This paper was presented at CUMREC '99, The College and University Information Services Conference. It is the intellectual property of the author(s). Permission to print out copies of this paper is granted provided that the copies are not made or distributed for commercial advantage and that the title and authors of the paper appear on the copies. To copy or disseminate otherwise, or to republish in any form, print or electronic, requires written permission from the authors.
Planning for Technology Replacement; Is it Possible?
Mark R. Ritschard, Systems Integrator, Desktop Team Leader
Eugene L. Spencer, Associate Director
ISR/Computer & Communication Services
Bucknell University rises on a hill overlooking the Susquehanna River in central Pennsylvania. Primarily an undergraduate institution, Bucknell University enrolls approximately 3,500 highly selected scholars (3,300 undergraduates and 200 graduate students), who benefit from the wide curricular choices of a comprehensive university and the personal attention received from more than 260 faculty members. Bucknell offers a strong liberal arts program, with 80 percent of the students enrolled in the College of Arts and Sciences, and the remainder in the College of Engineering. In addition to engineering, professional programs are offered in management, education, and music. Interdisciplinary programs, such as those in animal behavior, environmental studies, and Latin American studies, are encouraged. Students work closely with faculty, and many assist professors with special undergraduate research projects.
Replacing computing equipment before it is obsolete is a daunting challenge. Hardware vendors offer the latest technology at a rapid pace and new versions of software quickly take advantage of the new capacity. Are the effective management of technology, a constrained budget and limited technical staff resources mutually exclusive? How are technology replacement plans and "Total Cost of Ownership" related? Is it possible to manage the expectations of faculty and staff?
Many institutions have a plan to replace desktop computers on a periodic basis. Other institutions utilize a process whereby users request new resources against an annual budget or a periodic stream of irregular funding. When IT management steps back and looks at all of the activities resulting from these processes, many problems become evident. The replacement/upgrade processes are inherently costly, labor intensive for technical support staff and too frequently disruptive for users. Such plans also rely on decision-making processes that are often politically charged and too heavily weighted toward individual goals or departmental goals rather than institutional goals. In an attempt to satisfy as many needs as possible, IT organizations have created processes that are often a burden to all involved.
In many cases, institutions have based their decisions on a set of long-standing assumptions that should be reconsidered. Is there a better way? This paper reflects on the variables involved in technology replacement within a context of constrained resources in an attempt to vastly improve the process of computer installations. As a practical example, the Academic Desktop Computing Replacement Plan of Bucknell University is presented for consideration.
Planning for Technology Replacement; Is it Possible?
Replacing computing equipment on a large scale within a relatively short time frame is a fairly recent dilemma. In the not so distant past, it was not uncommon for colleges and universities to replace mainframes after seven to ten years, and only after several substantial upgrades.
Now, being more numerous and more widely distributed than mainframes, the replacement of microcomputers and associated resources poses a larger problem.
For some years, institutions simply purchased more computers to add to their existing inventory. Where newer computers were needed, older computers were replaced and then shifted somewhere else on campus. While this worked for awhile, it gradually became evident that the process of shifting computers begs the replacement question. How long should computers stay on campus? Should any given computer be replaced with new or used equipment?
In 1995, the Desktop Team of Computer and Communication Services (now part of the newly formed department Information Services and Resources) at Bucknell University began to look at the issues involved in technology replacement and the various areas of responsibility and oversight involved.
Why Have a Plan?
The plan described in this paper grew out of three major concerns. Faculty members of the Committee on Academic Computing (CAC) frequently expressed concerns that a formal plan did not exist for the ongoing replacement of desktop computers. The faculty wanted to see a replacement plan comparable to the original "Computer On Every Desk" policy, the original computer allocation plan of the university. Furthermore, the decision-making process for allocating constrained resources was difficult. Our senior administration was concerned with the rate of growth in the annual budget request for desktop computing; where would it all end? IT
staff were concerned with the growing workload of new installations and the endless stream of computers to be redeployed (often called "trickle-down").
Taken together, the above reasons encouraged the Desktop Team to develop a scheme for planned technology replacement. The resulting plan has helped solve the original concerns of the three constituencies mentioned above (faculty, senior administration, and IT staff) and has also provided some unexpected benefits.
