CAUSE the association for managing and using information resources in higher education The Crisis in Information Technology Support: Has Our Current Model Reached Its Limit? by Polley A. McClure, John W. Smith, and Toby D. Sitko CAUSE Professional Paper Series, #16 CAUSE Professional Paper Series, #16 The Crisis in Information Technology Support: Has Our Current Model Reached Its Limit? the association for managing and using information resources in higher education 4840 Pearl East Circle, Suite 302E Boulder, Colorado 80301 Phone: 303-449-4430 Fax: 303-440-0461 Publication of this paper was sponsored by Adobe Systems Incorporated. This print paper was prepared using Adobe PageMaker and Illustrator software. An HTML version of the paper created with Adobe’s PageMill software and a PDF version of the paper created with Adobe Acrobat will be available by the spring of 1997 on the CAUSE World Wide Web server at http://www.cause.org/pub/profess.html. Published by Copyright © 1997 by CAUSE. All rights reserved. A complimentary copy of this paper has been distributed to all CAUSE member representatives. Additional copies are avail- able to anyone on a member campus at $16 per copy. The paper is available to non-members at $32 per copy. Inquire about quantity discounts to CAUSE Professional Paper Series, #16 The Crisis in Information Technology Support: Has Our Current Model Reached Its Limit? by Polley A. McClure, John W. Smith, and Toby D. Sitko Contents Introduction 1 The Crisis Triad and How We Got Here 2 Elements of an Ideal Technology Support Environment 7 We Can Get There From Here 17 Seeing Our Future in an Evolutionary Context 19 Corporate Sponsor Profile 22 CAUSE appreciates the generous support of ADOBE SYSTEMS INCORPORATED who funded publication and distribution of this paper (see pages 22-23) About the Authors Polley A. McClure ( THE CRISIS IN INFORMATION TECHNOLOGY SUPPORT 1 Introduction rom the day the first computer was installed capacity to supply. Yohe sounded the alarm, in an institution of higher learning, identifying many factors contributing to the crisis Ftechnology managers have dealt with and suggesting some useful immediate steps to increasing demand for services. Continuous lessen their impact.1 growth of 20 or 30 percent per year in various We think the current crisis is not just a result of measures, from CPU cycles to help desk calls, were increasing numbers. Fundamental changes taking common. The campus periodically entered a crisis place in higher education make old models of state when specific capabilities became saturated. information technology support inappropriate and The typical response to the crisis was to present insufficient. In this paper we describe the current the obvious inadequacies to the institution and problems and suggest some ways in which they are ask for additional resources. As soon as enough different from those we faced in the past. We capacity to last for a few years was added, the cycle propose some elements of a new support model began again. and suggest how we might evolve to it from our Today we perceive a sense of crisis in current environment. Finally, we present a brief information technology support, but it has a evolutionary history of information technology different twist. Our conversations with colleagues support in higher education to put the current describe the problems as more expansive, and the situation in historical context and summarize old solutions are not working very well. We appropriate directions for the future. cannot fix the insufficiencies by buying the next generation computer or by hiring five more people. In virtually every service we provide, 1 Michael Yohe, “Information Technology Support Ser- demand seems to be growing far beyond our vices: Crisis or Opportunity?” CAUSE/EFFECT, Fall 1996, 6-13. 2 CAUSE PROFESSIONAL PAPER SERIES, #16 The Crisis Triad and How We Got Here or years we have struggled with the need for respond to these exploding needs. Many of these increased quantity and quality of services problems are common at other campuses, and they Fand support. What tips the scales now is the are beyond the ability of a single institution to degree to which our underlying assumptions are resolve. inadequate. Three primary issues characterize the Non-linear, exponential growth is not a new current crisis: phenomenon in information technology. The basic ◆ Demands on central information technology elements driving demand for services, however, (IT) organizations are overwhelming provide insight into why today’s demand growth ◆ Support quality is deteriorating curve is so precipitous. ◆ Central IT organizations are becoming the scapegoat More customers need more services To create a solution, we must understand the A decade ago, fewer than 20 percent of our mechanisms behind each of these issues. faculty, staff, and students were active consumers of technology services and support. Today, almost all of them are, at least to some degree. Ten years Central organizations are ago, a handful of the campus population was overwhelmed by demand interested in dial-in access. Today, a typical Faculty and administrators on our campuses student package includes accounts for e-mail, dial- increasingly perceive information technology to in, the World Wide Web, and networked file and be critical to their work, and they want central print services. The educational potential of the technology organizations to promptly meet their Web alone has unleashed a firestorm of support changing expectations. At the University of demands, not to mention escalating printing costs Virginia, for example, demand for dial-in lines has in public labs. increased by more than 100 percent each of the last two years, as have calls to the help desk, Per capita demand for services has increased network traffic, and requests for UNIX accounts Users used to work for months to generate a few for Web pages. Administrators are demanding tens of thousands of bytes of information. When electronic forms and information warehouses; we they moved that information across the network, a need to replace our legacy systems with modern few kilobits per second of bandwidth delivered client/server integrated applications. Faculty