This document was contributed by Tom Moberg to the EDUCAUSE Information Resources Library. It is the intellectual property of the contributor. Permission to copy or disseminate all or part of this material is granted for non-commercial purposes and provided that the ownership and source are acknowledged. To republish in part or whole requires written permission from the contributor. See http://www.educause.edu/copyright.html for additional copyright information.
PLANNING A CAMPUS NETWORK: A SMALL COLLEGE PERSPECTIVE
Tom Moberg
CONTENTS
SECTION 1: Introduction
SECTION 2: Observations About the Network Planning Process
SECTION 3: Strategic Issues Related to Network Planning
SECTION 4: Planning the Ideal Campus Network
SECTION 5: Building the Ideal Campus Network
SECTION 6: Funding and Staffing Issues Related to Campus Networking
SECTION 7: Planning for New Activities and Unexpected Consequences
SECTION 8: Conluding Thoughts
SIDEBAR: Information About Mediated Roundtable
Campus networking is rightly perceived by most colleges and universities to be one of the most important issues facing them at present. A campus network is far more than just a physical infrastructure, and planning and building a campus network really means developing an entirely new information environment which will have a profound impact on almost every aspect of campus life. In building networks, campuses are engaged in a change process which is strategically essential to their very nature and existence. This paper offers some advice to administrators and faculty about things to consider as they engage in planning, building, and managing campus networks.
This basic ideas in this paper evolved from a series of talks given by the author at the 1992, 1993, and 1994 CIC/CAUSE/EDUCOM/C-CUE "Workshop on Information Technologies" conferences in Pittsburgh, PA, on topics related to planning and developing campus networks at small colleges. The CAUSE staff proposed using this material as the outline to guide an Internet "mediated electronic roundtable" discussion which would glean focused opinions about small college networking from people with a wide range of experience in that endeavor.
The roundtable discussion was carried out on a special listserve established and managed by the CAUSE staff. The discussion was guided by mediators and proceeded by rounds, with each round focused on a topic in a structured outline. The mediators were responsible for the general outline of the discussion, as well as the content and length of each round. Roundtable participants were expected to adhere to basic courtesies of polite electronic discourse. At the end of most rounds, a mediator summarized the discussion for the group. The original idea was to have each round last just a few weeks, but that rarely worked. After several years of discussions, the questions and responses from the roundtable were collected and incorporated with the material from the CIC/CAUSE/EDUCOM talks into this paper.
Some members of the roundtable were chosen after they responded to a CAUSE announcement, while others were invited to join because of their perspectives and experience with small college networking. The people who participated in the mediated roundtable discussion at one time or another, and the institutions where they worked at the time, were:
John Balling; Dickinson College
Paul Bishop; Washington College
Bill Doemel; Wabash College
Morris Galloway; Presbyterian College
Leo Geoffrion; Skidmore College
Carl Heideman; Hope College
Bob Hodge; Taylor University
John Langeland; Trinity College
Ron McCleary; Rockhurst College
Ed Meyers; Macalester College
Jim Nash; Albion College
Juliet Oeffinger; North Central College
Ken Pecka; Whitworth College
Ray Phillips; Colby College
Dave Rotman; Cedarville College
Gary Schlickeiser; Reed College
Dave Smallen; Hamilton College
Tom Warger; Bryn Mawr College
Tom Moberg; Kenyon College -- Discussion mediator
Al Essa; Austin College -- Discussion mediator and participant
Michael Zastrocky; CAUSE Vice President -- Staff liaison
Julie Rudy; CAUSE Director of Publications -- Staff liaison
The paper is written in a fairly informal style, following as it does from conference presentations and Internet discussions by a group of networking practitioners. The topics in this paper are ones which seem important to the author. Obviously, there is no single correct perspective on network planning. Much, perhaps most, of this material pertains to small colleges and universities since those are the kinds of institutions in which the author and roundtable participants learned their lessons about campus networking. It is also fair to note that some of the material gathered two or three years ago may be a bit out of date, considering the pace of technological change.
The author is most appreciative of the help provided by Julie Rudy, Michael Zastrocky, and Al Essa in managing the mediated roundtable, and extremely grateful to all the participants in the mediated roundtable discussion for sharing their ideas, opinions, hopes, dreams, and vast knowledge about campus networking. The CAUSE staff, particularly Karen McBride, deserve a great deal of credit for their peerless work in editing and producing this paper.
SECTION 2: OBSERVATIONS ABOUT THE NETWORK PLANNING PROCESS
Campus planning may be viewed by some people as an oxymoron. Nevertheless, it is impossible to build an effective campus network without a significant amount of planning.
Before addressing some of the specifics of planning campus networks, it is worthwhile to think about the planning activity itself.
�Planning is a process, not a product
Some colleges use formal methodologies for planning, while others do planning "on the fly." Whatever the procedure used on your campus, it is important to stress that planning is a process, not a product. This is one of those "trite but true" statements about planning. It may be especially true (or especially trite) when applied to network planning. But a network is an organic entity which will continually evolve, grow, and mutate. Thus colleges will need appropriate mechanisms (ie., organizational structures, committees, etc.) for measuring growth and handling change in the network.
�Network planning must be mission driven
The network planning activity must be connected to the strategic goals and directions of the institution. What is the primary mission of the institution, and how does this affect networking? For example, if residentiality is a key part of the college's mission, how does networking relate to that? As colleges develop strategic plans for the institution, or master plans for facilities, it is very important to include the campus network at appropriate places in the overall planning.
�Network planning is planning for the new ICE Age, where Information, Communication, and Entertainment will converge.
