Will The Last Central IT Person Please Turn Off The Lights |-------------------------------------| | Paper presented at CAUSE92 | | December 1-4, 1992, Dallas, Texas | |-------------------------------------| WILL THE LAST CENTRAL IT PERSON PLEASE TURN OFF THE LIGHTS Larry Conrad John Rome John Wasileski Arizona State University Tempe, Arizona ABSTRACT Ten years ago, John Naisbitt articulated the coming of a number of societal changes which were to have enormous impact on organizations at all levels. In this paper, three of Naisbitt's trends are revisited and shown to be closely interwoven with the pressures to move toward distributed computing. We also explore the chaotic behavior these movements exhibit and present the plans being developed at ASU to enable decentralization while maintaining the coordination required to provide continued support for executive level decisions. ------------------------------------------------ Will the Last Central IT Person Please Turn Off the Lights[1] CHAOS "It's not good. We will lose our jobs." --An East German checkpoint guard commenting on the fall of the Berlin Wall.[2] Also spoken by an ASU employee about the changes in AIT. This is the tenth anniversary of Megatrends[3], the book which articulated ten major organizational waves that were beginning to re- shape our society. Three of those are of particular interest for our discussion. They are the movements from: Centralization to Decentralization Hierarchies to Networking Representative to Participatory governance During the last few years we have seen how remarkably accurate these predictions have been. One need simply reflect on the collapse of the USSR to see the results of these trends on an international scale. On the national level, huge companies like Ford, IBM, and others are reorganizing in ways which look remarkably like the movements within the computer industry to "downsize" or "rightsize". The pressures to reorganize the information technology functions at Arizona State University (ASU) provide an example of these same phenomena working on a local scale. The effects of these changes have been nothing less than chaotic for our community as it struggles to adapt to a new order. By a remarkable coincidence, this is also the tenth anniversary of Benoit Mandelbrot's[4] popular publication of his mathematical discoveries in chaos. While not providing any relief from the uneasy feelings reflected in its name, chaos does provide a mathematical means of describing these multi-scale changes. The mathematics also reveals that such changes may well be the result of "...order melting smoothly into chaos."[5] That is, not as a result of some drastic alteration in the way things are done but just a small change in one or more 'parameters' which have input into the process. CAUSE OR EFFECT "What a curious feeling! I must be shutting up like a telescope!" --Alice[6] Information technology was one of the parameters which changed. In fact, Naisbitt and Toffler[7] both credit the innovations in information technology with providing much of the wherewithal to effect these changes. It is ironic that the very organizations which enabled this technology and acted as agents of change should find themselves so soon awash in the waves they helped create. To better understand how these multi-scale movements are affecting computer organizations, consider each of them and their inter-relatedness as they apply to technology organizations. The movement from centralization to decentralization is exactly the phenomenon of distributed systems. The decentralization movement recognizes a limitation of central systems in that they cannot know many of the little details ('small parameters') which have large effects locally. It removes the "high priest" aura many central authorities have cherished and elevates the importance of business concerns over technology concerns. Distributing systems also makes the trend of replacing hierarchies by networks imperative. It is a fact that bureaucracies slow down the flow of information and thereby risk the competitive advantage which information can give an organization. "Channels and cubbyholes"[8] is an excellent description of bureaucratic organizational behavior as well as an apt description of the communication barriers such organizations impose. The trend is to move away from a rigid hierarchy of authority toward an active network of peers. These peers must have expertise in a wide range of subjects, understand specific business concerns, be capable of dealing with informational issues from a technical viewpoint, and be able to keep a broad perspective on a problem. Such an employee is generally educated and highly trained. It is no surprise, therefore, that they demand more direct input into decisions and this is helping to drive the movement toward participatory management. This shift toward decentralization emphasizes the inability of our bureaucratic organizational structures to make quick adjustments when needed. Information technology organizations have become such bureaucratic structures and have suffered the same lack of flexibility. They have also made specific errors of their own. For example, many have: Provided technological solutions which took no account of business needs. Allowed development backlogs to grow to a ridiculous size (hundreds of person-years in some cases) in order to lobby for additional resources. Installed development methodologies which guarantee a three-to-five year development time frame and slavishly stick to them even when it has become clear that speed is essential. Been slow to recognize the opportunities provided by technological innovation and the empowerment these can provide for non-technical staff. During this economically difficult time, executives have responded to this suite of disappointments by making