To aid other colleges and universities that are interested in a planned replacement scheme for technology, factors will be presented that are important to consider. Bucknellís replacement plan is then presented with a word on how each of the factors is addressed.
The following factors are key players in replacement decisions. Each one should be examined when considering implementation of a replacement scheme.
Funding. What is the current funding commitment from the administration and trustees? Does the administration understand the impact of varying funding levels for replacement? Is funding centralized or decentralized? Do the current funding mechanisms match the goals of technology replacement?
Budget. Once a budget is in hand, how should it be managed to carry out the plan?
Jurisdiction. Are computer funding and purchasing aggregated for the entire university or is there departmental autonomy? If budgets are separate, should purchasing be centralized? How is the university as a whole affected by the current strategy?
Computing Power. How does the institution ensure that faculty and staff have the necessary computing power?
Acquisition. How do various members of the campus community acquire computing equipment? What role might hardware standards play in the acquisition process?
Allocation. What group(s) weigh(s) needs and resources to wisely allocate computing equipment to best meet institutional goals?
Distribution & Installation. Where does equipment arrive on campus? How is it set up and transported for installation? Who is involved in the process?
Reputation. How is the institutionís reputation affected by installed technology among prospective customers, prospective faculty, and other constituencies?
"Junktion". How does the institution dispose of equipment that "drops off the bottom"?
Although a budget limits technology purchases, a budget should not limit planning for computing equipment replacement. Remember that a budget is a fixed amount, usually set annually, that determines purchases for a given year. However, implementation of a planned replacement schedule will depend on the broader issue of technology funding -- the process of setting long-term commitments for amounts spent on computing equipment across the institution. First, a broad look should be taken at current funding methods and levels. Some questions to consider: Do individual departments or divisions fund their own equipment? Who is responsible for technology funding levels across the institution? Do the trustees and the senior administration understand the impact of various funding levels? Implementation of a successful replacement scheme will depend on cooperation from many members of the institutional community and the willingness to take a fresh look at institution wide funding commitments for technology. However, once a successful case is made regarding the benefits of a replacement scheme, funding discussions will go more smoothly. A technology replacement plan makes long-term funding more predictable and stable.
Funding solutions will, in the end, produce an annual budget. Although funding commitments should be consistent, the annual budget may change, either in amount or in distribution. Variables to consider when developing the budget include average replacement costs, anticipated upgrade costs, funds for unanticipated new equipment, and funds for new faculty and staff. Careful analysis of the budget and annual needs will facilitate the replacement scheme.
Regardless of the where the current responsibilities lie for funding and purchasing, it may make the most sense to take a step toward a replacement plan within the current framework. Obviously, if the institution currently has centralized funding and purchasing for computing equipment, an institution-wide transition to a replacement plan makes sense. Conversely, if there exists a fair amount of departmental of divisional autonomy, it may be more appropriate to make steps toward replacement plans within those divisions. However, it is important that all of the factors are considered and to move toward a plan that is commensurate with institution-wide goals. Take a hard look at the effect of the current model. If changes should be made, present a good plan with good justification to those who are able to effect changes. Demonstrating how the institution is affected in each of the factors discussed here, will help guide those in the decision-making role.
Any successful technology replacement scheme must take into account the need for individuals to have the necessary computing power. To state the obvious, necessary computing equipment is not always the same as desired computing equipment. Clearly, there are faculty and staff who need greater computing capacity than others and computing equipment for such individuals will need to be replaced more often. Herein lies one of the greatest challenges in the world of computing. The power of yesterday is the weakness of tomorrow. The trap of obsolescence is open wide, but a planned replacement scheme can accommodate needs if wisely implemented. The challenge is to meet the needs for more frequent replacement within the framework of a replacement scheme that is realistic for the majority of the equipment on campus.
Though it appears straightforward, the acquisition of computing equipment should not be taken lightly. Most obviously, aggregation of purchases leads to cost savings. However, before purchases are made, research is important to ensure the institution is making a wise commitment in technology. Organizations like The Gartner Group can help institutions think through the long-term issues involved in technology acquisition. Another long-term consideration is hardware standards. Developing some broad standards for various pieces of hardware may provide the institution with an established framework that will expedite some purchases. Other variables include vendor negotiations, when and how often purchases are made, and coordination of purchases.