It is important to understand that network planning really means planning a whole new environment which will encompass information, communication, and entertainment on the campus. The network will provide a powerful, dynamic, new channel of access to the world of information. It will also enhance and fundamentally alter the communication mechanisms on campus. And whether we want it or not, the network will provide a wide range of new entertainment opportunities.
This new environment will encompass or touch all the information technology areas (e.g., computing, library automation, audio, video and multimedia technologies). This will challenge, compete with, and sometimes replace existing modes of campus communications, information access, administrative work, power hierarchies, budgeting systems, etc.
�Network planners are change agents
People who plan and build networks are change agents. People who cause change are often suspect, sometimes barely tolerated, occasionally feared, and rarely praised. Network developers should be prepared for some resistance and some enthusiasm, from strange quarters. Don't be surprised by reactions of any sort, from both your friends and your enemies. Some people will initially support networking until they see that real change is going to happen, that hierarchies will crumble, that power bases will shift and administrative lines blur. Depending on your personality, you may even relish this role.
� Observation and reflection are crucial
Change has always been a characteristic of technology. Now, the rate of change is increasing. New technologies and combinations of technologies seem to erupt in front of us almost constantly. It is difficult to even know about the technology, much less understand the implications of it for our campuses.
Our job as campus network builders is to be both scouts and wagon masters of the Information Age. Thus it is important to find ways to step back, reflect, and observe the changes taking place. We need this to get a perspective, perceive subtle directions and trends, and respond to new imperatives before they become crises.
DISCUSSION SECTION
The roundtable participants were given a set of questions related to the five points about the planning process to see how the points correspond (or don't) to the planning processes on the participants' campuses.
*********************************************************************************************************
Questions: What methodology, if any, do you use for network planning? Do you use surveys? If so, for what? How frequently?
*********************************************************************************************************
David Smallen:
My observation is that most institutions do not have a well understood planning process (regardless of whether it is formal or not), especially when it comes to relating program to resource allocation. The result is often that planning becomes "wishing". In a small institution, with a shortage of staff, it is easy to let "measuring" things fall by the wayside. I now realize how important this is and have begun to count everything that we can to track growth.
*********************************************************************************************************
John Langeland:
In a small institution, much of planning task is driven by fairly obvious factors -- greater number of machines in locations (e.g. dormitories), new software likely to make a significant impact on network capacity (e.g., WWW), changes in available hardware (the vast majority of this year's students have Ethernet hardware -- not Appletalk on their machines), new instructional initiatives, and even in-passing comments from faculty about how much they have come to rely on e-mail as a mechanism for increasing discussion with students in their courses. Each of these factors (and many others) are part of the background for network planning, and some form of measurement is always useful in thinking about what aspects of the network require additional resources.
*********************************************************************************************************
Bill Doemel:
At this liberal arts college, if there were a metric to measure the faculty's perception of technology that would be the metric to use. Our faculty has been the driving and impeding force that impacts the introduction of technology onto the campus. Perception is more important than actual use either personally or in class. For if a faculty member perceives that computers are important for his or her teaching or research then he or she will begin to play a role in the process that leads to change. For example, at most institutions computers and then networks appeared in the science departments. At Wabash, the primary pressure to advance our technology came from the English Department. They decided that computers were important to the teaching of English and that this could be best accomplished if there were a LAN so at our institution, the first network linked the English Department and a classroom of computers to a server in the computer center. Since then, we have been following a plan for expansion and innovation which, being governed by a Committee of faculty and staff, responds to changes in those faculty perceptions. While we have maintained internal statistics on use of bandwidth, hard drive space, etc., the decision makers have been more impressed and responsive to testimonials by faculty members who share their perceptions about the importance of the computer to their work.
*********************************************************************************************************
Questions: What is the role of advisory committees at your institution? Who participates in these committees? How often do the committees meet?
*********************************************************************************************************
David Smallen:
We do use an advisory committee (4 faculty, 3 administrators, two students) which meet several times a semester. However, from experience, if people do not see a direct relationship between advising and resource allocation they will not be willing to devote time to the effort.
*********************************************************************************************************
John Langeland:
The faculty advisory committee is primarily concerned with instructional uses of computing technology; they've come to assume that whatever is required in a networking sense will "simply happen." While I take this as evidence that we must be doing at least a passable job of network planning, I believe that faculty and administrative staff may fail to consider opportunities that enhanced networking make possible because the network is "out of sight, out of mind."
As technology has become a part of what nearly everyone "does" in their daily routines, I've increasingly found it useful to impose the role of technology planning within other on-going planning contexts. There was a time (I think) when technology planning could serve as an organizing and planning opportunity for other aspects of the institution's life. I've increasingly come to believe that the best way for an institution to exploit technological resources is to ensure that all planning venues within the institution have at least some resource available to encourage a consideration of the role technology can play.
This means for us that the only planning committee for networking is within the staff of the computing center; our sense of direction comes from the Computing Center's involvement throughout the life of the institution.
*********************************************************************************************************
Morris Galloway Jr.:
We have a campus networking committee, composed of faculty from several disciplines (currently business, religion, chemistry, computer science), the library, and computer services. It "meets" via e-mail several times a year. A face to face meeting is held once a year, or if anyone requests a meeting. A technical working group handles bits and bytes. We meet in the corridors or over lunch. This group is all CS types.