need adequate service. Today, a user with a scanner can support in introducing technology into their generate hundreds of megabytes per hour. When a classes; they want properly designed and equipped thousand people attempt to view those electronic classroom facilities in which to work. Record images via the Web, even 100-Mbps networks are numbers of students bring computers to campus, stressed. In forty hours of instruction per semester, and they want Ethernet cards installed and we can light the spark that will make a student configured as soon as they arrive. When Internet want to use information technology resources and performance decreased last spring, research faculty services forty hours a week to write papers, run lab demanded improvements in this fundamental simulations, and interact with instructors. resource for their work. Budgets have not increased significantly for the past five years at the New users are mainstream University, nor has the number of people to The factor that has most dramatically escalated THE CRISIS IN INFORMATION TECHNOLOGY SUPPORT 3 the demand for support is the new breed of user. develop problems when interconnected. The Two decades ago, our users were a hardy group, number of potential problems increases knowledgeable about and seriously interested in multiplicatively as more and more heterogeneous computing. They were tolerant of system components are added to the mix. Simply moving idiosyncrasies and failures. Adequate support a linked file into a different subdirectory can break meant posting signs in the computer center with applications campuswide. Common user examples of the control cards users needed in applications, such as library bibliography access, order to run different kinds of jobs. We wrote pass through technologies managed and controlled documentation and they had the motivation and by different administrative units. Thus we must expertise to decipher it. Recent users of deal not only with technical interactivity, but also information technology are often not particularly cultural and administrative diversity. interested in the technology itself, and they are willing to spend only minimal time and effort to Funding models are inappropriate learn to use it. How we support the last 20 percent Many of our institutions are still operating of the population that we are bringing into the using library models of providing “free” technology environment is very different from the computing resources. This model was probably way we supported the early adopters. New users appropriate when the computer was a fixed-cost 2 want “complete products.” mainframe and we were trying to promote use of the network. Now that the value of information Multivendor, distributed technology technology is well established, many of us still requires high-level support have not shifted into other economic models that The support burden derives from the increased will help to better manage demand. Outmoded complexity of desktop applications themselves, models cause significant problems, including and, increasingly, from how those applications making it difficult for users to match costs and interconnect. A typical end-user application today benefits, promoting excessive consumption of might involve a desktop computer, a departmental resources, and contributing to the support crisis network and server, the campus network, and a we are experiencing. mainframe information source. The end-to-end information path involves multiple systems and Support quality deteriorates several administrative units. New technologies, such as object linking and embedding, In the absence of a significant change in approach, significantly increase the complexity and the mismatch between demand and capacity interdependencies among applications. Users are produces a death spiral of decreased quality of not satisfied with accessing and manipulating only service and support. Continuing the University of numbers and text. They expect images, sounds, Virginia example, contention for dial-in lines is so and full-motion video. Users also demand fierce that individuals set up attack dialers and transparent interoperability between applications, then camp on in an attempt to assure continued regardless of the operating system or vendor. Two access. During the first month of the academic systems that work just fine by themselves can year, the call-waiting queue at the help desk is longer than the line at a Grateful Dead concert. Information technology staff are frustrated because they cannot meet needs, despite the long 2 Geoffrey A. Moore, Crossing the Chasm: Marketing and hours they invest at work. Once viewed as heroes, Selling High-Tech Products to Mainstream Customers (New York: HarperBusiness, 1991). these same people are now considered 4 CAUSE PROFESSIONAL PAPER SERIES, #16 incompetent because they cannot handle the flood population of users has changed, this practice has of requests. In an attempt to serve, help desk staff become both ineffective and inefficient. It is implement automated solutions to providing help ineffective because no single consultant can that irritate customers who want the personal possibly know all the answers that the wide touch. We do not have time to plan successful spectrum of users requires. The variety of user implementation and support of new technologies information needs spans such a broad range of or to communicate with customers about changes. technology services and computing applications Instead, we rush new services out the door. that users, especially new users, often have Because they are not production-quality services, difficulty communicating the critical elements of they further increase pressure on the help desk, their environments in terms that experts can use to and the spiral continues. answer their questions. As we scale up the number It is easy to understand that if demand is of these questions, it also becomes inefficient to drastically increasing and support resources are use highly skilled experts to provide primary not, both the quantity and quality of support will consultation, such as “Have you turned on the diminish. Given the difference between demand power switch?” Instead we need people on the and supply on