Migrating from Host-based Computing to Client/Server Copyright CAUSE 1994. This paper was presented at the 1994 CAUSE Annual Conference held in Orlando, FL, November 29- December 2, and is part of the conference proceedings published by CAUSE. Permission to copy or disseminate all or part of this material is granted provided that the copies are not made or distributed for commercial advantage, that the CAUSE copyright notice and the title and authors of the publication and its date appear, and that notice is given that copying is by permission of CAUSE, the association for managing and using information resources in higher education. To copy or disseminate otherwise, or to republish in any form, requires written permission from CAUSE. For further information: CAUSE, 4840 Pearl East Circle, Suite 302E, Boulder, CO 80301; 303449-4430; e-mail info@cause.colorado.edu MIGRATING FROM HOST-BASED COMPUTING TO CLIENT/SERVER Jay Blum Computer Center Director Thomas More College Crestview Hills, Kentucky and Richard Reno CARS Information Systems, Corp. Cincinnati, Ohio Abstract Thomas More College is working to successfully evolve the CARS System from a host-based computing system to client/server architecture. The CARS System provides the institution with the ability to migrate individual offices or office users at a time best suited for them, independently of the migration plan for other offices or users of the administrative system. Technologically, the key benefit is the PC utilized as a workstation on the administrative system versus as a character-based terminal through a terminal emulation package. Thus, the workstation provides for the following: * Initial balancing of the processing load * An easy-to-use interface to the host-based applications * Efficient and effective PC tools that are familiar to the user This case study of Thomas More College involves the prototyping of the practical path from host-based computing to the client/server architecture of the CARS System. The thinking and the planning for the right architecture at the right time are critical. The implementation and support of this architecture must be feasible for both the institution and its personnel. The appropriate architecture allows the Computer Center to realize the benefits of using the client workstations with the server/host. The CARS System, running on a host computer, acts as the server for the workstations, which are acting as clients. This enhances the users computing resources and CARS as the vendor server enhances the computing of the institution as a client. INTRODUCTION: PARTNERING THE CLIENTS AND THE SERVERS WITHIN THE CAMPUS INFORMATION SYSTEM One can hardly pick up a computer magazine today without seeing the term "client/server," and many institutions are evaluating the possibility of moving toward that type of system. We hope that our experience at Thomas More College (TMC) will be informative for other institutions considering moving toward client/server architecture. For our continued discussion, we define "client/server" as follows: The "server" is the computer where the data and maybe the executable program resides. The "client" is the PC or workstation where the executable program is loaded at the time of execution. This client only accesses the server when the program loads or when there is a need to access data. For Thomas More College, this process began in February of 1991 when the President commissioned the VP of Finance and Administration, to form a task force to develop a long range plan for computing on campus. The group was made up of computer center personnel, faculty, staff, students, administrators, and representatives from IBM. At the time, the TMC administration was running CARS software on an HP 9000/832 connected via serial connections. Academic Computing was using a Micro VAX with twenty-one networked PCs. As members of the task force, we first assessed the current situation at TMC. We distributed a written needs assessment survey to faculty, staff and students. Based on the responses from that survey, we developed a list of campus needs. From this needs list, we determined the consequences for the campus if these needs were not addressed. We then met with users in small groups to get feedback and begin finding solutions. Armed with this user information, the group then developed a solutions list. With this list, we proceeded to list the benefits. From this work, TMC developed a three-year plan for the installation of a campus- wide network. Total cost of the project was estimated at $1.5 million. In 1991, TMC was awarded a $750,000 grant to begin the installation of a campus-wide network. Installation began in May of 1992. TMC installed 1400 individual connections using Level 5 UTP, with at least two connections in each room, office or classroom. Some offices and classrooms had as many as 25 connections. Installation included the Computer Center's central concentrator and five (5) other concentrators on the campus connected via a fiber optic backbone. The grant also allowed us to upgrade the PCs on campus. When examining the direction of computing, we found we needed powerful workstations. Our original plan called for the purchase of forty (40) 386 PCs for users. As an example of how fast technology and pricing can change, in only eight months from the plan's acceptance to completion, we ended up purchasing sixty-seven (67) 486 and thirty (30) 386 PCs for the campus. Forty-six (46) of the 486 PCs went