Distributed Computing with Centralized Support Works at Brigham Young University Copyright 1992 CAUSE From _CAUSE/EFFECT_ Volume 15, Number 4, Winter 1992. Permission to copy or disseminate all or part of this material is granted provided that the copies are not made or distributed for commercial advantage, the CAUSE copyright and its date appear,and notice is given that copying is by permission of CAUSE, the association for managing and using information resources in higher education. To disseminate otherwise, or to republish, requires written permission.For further information, contact CAUSE, 4840 Pearl East Circle, Suite 302E, Boulder, CO 80301, 303-449-4430, e-mail info@CAUSE.colorado.edu DISTRIBUTED COMPUTING WITH CENTRALIZED SUPPORT WORKS AT BRIGHAM YOUNG by Kelly C. McDonald and Brad Stone ************************************************************************ Kelly C. McDonald is Executive Director of University Computing Services at Brigham Young University, a position he has held since September of 1990. Previously he served as Associate Executive Director and Director of Planning and Development. He has worked and published in the computing and data communications industry for the past twenty years. He occasionally teaches data communications and computer science courses and holds an MS degree from Brigham Young University in computer science. Brad Stone is the Computing Support Coordinator at Brigham Young University, a position he has held since February of 1991. Over the past year and a half, he has organized computing support campus-wide and has tied various support organizations together into a support model called the Computing Support Representative (CSR) program. He holds a BS degree from Brigham Young University in electrical and computer engineering. ************************************************************************ ABSTRACT: Brigham Young University, like most institutions, has experienced rapid growth in the use of distributed computing systems. However, the maintenance and support of distributed computing has not kept pace with its growth. Last year a support method known as the Computing Support Representative (CSR) program was implemented. This program is patterned after a national business franchise, which provides the support and training of a centralized administration, but allows each unit to operate much like an independent small business. At Brigham Young University (BYU), we have observed an interesting evolution in computing support. In the late 1970s and early 1980s, campus computing revolved around one or more central computers. Computing support mirrored our highly centralized computing technology with computing personnel concentrated in a centralized support organization. The advent of the personal computer as a serious research and administrative tool and the evolution of departmental information systems has brought a migration of computing facilities into the various organizations and departments. Likewise, computing support of the mid 1980s mirrored the decentralization of computing technology. However, a small centralized group was left to tend the remaining systems that continued on the campus mainframe. In most cases, the departmental support groups were autonomous. They made their own purchasing decisions and supported their employees with little if any University coordination. When sharing of databases or other resources was necessary, these groups interacted directly with each other and invented arbitrary interfaces to deal with the interchange. In 1987, the University began to install a computer network to support some of these resource-sharing applications. This network evolved into a campus-wide network which we call Y-Net, or Brigham Young University's internet. In the early days of this network the support structure was very much decentralized in that Y-Net simply interconnected these diverse and autonomous departments. However, as the network became more and more useful, the computing environment that the average faculty or staff member dealt with became more complex. The answers to questions could no longer be found in word processing or spreadsheet manuals because the interaction with network resources complicated their environment. We saw the support problem growing more complex. In 1990, we realized that more and more of the critical University business was being conducted via the campus network. We needed to solidify and "productionize" the support that was given to the networked computing environment. One alternative was to re-centralize the computing support personnel to provide the coordination and economy of scale necessary to accomplish the task. However, many support organizations had become quite autonomous and supported many employees and stable applications within their organizations. Such a radical re- centralization would be disruptive to University business and would probably cause more problems than it solved. Because of these decentralized support groups on campus, we chose to model our computer support structure after the computing technology that we were dealing with. Thus we developed a networked computing support team that mirrors the computing environment that is rapidly growing at BYU. Networked organizations are prevalent in the service industry. An example is a national franchise, where advertising and distribution are supported nationally, but the local business is independent and autonomous. We have used these networked organizations as the model for our computing support. Networked Computing Support Team BYU has chosen the client-server computing model to create its networked computing environment. This means that each computing user utilizes a client computer (such as a personal computer) attached to the campus network to access resources provided by server computers that are also attached to the campus network. Using this same client-server model for the support structure implies that each computing user utilizes a specific support person (Computing Support Representative or CSR) who has access to the computer support network and, through it, access to all the other