most campuses, perhaps the most front lines who are trained in diagnosis, more like amazing observation is that our support primary care physicians, rather than specialists. A mechanisms have not entirely collapsed—perhaps single pool of staff will be less good at answering because some users have developed their own front-line questions for a heterogeneous support mechanisms. population of thousands of users than the same Some say we are approaching crisis, when the number of staff who are divided among the users data indicate that we should actually be in in smaller sets, with the opportunity to learn the meltdown. In addition to the simple mismatch capabilities and environments of their individual between supply and demand, qualitative changes subsets of customers. Thus, we believe that the in the technology environment of our institutions current centralized primary support model in also contribute to the deteriorating level of many institutions is doomed to collapse. support. Exploring the mechanisms and One exception we see is centrally provided, characteristics of technology support will provide subject-specific consulting and service centers, a better understanding of the problem and will such as the Center for Mathematical and Statistical ultimately lead us to a possible solution. Consulting at Indiana University, the Academic Computing-Health Sciences Center at the Centrally provided primary University of Virginia, and the High Performance customer support does not scale Computing Center that is now getting started at Primary support consists of answering user the University of Houston. These centers provide questions directly with the customer. This primary and expert consulting and often system contrasts with secondary or expert support in and network administration, and they seem to be which a group of staff provides information very successful in meeting their users’ needs. resources, tools, deep expertise, and backup for However, unlike general help desks or consulting those who provide the front-line, primary pools, their domain is very focused in content and response. they serve a small subset of the entire range of Historically, when users were fewer and more users on campus. These centers seem to be the homogeneous, the computer center provided most exceptions that prove our rule. of the primary support, often employing expert consultants and programmers to do so. As the THE CRISIS IN INFORMATION TECHNOLOGY SUPPORT 5 Assignment of support responsibility is Every machine in the institution is different ambiguous In the early 1980s, microcomputers were As central resources fell behind in their ability limited in processing capability. Software was to meet demand, users responded in two ways. likewise very limited in functionality. Under these Those who assumed it was the responsibility of the conditions, the choices we had to make central technology organization to provide all of significantly determined the degree to which our the answers and help they required took the machine actually met our needs. No machine position that it was the IT organization’s job to could solve all of our problems, so we chose the teach, rather than their responsibility to learn. one that came closest and that we could put on our Departments and individuals purchased hardware desk soonest. In addition, we funded our and software that did not meet their requirements equipment through donations, grants, begging, and then expected central support to make it all and every one-time method we knew. work. In many cases, the IT organization has Compounded over a decade or so, the result of promoted these misperceptions by attempting to these practices—especially at large research convince the institution that it had complete universities—is an assemblage of equipment, responsibility. Some distributed units, however, software, and configurations that is nearly responded by developing their own support insupportable at any reasonable cost. mechanisms that functioned quite independently of the central services. Many of these have not Central units are merging been able to keep up with the increasing There has been a trend over the past five years complexity of technology or to interconnect their to consolidate academic and administrative idiosyncratic environments with the rest of the computing organizations and, in many cases, also campus and world. The central organization’s telecommunications, media services, and libraries response to these units has often been to simply into a single organization. These mergers are ignore them, relinquishing the ability to influence probably the right thing to do, but they require users’ decisions and learn from their experiences redefinition of identity and responsibilities for (both bad and good). In either case, both central staff and users and major reorientation—both technology staff and the users perceived the technical and cultural. Precisely how this affects quality of support to be significantly diminished. the effectiveness and efficiency of the support staff and the user’s ability to get needed support is not Distributed systems need special support clear, but it may contribute to a sense of Problems with local area networks, desktop frustration. On the other hand, there are synergies client applications, and remote access are and new knowledge that result from these sometimes impossible to troubleshoot from a organizational changes that promise, at least distance. With no one else to turn to, the user eventually, to outweigh the transient problems. resorts to central technology support providers. This no-win situation causes frustration for the Central IT organizations are the customer and serious demoralization for the scapegoat support staff who try valiantly to find solutions from afar. Unable to do so, they eventually make Faculty, administrators, and students have office calls or house calls. This wonderfully followed our lead by incorporating information personalized service typically alienates other technologies into their daily work. We provided customers unless there are enough support extensive personal support to enable faculty providers to handle everyone’s needs this way. pioneers to enhance instruction through