to student classrooms and labs, and the rest to staff and faculty offices. Because of this installation, we were immediately able to consider moving our administrative computer system to a client/server system. Because of our network, CARS Information System Corporation asked us to be their Prototype Site for the Graphical User Interface (GUI) for the CARS System. CARS has developed a two-step approach to moving to client/server architecture. The first step is to convert the interface for the user from the current character-based interface to the Graphical User Interface (GUI). During this phase, the programs still reside and load on the central computer, whether mini or mainframe. The user interface changes to GUI, allowing the users to learn to use the Windows interface without a major disruption of their work. Step two would be the expanded migration to the client/server architecture. I. THINKING: THE CURRENT ISSUES OF EVOLVING COMPUTING FOR A SMALL INSTITUTION When thinking about the direction in which campus computing should evolve, the following important issues must be considered: * The available technologies, both in the present and in the immediate future * The limited resources that the institution can allocate to the project * Any proposed change should be developed within the framework of a strategic plan for information technology utilization * The current use of PCs by the staff, since these may be considerable A. Consideration of the New Technology within Computing When considering an upgrade for your institutional computing environment, consider the rapid rate of development in technology today. Administrators have a hard time understanding that something that was top of the line only two or three years ago is now literally unusable to carry out current tasks. To ensure that fundamental needs are discovered before implementation, the institution must consider the rapid rate of technological changes in its plans. Because hardware prices are plummeting so fast and the technology is developing so rapidly, an institution benefits from lower prices overall or greater computing power through the implementation cycle. The most important aspect of considering new technology is to try to foresee the future as much as possible. Institutions must also research new emerging technologies to decide whether they will be important to the institution within the lifetime of the current plans. Good technological examples today are real-time audio and video transmission across the campus networks and presentation of the transmissions to the desktop. This technology promises to be very important to academic institutions. Other technologies of interest include document storage and virtual reality systems. We believe that the single most important component for any computing model today is the campus backbone network. An institution could get very detailed in terms of anticipated network usage to decide which cable to use. However, if such consideration is based upon current usage plus reasonable growth estimates, and if the network is expected to last ten years, then any such detailed consideration will surely be wrong. These considerations do not consider the future technologies involving the distribution of images, audio, and video across the network. The network usage demanded by these technologies will dwarf the use of the network for "standard" computing operations. B. Realization of the Limited Resources Allocated within the Institutions Budget Most academic institutions today are facing shrinking budgets, while simultaneously receiving increasing requests for computing resources from various constituencies. The requests become even greater after a network is set up, and users begin to see what others can do. When the user base encompasses most of the campus, computer center staff can more easily justify purchasing the necessary computing resources. The institution must consider realistic funding estimates when planning. Because Thomas More College experiences the same budgetary constraints as many other colleges, we decided we needed to incorporate our existing resources when moving forward. C. Recognizing the Need for a Strategic Plan for Information Technology To change an institution's computing environment, a large segment of the campus community must support it. One way to accomplish this goal is to provide a strategic plan for information technology. In preparing a strategic plan, the computer center must solicit opinions from as many constituents as possible. When users participate in the process, they are more likely to support the plan. If most of the campus supports the IS strategic plan, administrators are more easily able to carry out components of the plan. D. Using PCs within Administrative Processing As previously mentioned, TMC started by evaluating the current campus situation and developed a three-year plan. Because of the participation of the whole campus, we had a great deal of support for carrying out our plan. In the plan, we recognized that utilizing PC workstations would be an important component. For budgetary reasons we decided initially that only users using wordprocessing or spreadsheets on a daily or weekly bases would require a PC. The grant allowed us to expand usage with our original purchase, but we still need about seventy (70) PCs to meet the demand. II. PLANNING: THE RIGHT ARCHITECTURE FOR ADMINISTRATIVE COMPUTING ON CAMPUS In planning for an evolutionary change in the computing