Turning the Vision Into Reality |-------------------------------------| | Paper presented at CAUSE92 | | December 1-4, 1992, Dallas, Texas | |-------------------------------------| TURNING THE VISION INTO REALITY Roy A. Gruver Director, Administrative Systems & Telecommunications Randall Wambold Manager, Systems & Programming, Administrative Systems Lehigh University Bethlehem, Pennsylvania ABSTRACT The VISION was the easy part...what was the role that information technology (IT) could play in supporting the University's mission? The vision is an exciting depiction which no one can argue with--and no one can implement. Plans, however, are implementable. How did we get from the Vision to the Plan? Our vision included eliminating 7 operating systems and our mainframes, networking the remainder of our campus, establishing a standards-based environment, running administrative applications on Unix¨ workstations, and reducing the percentage of institutional funds spent on IT. This document is not a tutorial on planning, but will generally describe the process and outcomes of translating a vision into a plan and into reality. It then focuses on specific impacts that implementing the plan has had on staff, budgets, productivity, management, and the computing and communications environment. Much of the information presented will relate to administrative computing and networking issues. THE VISION Have a vision as your starting point. A vision is continuously shaped by new needs and new technologies, by evolving standards and applications, and driven by the imagination and practical experience of the entire institution. Our vision in the mid-1980's was brought to life with the decision to expand our computing (in those days we just called it computing) infrastructure to include large numbers of university maintained microcomputers (over 600), a digital PBX, a campus-wide data network, and a campus-wide information system. The information system included electronic mail, bulletin boards, library catalog, file transfer, and site-licensed software products for word-processing, spreadsheet, graphics, and communications. We provided low-speed (9600 baud) access from every office, student residence, classroom and lab on campus. Many looked upon this as quite a dramatic change in environment, but we recognized that this was only the beginning. The change in our infrastructure in the mid-80s provided an environment which fostered positive changes in campus communication and access to technology. Literally, no part of campus life was left unaffected by the use of the campus network and information system and by the availability of microcomputers. We saw clearly how a general change in infrastructure could lead to a whole range of changes in the use of IT. We realized that providing the environment was sufficient to cause specific reactions. And, we might be able to control those reactions by carefully planning that infrastructure. So our new vision would carefully consider infrastructure. THE BEGINNING OF A PLAN For many years, Lehigh University departments have been producing annual progress and priority statements. Departmental plans are aggregated to produce a university-wide document. With the emphasis on annual planning, we did not create longer-term plans. While we had a tight grip on the vision, we produced no long-term comprehensive plan to support it. Several things occurred that changed the way we developed plans. First, as funding tightened, it became almost impossible to initiate and finance IT projects in a one-year time frame. Greater lead times and longer amortization schedules were required to fund new projects. Second, reallocation of existing budget dollars became the primary source of new funding. Each vice-president was responsible for funding most projects with monies already available in his or her respective stems. In this process, IT lost money to other priorities. Finally, our Computing and Communications Visiting Committee reported that a long- term comprehensive plan for Computing and Communication Services was essential to ensure effective management in the coming years. Thus, the first comprehensive plan for computing and communications was initiated. THE PLANNING PROCESS The vision as we perceived it was to use information technology to support the instructional process, research activities, and the business and administrative activities of the University. Translating this vision into the beginnings of a plan, we set three broad goals: (1) Provide desktop, public-site, and classroom access to all computing and information resources on campus and to similar external resources, especially those available on the national Internet. (2) Provide necessary computing and information resources to enhance personal productivity, and to improve education, research and the administration at the University to ensure that Lehigh remains a competitive educational and research institution. (3) Provide sufficient support to enable the user to simply, effectively, and efficiently utilize available computing and information resources. The planning team consisted of the Assistant VP for Computing and Communications Services, the Director of the Computing Center, and the Director of Administrative Systems and Telecommunications. Overcoming the inertia of inactivity and indecision was difficult. We had never done this, so starting was no simple matter. The CAUSE Library had several plans available which served as our guide to the format and language of our plan. We used these plans as a model. The next step was easy. We sat down and made it up. We blended our experience and the experience of others with a reasonably good understanding of the current and emerging technology. The purpose of our plan became, then, to develop the infrastructure needed to support the probable future computing environment. There was little input from outside the organization, and not much more from within. The entire process took less than one month. Although, it was a very busy month. The final task was to plan for evaluation and acceptance. We already had top management's support to develop the plan. Since the planning process included very little outside influence, we now wanted to develop consensus. The plan was shared directly with the president's council and the provost's council. These groups include all of the top management of the University. Further, each faculty member and each administrative department head was encouraged to comment. Finally, each of the Computing and Communication Services advisory committees was asked to review the plan and make comment. We braced ourselves for a protracted period of plan review, critique, and rewrite. It never happened. We received minimal critical feedback. Most everyone understood the plan and liked it. We immediately started to focus our actions through the strategies, tactics, and