*********************************************************************************************************
Bill Doemel:
Terribly important here. We have several Committees, the Computer Advisory Committee, being the central Committee. Composed of 4 faculty, 3 staff (administration) and the Director of Computer Services with support people being present--library and Computer Services, this group meets regularly to prepare a budget proposal in the Fall. It also meets monthly/bi-monthly to consider issues that are emerging, like a new campus network. The primary role of this group is to prioritize the needs of the campus. Departments submit their requests, the Committee interacts with the Departments to be certain that all of the initiatives and needs are understood and then the Committee establishes priorities. There is also an Administrative Computing Committee that focuses on Administrative implementation of Datatel, our Administrative system. This Committee has one faculty member and representatives from all of the offices. Finally, we had a temporary telecommunications Committee with representatives from faculty, administration, secretaries and students to develop an RFP, evaluate the proposals, and recommend action to the Administration.
*********************************************************************************************************
Robert Hodge:
Our planning process for anything always follows the same hierarchical model - program, people, buildings and equipment, then finances. A network plan needs to be responsive rather than proactive. I agree that network planning needs to be mission driven. I do note that mission statements are usually broad statements of direction under which virtually everything can be covered and little can be evaluated. To the extent that the mission is defined in more detail through institution purposes or objectives, I believe that network planning needs that intermediate linkage.
The two statements of "responding to mission/purposes" yet being "a change agent" produce some tensions within the organization. Waiting for something to respond to is not workable in this environment. Yet, trying to change the organization which has not decided that it wants to change brings forth images of the computer/network czars that tend to be assassinated. This balancing act defines the role of many CIOs.
Placement and support for all of this is crucial. If we are truly speaking of making a strategic difference using IT, it is important that senior level support be present and visible. Indeed, I was once told that the value of information and IT will only be as great as it is positioned in the university. Hence, our CIO/VP position that reports to the president's office.
How to bring consensus and broad based thinking to this effort? Maybe a little case study here may be of interest. For the last twelve years, we have had an Information Steering Committee which was a faculty committee with responsibilities for oversight and policy making. The senior administration did not choose to participate because they had "their man" on the committee. The committee time was used primarily to bring everyone up to speed on what had happened last month with little time to discuss the future.
Several years ago, we established the CIO position which eventually became a VP position. Last year, the faculty, tired of a lot of committee work, did away with many committees and consolidated others, leaving all the duties of the steering committee to the VP. Quickly, I established an MIS advisory group of those faculty and administrators who probably could do my job in a pinch; a group of people who have been both faculty and DP managers and in whom I could take all kinds of issues and get instant wisdom and who will hold me accountable. At the same time, we established an academic high tech planning group which is all faculty except me. Another great group of movers and shakers, good listeners, of high esteem to the faculty. The next step is to combine the two groups. They are willing and desirous of that; it will be all faculty except for one administrator and myself. So, still a faculty group but not part of the formal faculty governance system.
*********************************************************************************************************
Questions: How do you handle the entertainment component of network services? In particular, do you have guidelines about students playing games over the network and the Internet? What about pornography? Should academic institutions handle privacy issues differently than corporations?
**********************************************************************************************************
David Smallen:
We are in the process of developing policies for appropriate use of the network. The only advice I can offer is that we must try to make network policies consistent with other college policies. The more that we can treat network issues in the same manner as other issues, rather than something "new" the better off we will be. One thing I don't want is to get into monitoring people's choice of information.
*********************************************************************************************************
John Langeland:
I have found it very useful to operate along the lines of a common carrier -- like the telephone company. "The Computing Center doesn't provide content" is one of the frequent refrains around campus -- and we don't "do" censorship either. Games, instances of sexual or other harassment, and pornography are generally not issues of concern. When any of these activities interfere with the use of instructional resources, they are referred to the appropriate authority (generally the Dean of Students -- but when the game playing or Internet abuse is by staff to the appropriate supervisor).
*********************************************************************************************************
Morris Galloway Jr.
We use the "Be polite" approach. Students are asked to limit their use of public terminals to 30 minutes when others are waiting, and to yield to others with academic work to do. Student lab managers may enforce this policy. In addition, talk and similar services are restricted during the day.
We do not carry pornographic news groups. Of course, there are ways to get such material, but we do not officially host it by carrying the news groups.
Our acceptable use policy states that we inspect and remove files only as needed to maintain the smooth operation of the system. In practice, we've never had the need to inspect, but we have removed some disk hogs.
**********************************************************************************************************
Bill Doemel:
We inform students of our priorities--class work, research, personal education, communication and games. We expect the students to regulate themselves. Those that become hogs, are advised of their behavior and usually respond by allowing other users access to the system. We are not a corporation in that our users are not being employed to do work for some particular good--they are here to learn, to explore and to mature. We do not regulate their exploration, only request that they act within the guidelines of a civilized society. Indeed at Wabash, we have but one rule, always act as a gentleman (we are a school for men). The importance of understanding the meaning of that rule leads to many discussions in my office concerning the consequences of actions that one chooses to take. Some of those discussions lead to the Dean's office, most lead to a better understanding about living in a diverse society.
**********************************************************************************************************
Question 4: How is IT perceived at your institution? Does senior staff support your endeavors? What about faculty? As a change agent, what is the single most important message about IT you are trying to communicate to the campus community? What is the area of strongest resistance?
*********************************************************************************************************
David Smallen:
There are lots of ways to accomplish change, some more painful than others. I believe in the advice given by a college president that often the best way to accomplish change is to "navigate corridors of indifference". You have to know where you are going - you can't wander aimlessly, but you also don't accomplish much by running into walls. I feel the most important thing is to avoid raising expectations much beyond what is reasonable to deliver. Easier said than done when you live in the world of the possible!