environment, the architecture of the new system must be considered to find the best fit with the current resources and expertise. Capitalizing on users particular computer expertise and using tools with which they are familiar, the system must effectively tie together the users (clients) with the information sources (servers) for a system to be successful. A. A Plan for Integrating the Clients and the Servers into Administrative Processing At TMC, the primary server within the administrative system is the administrative minicomputer. TMC currently has an HP 9000/G30, with HP/UX operating system, running the CARS administrative software package. Before the installation of the network, there were about four users, which was 20% of all users, using a PC workstation with a terminal emulation package, while the rest used character-based terminals. After the installation of the network and the PCs, the percentage of users increased to more than 50%. This increase PC use allowed us to begin looking at client/server. Because the CARS System features an open system design, TMC could become the prototype site for the CARS System GUI software. This allowed TMC to retain the CARS software with which it was familiar and still move forward as fast as practical with the evolution to a client/server system. B. Migrating Administrative Systems that are Fully Operational Systems Today TMC wanted to migrate from the old system to this new technology without disrupting the entire campus. Besides requiring a substantial financial expenditure, such a migration would also entail a major training effort. TMC wanted the migration to take place over an extended period. To meet these concerns, CARS proposed the following two-step approach: (1) Allow users to experience some benefits of a GUI C/S system without requiring a complete changeover. With the graphical user interface, the campus can simultaneously carry out both traditional host-based computing and C/S computing. (2) Expanded migration to the client/server architecture. C. Allowing the Institution to Evolve Over a Time Frame According to the Needs of the Institution The primary new requirement for the CARS Graphical User interface is a PC operating as a workstation with a network connection. Because TMC had recently installed the fiber optic backbone on campus, the computing staff simply converted PCs to C/S ready workstations and hooked them to the network. This process can now continue as funds are made available until all CARS users are using C/S style workstations. Users have the flexibility to switch back and forth between terminal mode and GUI according to their individual preference. III. IMPLEMENTING: THE PRACTICAL PATH TO CLIENT/SERVER COMPUTING USING CARS GRAPHICAL USER INTERFACE Presently, TMC is well along the way to implementing CARS Graphical User Interface across the campus. Those people who are presently using this system, and who had previous PC background, are for the most part happy with the operation of the system. The MIS staff is in the early stages of learning to cope with the problems that will arise in a expanded C/S system. In particular, the staff is learning how to deal with the problems of more users accessing the network, and the increased load on the backbone. A. A Graphical User Interface for the CARS System The CARS Graphical User Interface is a CARS application specific tool that resembles an X-windows server. The GUI program server (currently only for Microsoft Windows) acts as a display manager for the software running on the central host. It uses a direct network connection to the host for communications. The CARS GUI allows the user to have multiple sessions active in different windows, move those windows around on the screen, and use a mouse for some selection operations. The most commonly stated benefit of using the GUI is that most users like the uniform operation and the reduced time hitting a return while doing data entry. "They love the point and click." B. Integrating Familiar User Tools Allowing the user to interface directly to the system using familiar standard PC tools is one strong argument for moving to the client/server model for the administrative system. The CARS Graphical User Interface allows TMC personnel to get some experience with this now and to prepare for a client/server architecture solution in the future. With the CARS GUI, the user can copy and paste into other GUI products. In addition, the CARS GUI is expected to support file transfer between the host and the user's PC. Users will be able to easily enter scanned images into the database (documents, signatures, photos etc.). Similarly, users will be able to retrieve these stored images and paste them into other applications. C. Introduce New Software over an Evolutionary Timeframe TMC hopes to eventually install the link software that will extend the database access to the PCs. One future example might include Q+E software that will link any PC software that allows OLE connectivity to have access to the database. In particular, TMC expects to use spreadsheets and third party report writers. We hope these report writers will relieve some pressure on computer center staff to create special reports. Frequently, the spreadsheet will be the natural interface to the database for the user: for example, applications running projections and statistics. Using a spreadsheet, users who are currently running reports on the system and uploading or hand-entering the data into