Limited funding always leads to competition for computing resources. How the resource allocation decisions are made are critical to a successful replacement plan. When distributing equipment, it is crucial to consider broader goals rather than focusing on individual computing power needs. How can equipment be allocated to facilitate computing support? Is inter- or intradepartmental collaboration important? In what ways can technology be distributed to enhance the sharing of information? How should faculty computing equipment be allocated in light of computing equipment available to students? Answering the above questions provides insight into the allocation needs that will need to be incorporated into a replacement plan.
Distribution & Installation
Geography, campus transportation systems and available staff are key elements in the installation and distribution of computing equipment. Should the trucking companies deliver to one or more locations? What space is needed to receive and process the equipment? Who will personalize the equipment for your campus? Once ready for delivery to offices, how will the equipment get there? Who is responsible for delivery? Who is responsible for the actual installation and network connectivity? An efficient replacement scheme addresses each of these questions.
Though it may not be of primary consideration, it is important to look at the reputation of the school that is either desired from or delivered by the replacement plan. Will it attract good faculty and staff or encourage them to look somewhere else? How will it affect the way prospective students view the institution? A replacement plan will probably not redefine the institution, but it will certainly contribute to its overall reputation .
It canít be ignored: a scheduled replacement scheme needs to include a transition plan for "junk", the equipment that is no longer needed on campus. Is any of the old stuff needed for parts? Can it be used in classes for hands-on laboratory learning? Should it be sold, donated, or junked? What institutional policies or granting agency policies limit or define how equipment can be disposed? Old equipment can be a headache or an opportunity; having a replacement scheme in hand will increase the opportunity.
Combining the Factors
Technology replacement is potentially an overwhelming issue. However, Bucknell found that taking the time to examine each of the above factors and design a plan is much better than approaching replacement in a piecemeal fashion.
Institutions will want to weigh each of the above factors for relative importance and applicability. Look for ways to work as much as possible through existing structures and policies. An initial plan will be refined, and the refining process will help guide changes in other areas as people realize the benefits to be gained from a scheduled replacement plan.
The Bucknell Plan
The Desktop Computing Replacement Plan began as a cooperative effort between the CAC and the Desktop Team of CCS. The Desktop Team proposed a scheduled replacement scheme to CAC, a committee comprised primarily of faculty. Working with CAC was of key importance in implementing the new plan, chiefly because the committee provided credence for the plan among the faculty. Strong support from the faculty from the beginning enabled the planís development.
The current Desktop Computing Replacement Plan took approximately two years to develop and three years to implement completely. It is a four year cyclical planned replacement scheme for all desktop and departmental computing equipment. In brief, any given department is scheduled to receive new desktop computing equipment every four years and requests may be submitted for upgrades or replacements every other year. For the complete text of the plan, see the section entitled Academic Desktop Computing Replacement Plan.
A key component to the success of the replacement plan is replacing equipment by department. All of the desktop computing equipment is replaced in a department when it is up for new equipment every five years. Bucknell replaces equipment on a department-wide basis to facilitate interdepartmental communication and pedagogical parity for faculty. When developing an overall replacement scheme, departments were grouped by function to facilitate intradepartmental communication.
Not all of the replacement factors mentioned above were considered initially. However, each one has been considered as the plan has been refined. How Bucknell has dealt with each factor is recorded below as a potential aid in approaching other institutional situations.
Replacement Factors and the Bucknell Replacement Plan
Technology funding at Bucknell University is set by the senior administration within the context of the overall university strategic planning process.
Technology replacement information was provided by the Desktop Team to apprise the administration of funding required to maintain an annual replacement schedule for all faculty and staff of the university. One of the biggest advantages of a scheduled replacement plan is the predictability and stability of the annual budget, an important aid in long-term fiscal planning.
Centralized funding was an already existing function of the university, a condition that has played a major role in the success of the replacement plan for Bucknell. It allows for many different facets of technology at Bucknell to be considered together within the broad context of technology funding.