programs of our plan. CONTENTS OF THE PLAN The plan is mainly about creating an appropriate computing and communications infrastructure. In addition, the plan describes the services and support available, and the process for administering and financing the plan. Most of the plan deals with a macro view of information technology, however, it does consider some specific application needs. Target Infrastructure The target infrastructure described in the plan is comprised of four major components: (1) a campus-wide high-speed network, (2) a distributed computing and information environment, (3) computing and information standards, and (4) campus relational database technology. The plan call for the extension of the campus-wide high-speed network to all academic buildings and laboratories, administrative offices and campus residences. This network is a fiber optic backbone which is intended to bring high-speed connectivity to the desktop. Network access is described in the plan as an information utility, much as we describe telephone, electricity and water systems. The plan sets forth a distributed computing and information environment consisting of resources ranging from high-end compute and file servers to desktop computers, all readily accessible through the network. In the past, most computing requirements, ranging from the most demanding to the most elementary, were provided by individual mainframe computers. The focus now would be on powerful workstations and the eventual elimination of mainframes. It should be noted that the plan includes all computing activities, including administrative and business support, in the move to distributed resources. The University has made a major commitment to a wide variety of computing and information standards with this plan. The intent of this commitment is to mainstream our activities and create much greater vendor independence by avoiding proprietary solutions and adopting open standards wherever possible. The plan calls for the establishment of standards for operating systems, data management, software portability, graphical user interfaces, hardware, interfaces, and network protocols. Compatibility and supportability are critical characteristics of the new standards-based environment. When the plan was released we supported 9 operating systems on campus. How much simpler the support task would be if we only had two or three. The plan calls for the adoption of a single relational database product for use across campus. The use of large-scale relational database technology was limited on campus. We felt that relational technology could yield certain benefits for us. However, we wanted to restrict the use of relational technology to one product set. This was important for several reasons: we wanted to ensure that University data were accessible in a common way; we wanted to limit our support and training efforts to one product; we wanted to develop portable applications and databases; and we thought we could negotiate a more cost-effective deal with one vendor than we could with many vendors. Services and Support To ensure we were able to maximize the utility of the new environment, the plan called for the development of appropriate training and support structures. We would develop new classroom training for students, faculty and staff, on all campus standards, and expected hardware platforms. We would reassign staff to provide broader and deeper support for local and campus-wide networking. We would extend consulting support to the local computing environment. The plan also introduced the idea to create graduate assistantships specifically to provide departmental computing support and to provide research activities with extended professional services on a cost-recovery basis. If we were successful in focusing on a set of campus standards, we would offer to manage IT resources that met campus standards. By limiting what we would support, we expected to be able to expand the population served and the level of support provided, even to taking over the management tasks for other departments, if appropriate. While we had already implemented a fairly broad range of information services on campus, the plan wanted to satisfy the increasing demand for information services of different kinds. These included voice mail and fax services, as well as moving much of our existing information services--electronic mail, bulletin boards, conferencing, online library catalog, current contents databases--into the distributed environment. Applications We will continue to negotiate site-license agreements for widely used software products. Most of the software in this category has been for microcomputers, but we are increasingly adding workstation software. The wide range of functions includes wordprocessing, spreadsheet, symbolic math, statistics, and authoring languages. The plan itself does not specify software products by name, but the selections are clearly supported by the variety of training programs available. The need for several new administrative applications is noted. The plan includes the general notion that applications should be consistent with the provisions of the plan, and the specific idea that administrative applications were not an exception. The plan supports a common end-user query and reporting facility. Our customers had a great need to easily construct requests and retrieve information, and they needed much more control over local printing. Plan Administration and Finance Programs within the plan specifically describe the importance for management to commit to the plan. These programs also describe how the plan is to be reviewed and updated on an ongoing basis. Each year we produce a progress report and recommend revisions to be made to the plan. The Computing and Communications Services advisory groups are requested to review and evaluate the plan annually, but can make recommendations at any time. The plan needs to be integrated with other plans. To achieve maximum effectiveness the plan cannot be considered to apply only to Computing and Communication Services--it must be accepted as the foundation plan upon which others would build. Likewise, this plan cannot be considered to be the only plan needed for computing and communications. The colleges and research centers should create their own plans describing how they will use the infrastructure utility created to support their activities. There are a number of financial implications discussed in the plan. For the most part, the number of staff will be static. The plan noted in which areas additional staff support was likely to be needed. The plan also noted that most of those staff positions would be reallocated from other areas within Computing and Communications Services. Reallocation is a key word in this section of the plan. While the plan called for new funding to support its aggressive