*********************************************************************************************************
John Langeland:
Again, there was a time when I thought this was true. I now believe that while we still are essential and central to much of what is being re-invented within out institution, the days in which we somehow "caused" the change have passed. Instead, I think the new role is more one of enabling others to make changes. Clearly, without robust networking we can't see networked information resources as a reliable component for instruction. But creating a robust network doesn't make change happen -- faculty and students need to engage these resources and make them a part of their daily routine. The true change agents are those who attempt to use these new resources and technologies.
There was a time not very long ago in which the computing power of the CPU or other hardware actually limited the ability of people to complete an analysis or even format a wordprocessing document. As CPU power on faculty, staff, and student desktops has grown beyond that of mainframes of only a few years ago few are constrained by the hardware or software technology. As costs of technology have declined, most workstation technology in use within higher education is significantly less than five years old; as such the capabilities of the hardware and software that it uses are significantly greater than the demands placed upon it by its users.
So too with networks. With some effort and some significant resources, it isn't too hard to guarantee that network capacity is more than adequate for the needs of the community. In the same way that computing power grows and costs decline, so too have the costs of networking. And the promise of future technologies and greatly expanded bandwidth at ever decreasing costs are at least as secure as those of further declines in costs for CPU power.
In a robust computing and networking environment we can still encourage change by illustrating the kinds of uses that can be made of the new technology -- and in that sense we still function as change agents. But increasingly I suspect our own computing and networking organizations sometime limit change by not understanding that the more critical factors for promoting change are the strategic needs of the institutions we serve rather than the capabilities of the technology per se. That means that we must find ways to make others the champions of technological change within the institution.
*********************************************************************************************************
Morris Galloway Jr.
IT is perceived relatively positively. The Dean of Administrative Services is an officer of the college, so IT is represented in the president's cabinet. The message I am trying to get out is "Appropriate involvement across the college." For example, Web searching techniques should be the bailiwick of the library. Why should it matter whether the information is in Dialog, on the Web, or in a printed index? Again, editorial expertise is usually found in the Public Relations office. They should take the lead in Campus-Wide Information Services and, of course, P/R use of the Web.
**********************************************************************************************************
Bill Doemel:
Positively--because we are responding to others and trying to help them to achieve their goals. The Administration has been very supportive of the interests of the faculty, hence our budget has increased substantially while other budgets are shrinking or remaining the same. It is easier to describe what we are not communicating: we make it very clear to everyone that computers do not save resources. People who use computers print more documents, expect more computing power, and expect more support than people who don't use computers. Computers enhance and change the lives of users but they don't reduce the costs of teaching or the costs of administration. Those savings are often lost because we can find more things to do with computers than we could do without computers.
**********************************************************************************************************
Robert Hodge:
"Network planners as change agents" - I believe that applies to any IT person. In our situation, we moved from traditional DP to IT, and along the way the president suggested that since we had a lot of data, maybe institutional research should be a part of that. And, maybe my office should help determine what all the information meant. So, my office assumed responsibility, (from an information resource and systems mentality perspective) to coordinate long range planning for the university, much which increasingly has a high tech/information systems component. Good plans contain an assessment component, so the university assessment process is now coordinated by this office as well, working with academic assessment as a related but separate part.
I wonder why IT/network planners are change agents. Certainly the technology is a significant part of that. But, I believe that possibly more importantly, it is the systems thinking that is more necessary as the university processes are reengineered. IT may be a secondary part to the change process; it is hard to separate the two anymore.
*********************************************************************************************************
Questions 5: How do you make the time for observation and reflection? Do you plan activities for your staff (e.g. staff retreats) to gain perspective?
**********************************************************************************************************
Dave Smallen:
I can attest to the value of getting away. We try to do that once a month (with everyone on the staff) if only for a few hours, and afternoon, or a whole day. We do the longer ones during times when school is not in session - not because we don't have things to do, but that's the best time. Getting away is very hard to do if only because when we have so much to do the natural tendency is to keep plugging away. When that happens I have to keep reminding myself (and others) that if we were sick on a particular day, the world would go on without us. If that's the case, why not plan a day away and do something useful, rather than wait until you have to spend it in bed! Finally, I'd say that the process of planning is most valuable for those who do it, not as much for those who observe the results.
********************************************************************************************************
John Langeland:
Only in the most unstructured of ways. Late Friday afternoon one will likely find relaxed conversations over coffee or other libations where planning and assessment are at least one part of what's going on. We've never had a staff retreat, but we do try to get the entire staff together over lunch four or five times a year to suggest directions and encourage questions.
**********************************************************************************************************
Morris Galloway Jr.:
I work at home some mornings. I have a terminal there. My senior associate does the same. People are encouraged to go to appropriate conferences and visit other schools.
********************************************************************************************************
Bill Doemel:
We have had little time to reflect as a group but I have been very supportive of training for staff and workshops. Since we have to spend the weekend to save air fare and since usually 2 or more go these serve as mini-retreats and sources of new perspectives for everyone. Every member of the staff is usually gone for at least 3 days to one of these functions each year. We need to find other times to work together, but you have to remember that we are small--8 and so we do spend substantial time talking about where we want to be and where we are.
SECTION 3: STRATEGIC ISSUES RELATED TO NETWORK PLANNING
It seems to be characteristic of higher education today that many senior administrators are ambivalent about information technology: they aren't sure whether information technology is contributing to the problems of higher education, or whether information technology can be part of the solution to those problems. This is much less true today than it was a decade ago, but it still seems fairly typical that the people pushing networking have to convince senior administrators of the value of a network as part of the planning process. I have found it useful to explicitly describe ways that a campus network is fundamentally important to a college. This may not seem like a pragmatic topic, but it is essential for the whole campus to understand why a campus network is important.