CSRs who may be serving other clients. The CSR is responsible for solving computing problems that cannot be solved by the faculty or staff user. The key concept is that the CSR has "ownership" of these problems and uses his or her access to the support network to resolve them. Historically, many departments already have individuals on their staff who provide computing support but do not have formal access to a support network. Others have utilized the Computer Consultation Center (the BYU- provided computer support center) as their CSR; however, they do not develop continuity with any individual consultant and thereby spend much time rehashing problem information. Finally, University Computing Services (our central computing support organization) provides contractual support to other computing entities who do not want to hire individuals, but would rather contract their CSR service. The client-server model ties all these CSRs together into a formalized service organization which provides first-level support, as shown in Figure 1, with a "one call does all" philosophy. It emphasizes the concept of problem ownership by the CSR in order to dispel the unhappy aura which surrounds employees who do not know what is wrong or whom to call. And, finally, it creates a knowledge network which makes the "whole" greater than the "sum of the parts." [FIGURE 1 NOT AVAILABLE IN ASCII TEXT VERSION] The second level in the support team consists of expert problem solvers in specific areas of computing. These are individuals who can find solutions to problems that the first-level CSRs are not able to solve. They also have the skills necessary to deal with the technical support staff of vendors, to completely resolve problems that may involve that vendor's equipment. For the CSRs to find answers to tough problems, they must work with third-level support and off-campus vendors. At BYU, we have a number of third-level support organizations: Data Communications Systems (responsible for network wiring), Electronic Media Department (hardware service contracts and repair), and University Computing Services (mainframe, database, Unix, etc.). The major advantage of this three-leveled support team is to reduce the burden of the technical support people and give them more time to resolve problems. This keeps them from having to field so many trivial questions. The clients of the second-level problem solvers are the CSRs at the first level. This effectively reduces the support burden of the second-level experts from several thousand employees to less than 100 CSRs. An informal and unofficial structure of this nature has actually existed on our campus for many years. Support staff in various departments have had an unofficial list of whom they might call to obtain answers to technical questions. However, coordination has been haphazard, and isolation has still been prevalent. Therefore, another important concept of the networked computing support team was the addition of a Computing Support Coordinator to ensure that effective communication and problem management occurs between all support levels. The Computing Support Coordinator The duties of the Computing Support Coordinator, an administrative position in University Computing Services, are to: 1. Act as a problem manager to track problems that get escalated from the first level to the second level. This is to ensure that no problems get "lost" and that solutions are properly cataloged for future reference by first-level CSRs. This is accomplished by utilizing a problem-tracking system that keeps track of work accomplished on a problem. It also sends status information to all interested parties via electronic mail. 2. Coordinate communication between the CSRs and the technical support staff. This does not mean that all problem calls from CSRs must funnel through the Computing Support Coordinator. Rather, the Computing Support Coordinator acts as a "nameserver" in the network of CSRs when the CSRs don't know whom to call with a given problem. This communication is accomplished through electronic mailing lists and bulletin boards, as well as telephone contact. This provides first-level CSRs with two-way communication and solutions to problems or other contact to give them the feeling of membership on the computing support team. 3. Train CSRs in how to interact with the computing support structure of the University. This is done through monthly CSR meetings that focus on a topic of general interest. Additional specialized training is provided through other sessions that focus on more specific topics. These meetings not only help train the CSRs, but also bring them together so they can learn more about the various skills of other CSRs. This helps to foster a more networked support environment. 4. Publish periodic newsletters. This helps to keep the CSRs informed of new methods of support and appropriate use of computing technology for their clients. The newsletter also provides a forum so that CSRs with particular specialties can contribute to the computing support network by publishing articles about their own expertise. 5. Follow up on problem resolution. The Computing Support Coordinator contacts CSRs, as well as employees, to get feedback on how computing support is being perceived on campus. This acts as a quality control indicator to keep the computing support structure service- oriented. 