needed for other Macintosh applications and negotiations for its procurement are already underway, regardless of whether Aurora is adopted or not. Similar to the Macintosh situation, Windows clients require WinSOCK compliance. Since CWRU supports a non-standard network software suite for DOS/Windows clients, WinSOCK compliance is not purchasable package. While WinSOCK is available free of charge in the Public Domain, it requires reconfiguration and rebooting to jump back and forth between Mosaic and other CWRUnet software. As with Macintosh clients however, WinSOCK is required by other software packages. INS developers have estimated that development of CWRU-PC/IP WinSOCK will require approximately 6-FTE months and is already underway. I recommend that this development continue post-haste, and LIT has already offered its assistance in the beta-testing process as we have in previous releases of CWRU-PC/IP.

Conclusion

The Aurora Project presents CWRU with a great opportunity to continue its leadership role in the electronic information delivery arena. But it is an opportunity whose window is measured, not in years or months, but weeks. Well-developed Mosaic systems already exist at the Honolulu Community College, Mississippi State University, Trinity College and the Georgia Institute of Technology. Norwegian efforts, where every university in the country is available from a single Mosaic home page, far surpass anything currently available in the United States. This would be an ideal model on which to base the university Mosaic efforts within the OhioLINK consortium, followed by expansion to all the universities within this country. As with the initial implementation of OhioLINK, CWRU is poised to be the leader, but only if the initiative is taken...