So why is the campus network a strategic asset for a small college? What does this mean? Is this just jargon? To me, this means that the campus network is central to the basic purpose of education and can make the institution better and stronger. For most colleges, the central mission is to provide a high-quality teaching and learning environment for undergraduates. Beyond that, colleges need to be well-run, efficient, dynamic institutions. Here then are some general ways that I believe a campus network can improve and strengthen an institution.
•Full access to information
The existence of a campus network enhances teaching, learning, and research by allowing faculty, students, and staff to have participatory access to information and technological resources on the campus and in the community, region, nation, and world. The network eliminates physical isolation, and allows students at small colleges to have the same kind of information access that any student anywhere has. Without this access, institutions of higher education could become road kill along the information highway. An institution which does not participate in the Information Age will simply not be viable in the 21st century.
A national telephone survey conducted by the Chronicle of Higher Education asked 1000 adults about their preferences for services on a national information network. Of the people responding, 59 percent gave the highest desirability rating to the concept of interactive educational programs which would allow people to attend classes at home, while 70 percent said they would back government support for interactive, at-home educational programs for children and adults. Imagine the impact on traditional residential colleges when truly effective educational opportunities are available in the home.
•Enhanced communication
A network enhances the campus atmosphere by improving communication among faculty, staff, and students. For example, faculty and students can have closer contact and better communication, even on small campuses which pride themselves on this sort of educational interaction. Multi-disciplinary courses can be planned, facilitated, and even partially taught by faculty-to-faculty and faculty-to-student electronic mail conversations. Faculty and staff can more easily contact each other, without the typical situation of telephone tag. Student groups can arrange meetings, faculty can distribute assignments to students, students can turn in assignments, library books can be ordered from inter-library loan - just a few of the thousands of ways that campus activities are facilitated by communication over a campus network.
•Support for student services
A network supports student services by allowing on-line registration, requests for transcript information, scheduling appointments and meetings, and submission of electronic forms for almost anything students formerly used paper forms to do. Through Web pages and Internet forms, colleges can distribute information about the college which is useful in recruiting new students. Faculty and staff at colleges with off-campus programs use the Internet to maintain close links with the students studying away from the campus, thereby helping the students carry out activities such as course registration and communication with advisers and friends.
•Administrative efficiency
A network promotes administrative efficiency by providing access to institutional data, allowing better decision making, improving productivity, and facilitating reengineering of operational processes. The use of campus-wide databases for student and employee records, scheduling, procurement, and other business processes leads to more efficient work, flattens hierarchies, and removes departmental barriers.
•Integrates information technologies
The network integrates various information technology areas by providing a central focus for management of resources and services. The network makes it easy (and sometimes imperative) to coordinate the management of computing, library automation, telecommunications, media technology, and other networked information resources. Such coordinated management allows better use of resources, helps eliminate turf issues, clarifies confusion about service sources, and allows much more efficient strategies for handling things like cable systems.
•Supports institutional advancement
A network facilitates fund raising and good relationships with external constituencies (e.g., alumni, parents, Trustees, business partners) by providing new options for communicating with diverse and dispersed groups through electronic mail and World-Wide Web technologies. Such linkages allow the college to build and strengthen these important ties and also provide a useful service to the college "family."
DISCUSSION SECTION
The roundtable participants discussed a number of issues related to campus strategic planning.
**********************************************************************************************************
Questions: Does your institution have a strategic plan document? If so, please describe briefly its scope and purpose. Has it been endorsed by the senior administration? Who participated in its formulation? How frequently is the strategic plan revised? Is residential networking part of your strategic plan? Is Information Technology recognized explicitly in the overall strategic plan of your institution? Does your institution have a master plan for the campus network?
**********************************************************************************************************
David Rotman:
1. Does your institution have a strategic plan document? If so, please describe briefly its scope and purpose. Has it been endorsed by the senior administration? Who participated in its formulation?
Our college annually produces a campus wide strategic plan. Technology has been a high priority (one of 5 or 6) for the last four years. This strategic plan is influenced by the work of a technology task team. The team met in 1990-91 to develop a 3-year plan. The task team is meeting again to see where we go from here. These plans are reviewed by a college planning committee (mid-level managers and faculty), and then approved by the administrative council (vice-presidents) and the board of trustees.
2. How frequently is the strategic plan revised? Is "residential networking" part of your strategic plan?
We revise the plan annually. Residential networking is essential to our plan...we provide a college-owned computer in each dorm room.
3. Is Information Technology recognized explicitly in the overall strategic plan of your institution?
Yes, one of the top priorities.
4. Does your institution have a master plan for the campus network.
This would mostly be a technical document establishing network standards and priorities for new installs and upgrades.
Each change is planned according to prevailing standards (e.g., category 5 cable), but there is no formal document outlining all of these requirements. We do, though, place such requirements in RFP specifications.
**********************************************************************************************************
Bill Doemel:
1. Does your institution have a strategic plan document? If so, please describe briefly its scope and purpose. Has it been endorsed by the senior administration? Who participated in its formulation?
Yes. We are working on a series of projections that started as 5 year projections and gradually shrank to three year projections about where we want to be. These projections have formed the basis of our budget requests for each year and are supported by the administration. The plans are developed by CS in cooperation with the Computer Advisory Committee--3 faculty, 3 staff and the Director of CS. This plan is integrated in the campus wide strategic plan which is being developed by all Departments of the campus. This process is in its infancy so it is difficult to describe. It involves developing a mission statement, evaluative tools and proposal to address the deficiencies.