substantial reductions in technology budgets. PATHS TO THE FUTURE "Would you tell me which way I ought to go from here?" "That depends a good deal on where you want to get to." "I don't much care where." "Then it doesn't matter which way you go." --Alice and the Cheshire Cat[9] While it may have been only small changes in certain forces which precipitated things, the changes now required in IT organizations to respond must be dramatic. Every organization wishes to have some part in deciding its future and 'knowing where you want to go' is an essential part of selecting the right path. If your goals include total dedication to customer satisfaction and working to eradicate errors, then help in selecting a path has been provided. James Penrod and Michael Dolence[10] have done an excellent job in capturing the essential principles of change which must occur within higher educational enterprises (and the IT organizations within them). Similar principles have been put forward by Richard Murray and Richard Hardin[11] for changes within IT organizations in business. We provide a joint summary of these in the following: - Provide a strategic plan for IT which is founded upon organizational goals and values and provides a clear and compelling vision for the organization. - Build an IT organizational culture which supports: Innovation Entrepreneurialism Sharing Continuous learning - Maintain a holistic view of the enterprise. - Create a new style of leadership which encourages participative decision making and empowers the people responsible for getting the job done. With this as a backdrop, let us turn to the ASU experience with its recent series of organizational changes and efforts to "decentralize" the academic computing portion of its IT organization. Arizona State University is in the throes of dealing with the results of these movements. In particular, with the shift to distributed academic computing and the shift towards a distributed staff approach to supporting academic computing. These two shifts are only casually related since one could have any combination of them. For example, distributed computing with central support, central computing with distributed support. However, both these changes are targeted at achieving the same general goal: customer empowerment. We are trying to find ways to provide services that give our customers more control over their environment. The present budget crisis has only served to accelerate the need for changes since we must do a better job leveraging the resources which are available. Up to this point, we have been structured in a very typical way: central academic computing with central support. The faculty have been reasonably satisfied with the support delivered over the years but their own budget problems have heightened their desires to have more control over their technology environment. The two key budgetary forces driving these changes are: * The Information Technology department has roughly half the budget for capital acquisitions that it did just two years ago -- we no longer have the funds to buy mainframes; and * The need for computer support locally in the colleges is mushrooming while the ability of the colleges to deliver that support is sharply curtailed -- they are cutting faculty and classes and simply cannot afford to provide additional support services. There are a number of different ways of responding to these imperatives that would maintain the status quo. For instance, we could simply move to a model of buying used mainframes near the end of their useful life cycle in order to reduce the capital requirements. (In fact, we've already done some of that.) Or we could simply say the support problem in the colleges is their problem and just ignore it. (We've done some of that, too!) However, the budget reductions combined with the social forces toward decentralization have encouraged an approach that will help further empower our customers as well as respond to the budget reductions. DISTRIBUTING COMPUTER SUPPORT "A technology that informates can have a corrosive effect on the hierarchical organization of work, but its transformative power finally depends upon a series of crucial management choices." --Shoshana Zuboff[12] New Support Structures. The Computing and Network Consulting Services department within IT (CNCS) is in the process of establishing a more decentralized computer support structure. This support will be centrally managed and academically integrated in order to better meet the academic computing needs of the colleges and to better support the move towards distributed computing. While many members of the CNCS staff will be physically distributed to the colleges, a reduced central staff will continue to provide a suite of key services. In order to ensure we are distributing to the greatest extent possible, we've tried to approach this with the assumption that all services should be distributed ... and only centralize where necessary. Appendix A, shows which services will be retained. Transforming support. The new support structure will be less hierarchical and more of a networked and collaborative structure. In the existing hierarchical structure, customers contact central support whenever they have a question or problem. If the central staff cannot help, they escalate it to the vendor. In the new networked structure, customers will contact their local distributed support person who will either address the problem or contact other distributed support staff for assistance. If they cannot address the problem, an assigned distributed support person will contact the vendor for further assistance with no involvement by the central staff. New tools. We are working on a tool that will facilitate the "networking" aspects of this structure. We need something that will allow customers to electronically ask