6. Focus the technical staff on problem solutions. Individuals with in- depth knowledge of particular computing technology sometimes become so distracted with investigating solutions that they tend to lose perspective as to what the end result should be. The Computing Support Coordinator helps to channel this energy in productive ways. The Computing Support Representative Employees have two basic computing support needs: to have their computer problems solved and to gather information about their computing environment. To satisfy the employee's need for solutions and information, the Computing Support Representative acts as a channel to network the employee to campus computer support. The CSR is appointed by the college dean or organizational head to fulfill the employee's need for solutions and information. The CSR needs to be someone the employee can trust for reliable information. This can be anyone--from the "local computer guru" to a skilled secretary. We have found that it is important that the CSR be a full-time employee. No two organizations do things in exactly the same way. For the CSR program to be effective in each organization, it must be specifically designed to integrate into the way the organization operates and include any existing support structure that may already be in place. If either the administration or employees feel awkward about using the CSR program, they won't use it and the program will fail. There are many considerations when attempting to integrate the CSR program into an organization, but the most important is to listen to the needs of the organization and ask, "What is it that these people really need?" If an organization has computers, then they will have already created some method for supporting them, even if it is inefficient and undocumented. Support may range from a faculty member who knows a particular program to a team of network administrators. By understanding what the organization has in place, the CSR will be better able to determine their future needs, and how the CSR program can benefit a particular organization. CSR as a channel for solutions The CSRs are the employee's resource for resolving computer problems. The CSRs do not have to be highly trained computer problem solvers. The main requirement is that they be people who can handle responsibility in a way that fosters employee trust. The CSRs should at least be able to understand the employee's computing environment. This means that they should know what applications the organization uses, what sort of network configuration the employees are connected to, and what kind of computers are being used. The CSRs are given strong support from the second-level support team to help them find solutions to computing problems, and should never be in the awkward position of being responsible for solving a problem, but having no way to find a solution. CSR as a channel for information In addition to being a channel for problems and solutions, CSRs need to be a channel for transmitting information to the employees in their organization. This information can be anything that will help employees be more effective in using their computers. We let our CSRs know about changes in the campus network, site licensing of software, computer vendor presentations, known software bugs, vendor recalls of defective hardware components, common problems and solutions, campus computing standards, and any other information that will help keep employees informed as well as compatible with the rest of the campus. This information can be distributed in a number of ways. We especially encourage the use of electronic mail because of the relative ease and speed with which the information can be distributed. We require that every CSR have a mailbox for electronic mail, and strongly encourage the CSR to use electronic mail to distribute pertinent information within the organization. We have also found that regular newsletters are a good way of distributing information both to the CSRs and within an organization. It is important to note that using the CSR as an information channel works both ways. Since the CSRs are located in the organization, they have a good idea of the needs of their employees. This information is very important to the campus administrators who make decisions that affect the direction of campus computing. As an example, we sent out a survey to the CSRs asking what software their organization planned to purchase. We found that the University could save a great deal of money by purchasing a site license and distributing the software to everyone who needed it. CSR networking and training Each CSR has a particular specialty that they know better than other CSRs. By networking their collective knowledge, we are able to solve problems much more quickly. We have set up a listserv electronic mail group and a Usenet news group to help facilitate this process. In addition, we encourage our CSRs to interact at the monthly CSR meetings. CSRs are also networked to third-level support and to venders. When the CSR needs specific help from a third-level support organization or an off-campus vendor, and doesn't know whom to call, they can contact the Computing Support Coordinator, who can refer them to others for additional help. Also, when the third-level support organizations need to give information to the campus employees, they can work through the Computing Support Coordinator. The CSRs are the heart of our networked support program; the more knowledgeable and experienced they are, the better the system works. The Computing Support Coordinator teaches a workshop for newly appointed CSRs that covers the basics of all the subjects, and familiarizes them with how the program works. Over the past year we have focused on ways to better train the CSRs. This was done by separating the monthly CSR meeting and the training workshops. The monthly meeting is kept as short as possible--always no more than an hour--to accommodate the busy schedules of the CSRs, and only the most important information is discussed. In the training workshops, the second-level support personnel teach a workshop in their specialty. The CSRs can sign up for a subject, and when that workshop ends they can sign up for the next most important workshop for them. In the future, we will attempt to have some of our most knowledgeable CSRs teach workshops to other CSRs. This will help to foster the networking concept and get the CSRs working with each other. CSR Setup Models How can a CSR program be set up in an organization? We offer the following suggested models, stressing that for this program to be effective in a particular organization, _it must be tailored to fit what the organization feels is most comfortable._ In a relatively small organization, where there are thirty to sixty employees, a single CSR can coordinate the functions of resolving problems and distributing