2. How frequently is the strategic plan revised? Is residential networking part of your strategic plan?
Our plan is revised yearly and as I said has shifted from 5 years to 3 years. We are currently implementing the residential computing part of the plan.
3. Is Information Technology recognized explicitly in the overall strategic plan of your institution?
Yes.
4. Does your institution have a master plan for the campus network.
Yes. We developed a plan about three years ago for the campus network--inverted star topology, fiber links between buildings, 10-base-t within (level 5 installation). This plan has been implemented gradually until this year. Now, we are extending the network to all campus buildings.
**********************************************************************************************************
David Smallen:
1. Does your institution have a strategic plan document? If so, please describe briefly its scope and purpose.
Our strategic plan "for IT" is in the final draft stage.
Who participated in its formulation?
Committee of 4 faculty, two students, 5 staff/administration. First draft sent out to entire campus via gopher server.
2. How frequently is the strategic plan revised? Is "residential networking" part of your strategic plan?
The plan will be revised annually. Residential networking is an important part of the plan
3. Is Information Technology recognized explicitly in the overall strategic plan of your institution?
The institutional plan is in the process of formulation but will involve IT
4. Does your institution have a master plan for the campus network.
This would mostly be a technical document establishing network standards and priorities or new installs and upgrades. We are in the process of completing the entire network now (all buildings and spaces). The network conforms to a standards document.
**********************************************************************************************************
Robert Hodge:
1. Does your institution have a strategic plan document?
Taylor University has an institutional strategic and a long range plan. We define strategic as what we can do within the resources we have. Long range is that which will take resources outside of our present control e.g.. fund raising, grants, etc.
That plan is a consensus plan in that every constituent has several opportunities for input and discussion. A long range planning committee of the institution works on it for almost two years. There is also a board committee that works on it too. Consensus takes a long time but in the long run, you are all going in the same direction.
We have in existence a set of operating guidelines for computing and telecommunications. We took three years to put together a broadcast/ video services long range plan. Presently, we are completing a communications long range plan in response to the broader institutional plan. It covers voice, data and video, much of which will take place using the same switches and lines between campuses.
As part of this iterative process, there is a major high tech component in the institutional long range plan. It calls for an endowment which would cover the replacement of all existing equipment plus a large one time injection to add some significant amounts throughout the university. Those amounts together are almost as large as our institutional capital campaign that was just completed.
A sidebar about this is the concern from other areas about this approach. That is, it is a good approach but some are concerned that it may take money away from other types of equipment such as pianos or microscopes.
2. How frequently is the strategic plan revised?
The institutional plan takes about two years to get through with a year of rest in the middle. The high tech plans are ongoing and are completed as we look towards the coming together of new programs, the declining lifetime of significant portions of the high tech systems, etc. For instance, this year we have a significant interest in interactive video between campuses, and we have a phone switch that just reached its expected depreciation lifetime. A single switch will handle both needs much easier than two different switches and networks. And, they will use the same lines between campuses.
Residential networking is on the wish list portion of the plan as it is evolving. Presently, both faculty and students (they participate in planning too) see computers in the residence hall as inevitable but not for several years. The discussion is about whether they will be required or encouraged, and by major or the institution. Today, most students say they have access to a computer when needed in the hall; that means in the room, across the hall or on either side. There is an incredible amount of sharing done. My informal survey suggests that given an option of $125 more a semester as a high tech fee for more lab machines, or $250 a semester to lease to own their own machine, almost all would take their own machine. And, many have asked if they could sign up for that now. Their main reason, by the way, is to have access to equipment on Sunday when our library and labs are all closed.
3. Is Information Technology recognized explicitly in the overall strategic plan of your institution?
Yes, as mentioned above. The plan contains wording that high tech presents some of the best tools to improve education but that we are least prepared to fully utilize it.
It may be helpful to know that every element of the long range plan starts with a program. Everything is program driven, and every program must relate to an institutional purpose or the mission statement. We define the program, then the people needed to make the program work, then equipment and facilities to support the people to make the program work, and finally the finances to purchase the equipment, build the facilities, and pay the people to make the program work.
4. Is there a master plan for the campus network?
Mostly a technical document establishing standards and priorities.
I think those are two different things. We have a wonderful set of standards that have proven to be highly effective and efficient, for the IT people, the end users, planning for new buildings and remodels, and outside contractors. Our president hates the disfiguration of trenches on the campus. With that, we established a set of standards which would make trenching a one time thing, or at least work as long as the president is here and beyond.
We basically go with BICSI standards. We have often purchased the $200 set of manuals and simply give them to architects and require that whatever they do meet those standards. We have our own certified BICSI person on staff. It works very well.
Now, priorities for new things are different. Academic items are now prioritized through the dean's office, something we worked towards for ten years. MIS items are jointly agreed upon by users and the IS department. That works very well because it is a good team of people with a perspective broader than each department.
Major initiatives come from my office as a leadership role that I have. The president still introduces me as the czar and I always chafe at that. It is wrong positioning. Anyway, I prefer to work with ad hoc groups on major initiatives, and ultimately the budget committee and president do not approve any such projects without advice or recommendation from my area. And, most funding comes to this area to be administered, due to a good track record of coming in within budget and on time while meeting the needs of the users as they have defined them, and still being able to say "no" when the target moves too far off the original target.