questions or report problems, allow consultants to respond electronically, capture the responses in a data base, and allow the consultants to easily search the data base. This tool must be intuitive and easy to use. It must allow us to deliver services 24 hours a day, 7 days a week, and be available to the customers as well as staff. Today, our support staff can go to the office next door or down the hall to collaborate on a question or a problem; the tool is intended to serve as an electronic substitute for staff proximity. We hope it will help relieve the sense of isolation about which many of our staff have expressed concern. Phone consulting will continue to be provided centrally. We are planning to use students to staff a bank of "hot line" phones. The students will be trained to handle lower level questions and escalate the more difficult items to the permanent staff. All staff -- both central and distributed -- will respond to calls escalated from the hot line. Centrally Managed. For this concept to work, central management must add value to the structure; otherwise, it would be more effective to simply parcel out the FTE to the colleges. We believe there are a number of crucial ways in which CNCS can add value to the support function: --CNCS can provide consistency in problem solving thereby increasing the effectiveness of the distributed staff. --We can coordinate different approaches to a problem and assure a 'best solution' is communicated to the campus. --We can assure that technically 'right' solutions prevail over technically 'wrong' ones. --We can provide the variety of expertise which is just not possible to attain with two or three people. Some support needs will change over time. A centrally managed staff will allow us to respond to these changing needs. For example, a college may need a networking expert for 2 months to help plan and implement a network. However, after the network is in place, that college may want to concentrate on an instructional technology project for the next 6 months; followed by something else later on. Or a college may need access to someone with Macintosh expertise, but only part-time. Central management will allow coordinated staff assignments to be made with the right experts for the right amount of time and change assignments as needed. CNCS has many staff members with in-depth expertise in a variety of areas. While assigning them to work in specific areas will give them more general responsibilities, central management will allow us to retain and develop their expertise as institutional asset through project and development assignments. Finally, central management will allow us to offer career-paths for the technical staff. Opportunities in any one department or college are necessarily limited whereas we will be able to offer further professional growth and development across the institution and improve ASU's chances of retaining highly qualified personnel whose talents are in high demand. Separating roles. Under the current plan, there are 3 types of support services which can be provided. The appropriate mix of services will be determined for each college: 1) Full-time assignment of a staff member to a college or department to provide on-going local support. 2) Part-time assignment of a staff member to provide on-going local support on a limited basis; or 3) Project assignment of a staff member with in-depth expertise to support a project with a start, an end, and a specific deliverable. CNCS will provide the staff member a personal workstation appropriate to the assignment, basic documentation, and the formal management structure. Each college will provide office space, office furniture, a telephone, and an Ethernet connection. Each college will also provide a college-level coordinator responsible for providing local work direction and for resolving conflicts or issues internal to the college. The leadership in each college will be responsible for selecting the mix of support required for its faculty, staff, and students. In order to deliver the value-added support discussed above, 20% of each staff member's time will be retained to serve in a central capacity. This 20% reserve for central functions is intended to be flexible -- e.g., not one-day-out-of-five -- but can be attained as an average percentage over time with the details negotiated with the college or department. Distributed staff will have central functions. The 20% time reserved for the distributed support staff to serve in a central capacity will be spent in a variety of ways: * Providing "networked" support to other distributed support personnel in other colleges and departments; * Participation in development assignments (projects) in a particular technical discipline which will aid the institution overall and retain and hone their expertise. * Giving and/or attending training sessions in a particular technical discipline which will aid the institution overall and increase their expertise. * Attending staff meetings to remain in touch with other CNCS staff members and keep posted on the latest technical activities. and * Attending departmental meetings where knowledge about what is happening in each college will be shared and a context within which technology is being utilized can be provided. CONCLUSION "But in the time of parenthesis we have extraordinary leverage and influence - individually, professionally, and institutionally - if we can only get a clear sense, a clear conception, a clear vision, of the road ahead." --John Naisbitt[13] ASU has not embarked on a total reengineering program but it is taking its first, halting steps toward a new and uncertain future. Change is always traumatic and a change that involves shifts in paradigms is particularly difficult. Much patience and understanding by both management and staff are necessary to