Bucknellís annual technology budget is developed by a team of individuals representing the various technology organizations on campus. The Desktop Computing Replacement Plan budget has five components:
Bucknell is often asked, "How big is your budget?" or "How much do you spend per computer?". The clear implication of the question is that the size of the budget or the price of new computers determines the success of the replacement plan. While these two factors play an important role in the plan, they do not, by themselves, make or break a replacement plan.
When the Desktop Team implemented the plan, it was started on the existing computing budget. They key was how the purchased computers were distributed.
An institutionís current budget will determine an initial de facto replacement cycle. For starters, determine the capital value of currently installed desktop computing equipment (computers, printers, etc.), e.g., $1,000,000. Divide that value by the annual funds available for computer replacement, say $200,000. In this example, the replacement cycle is 5 years.
Now what? Take that number to those who can effect a change. Trustees? Senior Administration? the CFO? Is this the position the institution wants to be in? If not, what will it take to change? Thatís where the rest of this plan comes in. A well-designed plan can help justify increased funding for an appropriate life-cycle.
Historically, funding for computers and purchasing at Bucknell have been largely centralized. To move toward an institution-wide replacement plan, the Desktop Team started with all those departments already involved in the central process. As the merits of the plan became apparent, departments with autonomous funding have been folded into the plan as appropriate. In every case, purchasing has now been centralized and standards applied to all departments. Although funds come from different sources, all departments purchase standard desktop computing equipment through the annual process. Good communication and cooperation is key to encouraging separately funded departments to participate in the annual plan. As the plan becomes better understood by the institution, it may make sense to take another good look at current funding models.
Bucknell Computing Power
Based on committed funding levels, the Desktop Team developed the initial Desktop Computing Replacement Plan as a five-year scheduled replacement scheme. The plan was just recently reduced to a four-year cycle. Hence, every four years, any given department will receive new desktop computing equipment. (See the section entitled Academic Desktop Computing Replacement Plan for specifics.) However, the plan accommodates more intensive computing needs by allowing faculty and staff to request new computers as often as every other year. The "additional new machines" line in the above budget section provides funding for such requests.
One of Bucknellís initial actions was to make acquisition, as much as possible, a one-time annual event. Almost all new campus equipment (approximately 95%), including desktop computing equipment, public student facilities, and research equipment, is purchased together. Aggregation on a campus-wide basis has provided at least two benefits: purchasing larger quantities decreases the average cost per machine and purchasing once per year standardizes hardware for each year. In anticipation of needs that may arise during the academic year, extra equipment is purchased ahead of time. On the surface, this could appear to waste resources, by having "old" technology sitting unused for a few months. However, the support and repair savings involved in having a single platform each year have outweighed concerns about stocking equipment. Over the course of the five years of the plan, the result is to have, at most, five different kinds of "WIntel" machines and five different kinds of Macintoshes on faculty and staff desktops. Having seen the benefit of computing hardware standards, the Bucknell Desktop Team has applied the same idea to other hardware, such as printers and scanners.
Before the actual purchases, the Bucknell Desktop Team makes use of available research to help determine the best solutions for the upcoming year. Audio conferences and on-line resources have proved to be valuable tools.
In the past two years, Bucknell has purchased Dell and Macintosh computers, Hewlett Packard printers, and Sun workstations. However, several vendors of "WIntel" computers are considered each year, keeping in mind the support costs associated with switching vendors too often.
The long-term departmental allocation plan was matched with an annual request process. Requests are submitted by department, one request form per department. The requests are then reviewed for technical completeness by the Desktop Team. Final approval and allocation decisions are made for academic departments by CAC and for administrative departments by the Administrative Computing Committee (ACC). The request process wraps up in April so that all equipment can be ordered and ready for installation by the end of May.
Bucknell Distribution & Installation
Student employees are a key element in the timely set up, distribution, and installation of equipment on campus. Students are responsible for unpacking the equipment, personalizing it for Bucknell (installing standard software), and delivering it to its final destination. One staff member oversees the process. Once the computers are delivered, they become the responsibility of the IT organizations support staff for that particular department. Although students often work through the final installation and network connectivity, staff are there to make sure that files are transferred correctly and that a good transition is made from the old to the new equipment.