Overall, I am pleased with the evolution of the administration of high tech, especially with the new leadership provided by the assistant dean. That one aspect of leadership has done a lot to improve the academic use of high tech by leading the faculty to priority decisions.
I continue to believe that progress, at least here, is based more on the quality of working relationships than the breadth or depth of a plan. That is to say, the printed plan should be there but only to summarize the process which has already brought agreement in direction. Knowing that a printed plan is required brings accountability to the planning process, but ideally, I think it should be somewhat anticlimactic within the university. Externally, it is the document that is used for fund raising and proposal writing. Many have been amazed at the effectiveness of such planning, and the document, to inherently reallocate resources voluntarily, maintain direction when wild ideas tend to distract us, yet provide flexibility to change as technologies and programs change.
**********************************************************************************************************
Morris Galloway Jr.:
1. Does your institution have a strategic plan document? If so, please describe briefly its scope and purpose. Has it been endorsed by the senior administration? Who participated in its formulation?
A long range plan is prepared every 5 or 6 years, and reviewed halfway through it's life. It is used as the basis for annual administrative objectives. It has been reviewed by almost every constituency on campus, including the officers. A committee of faculty, staff and students prepared it.
We also link the 10 year self-studies to these plans, referring back to the last self study.
2. How frequently is the strategic plan revised? Is "residential networking" part of your strategic plan?
Residential networking is being considered in the current self-study, and will probably make it into the next long range plan.
3. Is Information Technology recognized explicitly in the overall strategic plan of your institution?
Yes.
4. Does your institution have a master plan for the campus network.
No, but we are doing some preliminary work on one now.
SECTION 4: PLANNING THE IDEAL CAMPUS NETWORK
The campus network provides the essential first link in connecting scholars to the information universe. As campus planners engage in the process of planning and building campus networks, they need a clear vision of what an "ideal" campus network should be like. Since every college is different, every campus network will use a different approach to envisioning, planning, and building its unique campuses network. The campus network will have a profound impact on the way the campus functions, so it is important to have wide involvement of all stakeholders in the planning process. It is also useful for campus planners to study the network strategies and designs of other colleges. The resources of CAUSE, including the CAUSE document library and the descriptions of the recipients in the CAUSE Networking Awards provide excellent sources of information about campus networks.
Regardless of the specific planning process or the particular network design used at a campus, there are certain dimensions (or aspects or characteristics) which exist in every campus network. The following model for envisioning the "ideal" campus network, incorporating five dimensions ranging from the physical to the cultural, may be useful to think about during the planning process.
1. Physical Dimension:
The ideal campus network is an information channel which reaches every place on campus where "knowledge workers" live and work, including offices, classrooms, laboratories, studios, student residences, student activity areas, etc. It includes a physical infrastructure which consists of high-grade copper and fiber cable, junction and termination boxes, communications devices such as fiber hubs, bridges, routers, terminal servers, and wiring closets to house equipment and termination panels. The ideal campus network provides a seamless interface to on-campus sub-nets and to off-campus locations and resources, such as faculty homes, metropolitan and regional networks, and the Internet. The ideal campus network has physical components which meet defined institutional standards, provide for modularity and expandability, and are well documented and mapped.
2. Protocol Dimension:
The ideal campus network handles multiple network protocols, such as TCP/IP, LAT, AppleTalk, Netware, etc. Therefore it should not have highly proprietary characteristics which preclude use of other protocols. The ideal network provides a seamless interface between protocols used on sub-nets and meets well-defined institutional standards for network connections and protocols.
3. Management Dimension:
On the ideal campus network, network activities and protocols are invisible to users. Network growth, while constant, is managed without disruption to users. The ideal network management structure includes appropriate staffing, budgeting, control, and security to provide for orderly and planned change. The institution with an ideal campus network has a funding program which covers continual growth of the network and replacement of functionally obsolete equipment.
4. Application Dimension:
The ideal campus network provides easy access from any access point to all information pools, including library materials, departmental libraries, non-print media collections, institutional databases, etc. It provides a variety of integrated information resources via a campus wide information system (CWIS). The ideal campus network incorporates a seamless electronic mail with a common user interface to all members of the institutional family, which may include off-campus constituencies, and provides easy sharing of electronic resources (data, text, images, sound, video) across the network. All members of the campus community find it easy to use shareable computing hardware and software resources such printers, plotters, statistical packages, programming languages, and databases, everyone has full access to Internet applications and information resources, and all applications are well documented and publicized.
5. Cultural Dimension:
Ideally, all campus constituencies use the network fluently as a natural and integral part of their communications and information exchange activities. On a campus with an ideal network, students use the network as an intrinsic part of their campus life, faculty actively seek to use the network in new and creative ways to enhance teaching and learning, and administrators and staff routinely use the network to improve operations and reengineer archaic administrative systems. The network provides a unifying concept for campus-wide integration of information technologies, resources, and services. The institution considers the network a strategic asset, and is committed to supporting the network as a vital strategic resource.
DISCUSSION SECTION
Most of the roundtable participants are the chief architects of the network developments under way at their campus, so their sense of the ideal is likely to be the vision which their college is trying to achieve. The development and management details which emerge at their institutions will be manifestations or realizations of that vision. The participants were asked to share their notions of the ideal campus network to see if there is a consistent vision for small colleges.
**********************************************************************************************************
Question: What is your vision of the ideal physical infrastructure for a campus network?