make the transition to the new order successful. Initial staff reaction was shock, followed by denial, a desire to negotiate alternatives, and, finally, acceptance. These are the phases of grief one feels with any major personal loss and it must be recognized that changes of this magnitude are a death-rebirth process. It is our hope that these changes will provide a rebirth for ASU's technology department and a better working environment for everyone. Appendix A: ONGOING CENTRALLY SUPPORTED FUNCTIONS The following functions will continue to be centrally supported: Network Support---Management and support of the network is fundamentally a centralized activity. Distributing management and support of the network would expose the institution to an unreliable and ineffective networking environment. Student Support---Operation and support of the IT Academic Computing Sites as well as the associated student consulting support provide a critically important base of computing resources for our students that cuts across disciplines and colleges. The ability of any student to go to any of the Sites and receive basic services -- regardless of their college affiliation -- and the availability of around-the-clock service make these Sites a key resources for our student community. This flexibility would be lost if the function were distributed. Training Coordination---Coordination and delivery of training is fundamentally a centralized function. Central coordination will provide the ability to pull together the best resources from across the institution for delivering training classes and/or for putting together classes utilizing off-campus resources for the best pricing. Electronic-Mail Support---E-mail needs to be a centralized function in order to ensure implementation of a set of e-mail packages that are adequately integrated and to provide reliable delivery of electronic information. Printing Distribution Support---As computing becomes more distributed, the need to provide convenient and affordable printing to the campus will also grow. All customers should have access to a high volume printer as needed on a convenient and affordable basis. A centrally managed set of strategically located printers will address this need. Site Licensing and Distribution---There will be an increased need for central coordination, negotiation, and distribution of site licenses here on campus as we move to a distributed computing paradigm. Discount pricing and an improved legal standing are the benefits of a centrally managed approach. Facilitate Information Sharing---Another need in a distributed computing environment, will be to provide adequate documentation for commonly used products. Plus, we need to better utilize technology to deliver information and consulting/problem resolution services electronically over the network. A central function is needed to coordinate preparation of the information, distribute the information, and monitor and maintain its overall quality and relevance. Presentation Graphics Production---All disciplines in all colleges share the need to prepare publications and presentations using the most effective tools available whether it be slide, display graphics, or video production. The equipment involved is expensive and requires specialized expertise. This is a function where the economy-of-scale argument still applies. Visualization Lab and Support---All disciplines in all colleges share the need to better understand and interpret the results of their research activities. Visualization techniques can often yield new insights and a greater understanding of computational results. The equipment involved is expensive and requires specialized expertise. This is also a function where the economy of-scale argument still applies. Research Programming Support---There is a critical need for discipline specific programming expertise in the colleges to support the research faculty. Ideally, this support would be provided by distributed staff. However, this kind of discipline specific background can only be provided by specialists in each field (e.g., physics, mechanical engineering) and the limited number of experts we will have also need to collaborate frequently on programming techniques and common technical problems. Therefore, we propose keeping this as a central function, but with a clear assignment for each expert to provide support to one or more specific academic units. Statistical Support Lab---The need for an on-campus focal point to provide consulting and support for statistical methods and software is a critical research support function. This need cuts across disciplines and colleges. This function will be overseen by the interdisciplinary Committee on Statistics. FOOTNOTES [1] Mr. Conrad thanks Rita Conrad for her many helpful suggestions and support in preparing this paper. [2] Newsweek, Special Report, The Wall, November 20, 1989, page 23. [3] John Naisbitt, Megatrends, Warner Books, New York, 1982 [4] Benoit Mandelbrot, The Fractile Geometry of Nature, W. H. Freeman, New York, 1982 [5] Douglas R. Hofstadter, Methmagical Themas: Questing for the Essence of Mind and Pattern, Basic Books, Inc., New York, 1985, page 364. [6] Lewis Carroll, Alice's Adventures in Wonderland, annotated by Martin Gardner, Bramhall House, New York, 1960. [7] Alvin Toffler, Powershift: Knowledge, Wealth, and Violence at the Edge of the 21st Century, Bantam Books, New York, 1990 [8] Ibid, page 165. [9] Carroll, page 88. [10]James I. Penrod and Michael G. Dolence, "Concepts for Reengineering Higher Education", CAUSE/EFFECT, Volume 14, Number 2, Summer 1991, pages10-17. [11]Richard J. Murray and Richard C. Hardin, "The IT Organization of the Future", Information Systems Management, Fall 1991, pages 68-72. Work and Power, Basic Books, Inc., New York, 1988, page 285. [13]Naisbitt, page 252.