programs, it provides over half of those funds through the reallocation of existing Computing and Communication Services monies. IMPLEMENTING THE PLAN Our campus-wide high-speed network has been expanded considerably over the last two years. We have completed the connection of all academic buildings to our Pronet 10Mb/sec TCP/IP network. We have started to replace segments of the network with FDDI segments. Our LAN environment is, for the most part, twisted-pair ethernet to the desktop with Novell's Netware¨ connecting workstations and PCs. The end-user "workstation" in academic areas is generally a PC with Microsoft's Windows¨ or DOS¨, or a RISC workstation with Unix and X- Windows. In most administrative areas, it is a PC with Windows or DOS. We have made good progress in moving away from mainframes to a distributed computing environment. In mid-1990 we had 4 mainframes with 5 proprietary operating systems. By January 1993, we will have one mainframe using 2 proprietary operating systems. That remaining mainframe supports a number of administrative applications, but the administrative computing environment is now a mixed environment--one mainframe and multiple RISC workstations. We have installed almost 200 workstations since the plan was accepted. We replaced several PC public sites with RISC workstation sites, installed several new departmental workstation labs, and placed a number of additional workstations on faculty desks. We now sell and maintain IBM RISC workstations through our Microcomputer Store. By focusing on IBM RISC workstations, we were able to negotiate special pricing with IBM, yielding even greater cost-effectiveness. Lehigh selected the relational products of the Oracle Corporation as our campus standard. Since we had a large variety of computing platforms when this all began, Oracle's wide portability was expected to be important. We are now focusing on the Unix environment. Oracle's RDBMS is available on a wide range of Unix-based platforms, with good connectivity, and has a large toolset for both application development and end-user data access. By focusing on one vendor's products, again, we were able to negotiate very favorable pricing for the products. In support of administrative systems users, a "standard" application and services implementation model has emerged. The major components of the model include: (1) an application server accessed through the high-speed network; and (2) a LAN in the user office, attached to the network and consisting of a RISC station as the LAN server with a mix of simple DOS and Windows nodes as clients. Under this model, the actual databases and RDBMS engine reside on the application servers along with the major application processes and services that need to be available to a broad set of end-users. Each application server is administered centrally, including such tasks as operating system maintenance, application maintenance, database and tools administration, and backup/recovery processes. The LAN server provides typical work group benefits such as disk storage, file sharing, printer services and access to common DOS and Windows applications. Further, the LAN server can provide specialized application components or services that require the processing and graphical capabilities of larger PCs working in conjunction with the application server, in a true client- server relationship. By locating these special application components on the LAN, we are able to maintain a standard central copy of the components. The administration of the LAN server is a joint effort between the end-users and the computing staff. We have engaged several pilot projects to test our ability to provide remote management of customer LANs with existing staff. We expect to be able to manage a few sites at most. Several departmental sites have been turned over to computing and communications services to manage. These sites have been ones which match the target environment in the plan. There has been limited advisory committee review and limited integration with other plans. We have been able to fund the projects so far, but our budgets are expected to decline (in actual dollars) over the next several years. While we expected to have to finance much of the plan through reallocation, we did not expect that those budgets would be reduced. We have been able to develop some new funding from operator services (zero-plus) commissions and voice mail revenues. CRITIQUE OF THE PLAN The plan has provided an extremely important management tool for Computing and Communication Services. Not only does the plan define a basic direction and design for computing, but it continually aids the decision-making process by providing us with criteria to measure each proposal or possible action against. Appropriate actions appear clear when the primary decision criteria is plan-compliance, and the inappropriateness of other actions are quickly revealed in their non- compliance. Additionally, it is important to note that others on campus more readily accept our decisions because they conform to the plan. Senior management has not questioned major initiatives or policies when it has been demonstrated that they are in support and compliance with the plan. Our vendors like the fact that we have a plan. They can focus their energies and can develop longer-term strategies for meeting our needs. We have been able to develop a number of partnerships which have been instrumental in our continued implementation of the plan. We have worked with IBM on examining the workstation environment for large-scale University application systems. We have worked with Systems and Computer Technology, Inc. (SCT) to develop improved applications for the distributed market. Finally, we are currently working with a partner to develop a suite of telemanagement applications for this new environment. We have been able to take advantage of special pricing because we were able to focus on workstation and relational technology. With such a clear understanding about where we are going, we can make commitments and decisions with vendors that support those directions. A clearly articulated plan has produced two somewhat contradictory situations. First, the plan has created substantial expectations among our customers. The campus is eager for enhanced computer and information resources, wider and easier access, as well as new tools to increase productivity. Within administrative offices, most are willing and even anxious, to move into the new environment--wanting a combination of their current application system's functionality and the benefits of the new environment: SQL-based RDBMS, and improved end-user query and reporting tools. But, while the end-users generally view the plan as attractive and worthy of pursuit by the University, it has caused considerable confusion and concern when they seek to apply its direction to their specific