a spreadsheet, will be able to add the queries in the spreadsheet to get data from the database directly. IV. SUPPORTING: THE FEASIBLE APPROACH IN COMPUTING ARCHITECTURE FOR THE CAMPUS PERSONNEL As an institution evolves its computing environment, the computer staff endures great stress to support this change. When institutions try to cut corners in staff training during the change, the success of the change is at risk. It is surprising how many institutions of higher learning do not believe in education for their employees. In the move to the client/server paradigm many, if not most, of the staff may not yet know how to use a GUI interface. Training must be a high priority. A. Implementing at the Right Time for Each Individual User At TMC the conversion is now well underway. One distinction of the CARS System Graphical User Interface is the fact that the computer center staff only has to maintain one screen definition file for both the underlying host computer and the CARS Graphical User Interface. The user is still running the same program on the host that they used to run when using a character-based terminal. The application program senses that the PC is running the CARS Graphical User Interface at startup and modifies its behavior correspondingly. Because the programs adapt to the situation on the workstation, TMC staff can on an 'as desired basis' move users to CARS Graphical User Interface. For example, an individual office may be moved, an individual may be moved or an application may be moved. If for some reason users want to suspend using the Graphical User Interface they can turn off the GUI server program and rerun the program; the program will appear in character-based mode. B. Updating the Administrative Software with Mandatory Changes from the Vendor Recognizing that institutions are constantly changing, whether to meet regulatory changes or to adapt to institutional policies, CARS realized the Graphical User Interface would need the same flexibility as the rest of the CARS System. When enhancements arrive from CARS, the computer center staff need only install one version. Because of the way that the Graphical User Interface operates, the same screen information a character-based terminal utilizes is also used by the GUI server program. Therefore updates must be made only once. V. EVALUATING: THE BENEFITS OF THE PRACTICAL PATH IN ENHANCING COMPUTING FOR THE CAMPUS A. How Did the Process Go? In evaluating this process at Thomas More, we found that the process worked very well. Realistically assessing where we were and where we wanted to go, TMC could map out a plan that incorporated the following: * Available technology * Our available budgetary resources * Our current investments in technology * Our strategic plan for information technology utilization Because the MIS plan incorporated the needs of the entire campus, the campus as a whole had a commitment to its success. We found users taking ownership of the process, because they were actively involved. We believe this better equips our campus in preparing for a client/server architecture administrative system. B. What Did We Learn Along the Way? A major lesson we learned was that any strategic plan must be flexible, to meet the constantly changing needs of the institution and also the emerging technologies. For TMC, the network provided an excellent example of this needed flexibility. Because TMC installed a fiber optic network, the institution was immediately able to begin testing CARS Graphical User Interface. Through this testing, TMC was also able to see how use of the Graphical User Interface increased network load. Because the CARS Graphical User Interface allows individual users to migrate on their own time schedule, TMC could carry out an evolutionary migration of the entire campus without disrupting its day-to-day operations. C. What is the User's Perspective? Our first user for testing the GUI was a data entry person. This person did not have any PC Workstation experience. After a week of Windows training, the user preferred using the GUI for entering information into the system. In the midst of this process, when the user switched offices, the first question to the Computer Center was, "Will I have the same GUI capabilities in my new office?" From Alumni and Development, another user used the CARS System testing the Graphical User Interface. This user prefers using the GUI, because she finds it easier to use. The data entry users like pointing and clicking on fields versus hitting a return during data entry. "They love the point and click." D. What is the Computer Center Perspective? Because both the GUI and the character-based CARS System are maintained by a single screen definition file, the GUI does not require any additional maintenance of the Computer Center staff. When TMC incorporates a new enhancement from CARS, these enhancements will be used in both character mode and GUI mode. Concerning training, we also found that users who were familiar with PC applications, were easily able to switch to using the GUI. For those users for whom the GUI was their first experience in PC capabilities, once they learned how to use it, they were able to easily learn other PC applications. While positioning us to migrate to a full client/server architecture in the future, TMC has found CARS Graphical User Interface to be a cost-effective solution that incorporates current resources while allowing users to enjoy the benefits of a PC workstation.