information (see Figure 2). For a medium-sized organization, where there are a number of employees in each department, we have seen two main models. One has a single organization CSR appointed with separate "department CSRs" working under her or him (see Figure 3), while the other has a separate CSR appointed for each department (see Figure 4). [FIGURES 2, 3 & 4 NOT AVAILABLE IN ASCII TEXT VERSION] In the first model, the Computing Support Coordinator can transfer information to the organization CSR who transfers it to the department CSRs. When the employees have a problem, they can contact the department CSR for solutions. If the department CSR cannot provide a satisfactory solution, he or she can work with the organization CSR who in turn can work with the campus CSR network. In the second model, each of the CSRs acts as an organization CSR who directly interfaces with the Computing Support Coordinator and the second-level support, rather than funneling through a single organization CSR. We have seen both models work well. At BYU we have a few large organizations with 200 or more employees. For this type of situation, the basic CSR model can be modified to have the organization CSR responsible for channeling information while an organizational support hot-line can be used to resolve problems (see Figure 5). [FIGURE 5 NOT AVAILABLE IN ASCII TEXT VERSION] In an organization of this size, the CSR has to use efficient means of distributing information, and we strongly recommend the use of both electronic mail and a regular organization computing newsletter to the employees. Since the CSR is the person channeling the information to and from the employees, they will view the CSR as a friend from within the organization, and feel secure with their computer support. When employees have a routine computing problem, they can call the organization support hot-line, which needs to be tied in with the rest of the campus computing support network. This will avoid taking valuable time away from the CSR, who has other important responsibilities. If the support hot-line does not give the employee enough satisfaction, the employee can contact the CSR to have the support hot-line problem resolved. We tried a model like this for three months and found we could support over 160 employees with one full-time CSR and a hot-line that was staffed with a trained student from 8:00 am to 5:00 pm. Advantages of the CSR Program Most of the advantages of the CSR program come from understanding the needs of the employee. By having the CSR in the organization, the employees feel more comfortable about asking for help. We have found that some faculty members feel uncomfortable about calling a student- staffed University-wide "computer hot-line," but feel comfortable talking to a peer in their own organization. In addition, CSRs understand the faculty's scholarly interests. For example, a biology professor may want to use a database program to process information sorted by genus and species. A support person outside the college might not be able to give as good advice as someone inside the college who understands basic biology concepts. The CSRs are also aware of hardware or software that may be unique to the equipment a specific organization uses. Finally, the CSR program eliminates the problem of whom to call for help, by providing one computing support telephone number to the employee. The employee knows that he or she can call one number and that the person on the other end will help find a solution to the problem. Summary and Future Efforts Computing support at Brigham Young University has evolved, just as computing technology has evolved. Computing technology is moving toward a more networked environment and an appropriate support organization for this technology is a network of computing support personnel. The key individuals in our networked support organization are the Computing Support Representatives, who ensure that employees' computing problems are solved, and the Computing Support Coordinator, who ensures that the CSRs receive the technical support that they need. We have seen that the number of employees a CSR can effectively support can be increased when the CSRs can rely on a networked support organization. The quality of computing support can also be increased by focusing the efforts of the support team with a Computing Support Coordinator. Although we have seen a lot of initial success with our support program, there is still some improvement to be accomplished. It is important to help employees feel more comfortable in dealing with their CSRs. The first step to achieving this is for employees to know who their CSR is. In several organizations we tried putting stickers on the front of employees' computers that told them who their CSR was and their support phone number. This helped employees become familiar with the CSR, and encouraged contact when help was needed. We plan to extend this project to the entire campus, so every employee knows whom to contact when they need help. We also plan to add a new field to the employee record in the University personnel database to specify who the CSR is for every employee. When the CSR program was first started, the organizations were happy to have some way of establishing computer support both within the organization and campus-wide. Deans were happy to assign the duties of the CSR to someone in the organization, and the CSR was willing to take on the extra burden of support knowing that now there was campus-level help available. This initial euphoria has worn off as networked computing support has become more common, and CSRs are now looking for an additional incentive to take on the extra responsibility. One of our recent goals has been to write a detailed job description of CSR responsibilities and use it to approve a salary increase for those who hold the CSR position. The results of those efforts, as well as more information to help other institutions set up the CSR program, are scheduled to be reported at CAUSE92 and will appear in the conference proceedings. ************************************************************************