As an idea of the effort involved in distribution and installation, an average of 40 machines a week over a period of 8 weeks were distributed and installed through the combined effort of 7 staff members and 8 student employees.
Although the reputation factor has been touched on briefly at Bucknell, there is still work to be done in considering the impact of the Desktop Computing Replacement Plan. To address the needs of new faculty, they are often provided with new equipment, provided by the "new faculty" line in the annual budget.
Once a year, equipment that is no longer needed on campus is offered to the campus community during a one day warehouse sale. Those who purchase equipment are required to sign a sale agreement to accept liability for any circumstances that may result from the use of old equipment. Items that are not sold at the sale are then aggregated into lots and offered for bid to used equipment vendors and resellers. In response to some problems associated with the campus warehouse sale, the Desktop Team is currently testing a sale on campus utilizing the World Wide Web.
Academic Desktop Computing Replacement Plan
To start the implementation of the Desktop Computing Replacement Plan, all computing equipment on campus was divided into three categories: desktop, departmental, and laboratory and teaching facilities. The Desktop Computing Replacement Plan was then implemented to address the planned replacement of desktop computing equipment only. It has since evolved to include departmental computing equipment.
Once the focus was on desktop computing equipment, the following steps were taken to develop the plan:
1. All of the academic departments were placed into one of five years, 1995-1999; this created the initial five year schedule.
2. Standard machines (Macintosh, "WIntel", Workstation, Printer) were determined for years 1990 through 1994 based on what was purchased during those years.
3. As the years progressed, new equipment was purchased for departments according to the initial five year schedule.
4. Departments not getting new machines were "standardized" with machines returned from the departments getting new equipment. Machines were only redeployed according to the standard determined for five years previous to the initial replacement year. For example, the Macintosh standard for a department whose initial replacement year was 1998 was a Macintosh Centris 650, the Macintosh model purchased in 1993. When Macintosh Centris 650ís were replaced prior to 1998, they were only redeployed to departments whose initial replacement year was 1998. This prevented multiple future moves for a single machine.
Using the above implementation scheme, the Desktop Team was able to fully implement the Desktop Computing Replacement Plan in three years. In other words, all departments had either 1) new machines or 2) machines that were equal to or better than the standard for their department.
The following five sections are the Academic Desktop Computing Replacement Plan of Bucknell University. Each year of the plan is named: Year 1, Year 2, Year 3, Year 4.
The Desktop Computing Replacement Plan does not imply an expectation that all computers will last five years, nor does it assume that computers can be installed and not moved or touched for five years or that all faculty and staff can get five years of usefulness from a computer. However, minimizing the number of machines that are replaced more frequently than every five years has great advantages.
Every four years, a department will be upgraded with new standard computers, printers, and software. That year will be considered Year 1. After Year 1, the department may submit requests two years later for upgrades, new computers, additional peripherals, or software. All scheduled requests, i.e., requests that are submitted as part of the Replacement Plan, regardless of year, must be submitted as a department. Any requests submitted by individuals will be considered out-of-cycle requests.
The Replacement Plan is for desktop computing equipment only, not for faculty/student research equipment, nor does the plan address classroom, teaching and open access computing facilities.
All full-time faculty and staff will automatically receive the standard new desktop machine and all departmental network printers will be replaced. Computers will be replaced with a like platform, e.g., Mac for Mac, unless otherwise requested. During the spring of Year 1, a request form will be sent to all academic departments, in which the department will be informed of the expected standard configurations for Windows computers, Macintoshes, workstations, printers, and accompanying software. Any requirements beyond the standard configuration must be submitted on the departmental request form. This includes more memory, larger hard drives, new peripherals, software, etc. If no request is submitted, each full-time faculty and staff will receive the standard computer for that year. Where possible, desktop computers for retirees, faculty associates, and adjuncts will be replaced with redeployed equipment.
It is very important that faculty plan ahead and attempt to request equipment that, as best as can be determined, will last for four years. At a minimum, faculty need to request computer hardware that will last at least two years without upgrades or additional peripherals, as little funding is available for upgrades in Year 2.
Replacement of existing peripherals and other special equipment will be evaluated on an individual basis. Justification will need to be provided if replacement for such equipment is requested.