offices. There is a belief that the plan will force them to move to new applications before the available products are mature. While we have reassured them that we will not force their movement to replacement applications that do not meet their needs, we have confirmed our intent to follow the plan. When an administrative application requires substantial improvement, we will seek and propose products that conform to the plan. The plan narrows a vast number of computing and information models to a smaller set. But this has not limited the real options available to us. We have found a strong set of satisfactory computing resources, hardware, software development tools, and applications, that comply with the plan. And, in the cases where the applications have not been fully satisfactory, we have found vendors willing, through partnership, to create highly functional applications for the new environment. So the plan, by making Lehigh an attractive partner, may have actually increased our computing options rather than having limited them. The plan does not deal fully with the difficulties of transforming our computing environment from the old to the new. This is especially true for administrative computing. The plan does not address the issue of how the migration of the administrative application stock can be accomplished. The plan simply notes that replacement applications that conform to the plan are currently available, or will be within the next several years. There is no plan, or funding, to acquire and implement these systems. During the interim conversion period, several transitional problems have emerged. First, as we move to our new standards, the required support effort has actually grown. We still have multiple operating systems, file systems and development environments. For the Administrative Systems group, the basic toolset has doubled. Unfortunately, if an application fails, there are now fewer people who are familiar with the required tool, and even fewer who have a substantial level of experience or proficiency with it. Second, the level of application systems integration has actually declined. This is an unpleasant consequence of having applications reside in two very different environments. Formerly integrated systems have been separated in order to facilitate our movement to the new environment, causing us to create new interfaces. We expect that as the new environment becomes our standard, these problems will ease. Although the plan did not deal fully with the difficulties of the transition, the new computing model has proven to be effective and flexible. Beyond our primary administrative implementations of an Admission system and an Alumni/Development system, we have used the new application development environment to quickly construct several reporting and analysis databases. These databases have included medical claims information and summarized financial data. We expect that additional data, currently maintained in the old environment, will be transferred to the new environment to permit the end users to access it with the new tools. We hope that this step will facilitate the full conversion of the applications in the future. One of the objectives of the plan was to create a more standard computing environment, which is easier to support. Prior to the plan, we had a few large computing resources with different operating systems and data structures. Now we have fewer operating systems and data management structures, but the productivity gains here have been more than offset by the added complexity of the new environment. This added complexity has at its root the sheer number of computing resources that must be supported. Each new compute, file, or applications server adds, proportionally, to the support work load. Each requires hardware, operating system, RDBMS, and tools installation and maintenance support. To control this support effort, we are seeking to regularize our support processes and to partition the services and tools on each resource. With the significant increase in computing resource complexity, there has been a parallel occurrence in the software development area. We noted earlier that there is an abundance of tools in the new environment, which adds to its general complexity. A point of concern with our plan is that on the issue of identifying the "right" long-term toolset, the plan remains at a very high level, offering little help. There is no mention of an appropriate application development methodology, no mention of CASE or Object Oriented development tools. The silence of the plan in this area adds to our implementation complexity, but it also adds to the flexibility of the plan by allowing those who are closest to the implementation to make the final choice of tools. Although the plan does not specify any changes within the Computing and Communication Services organization necessary to accomplish the plan objectives, it has been a consequence of our implementation. Each department has accepted new responsibilities as a necessary step toward our accomplishment of the plan. As an example, the Administrative Systems group has dramatically broadened its duties. Prior to the acceptance of the plan, this group's primary responsibility was to create and maintain application software on a mainframe, LANs and PCs. Now the group's primary responsibilities have expanded to include workstation and LAN operating system installation and maintenance, RDBMS and related products installation / maintenance / administration, and database administration activities for those applications that are currently functioning within the new environment. Twenty percent of the Administrative Systems group's effort is now spent supporting the new environment's infrastructure. The reallocation of effort to the infrastructure support function has lessened the overall ability of the office to create, deliver and maintain application systems and services, although the quality of the delivered systems and services is higher. Finally, the plan acknowledged that Lehigh's funding for Computing and Communication Services, as a percentage of the general budget, had been declining for several years. The plan proposed a funding scheme which depended on significant budget reallocation within the Computing and Communication Services organization plus an injection of new funds, both one-time and continuing. To date, internal reallocations, via the elimination of several mainframes, have provided the bulk of our implementation funding. In addition, there has been some new funding, predominantly one-time monies, available for classroom and lab upgrades to workstations and administrative system upgrades. Overall, the total budget of the Computing and Communications Services organization has declined, in real dollars as well as a percentage of the University's budget, during our implementation period.