Any individual may choose to not receive a new computer, with the understanding that a request to replace the current computer may not be submitted until Year 3, two years later.
Requests to fulfill emergency requirements may be submitted, but must be very well justified, as funds for Year 2 are extremely limited. Department Chairs will be asked to justify all requests.
Requests for memory, hard drives, other upgrades, software, or additional peripherals may be submitted. Some faculty need to keep abreast with the latest technology necessitating requests for new computers. However, requests for new equipment must be well justified as the resources for such equipment are minimal.
It is very important that faculty plan ahead and attempt to request any necessary upgrades, peripherals, or other computing equipment that will be needed in the next two years. Existing computing equipment should be augmented, if necessary, to last until the next new computer arrives in two years.
Requests to fulfill emergency requirements may be submitted, but must be very well justified. Funding for requests in Year 4 is extremely minimal because it is expected that faculty plan ahead in Year 3. Department Chairs will be asked to justify all requests.
Several benefits were expected from a scheduled replacement plan, the very reason the plan was initially developed. However, there were also some surprising unanticipated benefits. In addition, the institution as a whole also realized important benefits.
Were the anticipated benefits realized? Each anticipated benefit appears in boldface below.
Another factor that has contributed to the success of the plan is the high cost of switching to a new machine, both in terms of productivity and time. The knowledge that a new machine is arriving combined with the costs associated with replacement often makes individuals more willing to wait.
It is noteworthy that, since 1995, only one computer has been requested outside of the five-year cycle.
Due to the remarkable success of the plan, several benefits have been realized for the administration of the IT organization and for the senior administration of Bucknell:
Total Cost of Ownership
The Gartner Group recently introduced the term "Total Cost of Ownership", or TCO, that quickly caught the attention of the computing world. The phrase embodies the goal of reducing the overall costs associated with computing equipment for the entire time that it is used by an institution. Although hardware vendors offer some mechanisms for reducing TCO, The Gartner Group identified TCO as a way for those making purchasing decisions to emphasize the long-term costs to management and to help identify those long-term costs. During the audio conference, Faster, Better, More Cost-Effective Desktop Selection with The Gartner Groupís Tim Ogden on January 19, 1998, it was maintained that TCO is totally dependent on the actions taken by the purchasing organization and that vendors have little or nothing to contribute to reducing TCO. In light of these comments, the Desktop Team identified the following ways in which the Desktop Computing Replacement Plan has enabled Bucknell University to reduce its TCO for desktop computing equipment.
As time progresses, the Desktop Team will continue to look at other ways that savings can be gained through the Desktop Computing Replacement Plan.
Although many compelling arguments have been made in this paper, contemporary wisdom may say, "This plan will not work on my campus!" Indeed, there were significant doubts in the IT administration about the potential for success with this plan at Bucknell. However, faculty demand for a plan (any plan!) and staff demand for a "better way" caused the plan to grow from a grass-roots level.
Despite our initial concerns that such a plan might not be affordable, it is now clear that Bucknell can not afford to not have the plan. Our concerns about our faculty being too demanding to be supported by this plan have been relieved by finding that faculty demands actually change in the presence of a good replacement scheme. And finally, the time required to develop such a plan has provided time savings for all involved.
Departmental Computing Equipment -- computers (and any internal devices or boards), external peripherals, and printers that support the work of the department as a whole.
Desktop Computer --a computer that is used primarily by a single faculty or staff member.
Desktop Computing Equipment -- computers (and any internal devices or boards), and external peripherals, that are assigned to a faculty or staff member for his or her individual use. Networked printers supporting the ongoing mission of a department are also considered desktop computing equipment.
Laboratory and Teaching Computing Equipment-- computers (and any internal devices or boards), external peripherals, and printers that are used primarily in a laboratory or classroom setting for instruction, classroom activities, or faculty or student research.
Obsolete Computer Equipment -- computing equipment that can no longer perform the expected function because of hardware technology or the inability to run appropriate software. Obsolescence is not necessarily related to the age of the computer.
Redeployed Computer -- a computer that is removed from one location and installed in another.
Standard Computer -- the computer that is purchased en masse each year to satisfy The Desktop Computing Plan.