**********************************************************************************************************
Jim Nash:
I want to state a philosophical position before I answer the question. I believe the goal needs to be the inter-connection of all computing on campus, and then connection of the campus network to the Internet. Shared computers (servers) should be accessible to all campus clients. Having said that, I think the infrastructure can take a number of forms which accomplish the goal. Our choice has been to use fiber cable between buildings and copper within buildings. A star topology with hubs, we hope, will allow our one staff person to manage the network, and keep it running. So far it has worked. The fiber will allow us to increase from 10 MB to 100 MB on the backbone in the future. We will probably need that kind of speed in the buildings eventually also. We have switched our copper wiring standard to level 5 in new construction. Unfortunately we did not plan for faster speeds in the past, and may need to rewire some buildings.
I do not think that you can anticipate too far into the future what the infra- structure needs to be. We added a lot of coax to buildings years ago, which will never be used. So my answer to the question at this time is fiber between buildings, copper within buildings, management hubs, and good network management software.
**********************************************************************************************************
Bob Hodge:
1. We must have flexibility for change and growth with a minimum of difficulty and time in the future. That means that there is enough of our infrastructure in place to move quickly. That also means that the infrastructure is an institutional cost which is borne without any particular end user champion. Many end users only know that they want a network connection when they know they want a network connection. In physical terms, that means a lot of empty conduit, a few in-ground hubs in strategic places, reserved equipment closets with our name on them in various buildings, etc.
2. Standards - an ideal infrastructure will be based on common standards that we and architects agree upon. That has been difficult until recently. Too often, we talked to only electricians who had no vision or understanding of what we were trying to accomplish with data and video. Now, we use the BICSI standards and even purchase the several hundred dollar volumes for any contractor we deal with that doesn't know them or abide by them.
3. Internal strategy - for us, that strategy was premised on several statements from our president's office which has a greater mind for the look of the campus than the wire underneath it or behind walls. One, he didn't want to see another trench to a building during its lifetime. Second, he didn't want any walls to be opened for cabling once a new building was complete. With that, we provide our specifications to the contractors and the costs become building costs rather than communications costs. It works well.
In an ideal setting, we would have conduit to each office and room and possibly multiple locations, even to the point where there would be a non-power outlet next to each power-outlet in every office. There would be conduit to the roof built into each building when the roof was put on a new or reroofed building. Getting permission to penetrate a roof afterwards has proven to be difficult. Each building would have some place on the roof capable of supporting a satellite dish and maybe two. Yes, I know of direct broadcast satellites but they are not here yet and in an academic situation, where there is a satellite, someone will want to access it.
4. Reliability - in an ideal situation, we would have redundant pathways to each building similar in concept to what the power company does. We would have redundant pathways to our second campus and to Internet/Bitnet as well, similar to what we do with our long distance service.
5. Integration - in an ideal situation, we would at least integrate the copper, fiber, and coax into the same hubs, cable trays, conduit, equipment closets, etc. In the future, we will move to integrating the digital signals for video, voice, and data onto the same fiber.
**********************************************************************************************************
Dave Rotman:
Here at Cedarville, we spent over a year studying what we wanted to do with information technology...the first 3 or 4 months were heavily focused on foundational issues. The resulting report from our technology task team included the following "vision" statement which is a technological philosophy which seeks to enhance the educational experience we offer and increase its perceived value.
The Statement: Cedarville College is committed to providing its faculty, staff, and students with an integrated, broadly-accessible information, voice, and video communications technology infrastructure. To this end, the College should:
1. Assure that its personnel and students can access and maintain, through appropriate technology, the information necessary to fulfill their roles;
2. Continually assess information technologies and seek to implement appropriate hardware and software that enhances individual and organizational effectiveness;
3. Make available to personnel and students instruction and reinforcement in the use and application of information technologies;
4. Foster an environment that encourages responsible use of technology, yet maintains a sensitivity to technology's effects on its users and their environment.
On the basis of this vision statement, we have embarked on a 3-phase project to network our entire campus. The design reflects our concept of an ideal physical structure: fiber between buildings (the "physical" backbone); twisted-pair copper within the buildings (adequate for 100 Mbps); a collapsed logical backbone located in the network operations center. Building distribution is controlled by intelligent concentrator hubs. The collapsed backbone is currently a 16 Mbps token ring. All of the devices on the network can be managed via a graphical SNMP system, though we have found that a simple "telnet" to the device has been adequate so far.
The design decisions were influenced by the following:
1. Use of fiber between buildings allows for very high capacity and future utilization, while minimizing lightning problems.
2. Use of twisted pair media within the buildings provides an economical distribution system, with adequate bandwidth for individual stations.
3. The collapsed backbone provides easy network management from a single location.
**********************************************************************************************************
Morris Galloway Jr.:
Here are three suggestions.
1) Adhere to standards where they exist. We've hired a network integrator to design the physical network. They are members of BICSI, and follow the recommendations of that organization. We're also asking some colleagues at other schools and businesses to review their designs.
2) Get the best grade you can reasonably afford, but don't be the among the first to install a new wiring type or scheme. Wait until you know whether or not it will be popular. If lots of folks have a grade or type of wire installed, then new, fancy electronics to pump more over that wire will probably be available. We are installing FDDI grade multimode fiber between buildings, and category 5 copper within buildings.
3) Put in as much as you can afford while you have the ground/wall/ ceiling open. Since things seem to change so fast in the realm of networking, we are trying to allow for extra conduit space wherever we can. Our inter- building runs will consist of two 4" conduits, each prefabricated with four 1" innerducts. That gives us eight 1" pipes to each building. We're pulling 4-pair cat 5 copper. We can only afford one termination per workstation. However, we are pulling some extra copper to each wing of each building, leaving it unterminated and coiled up above the ceiling. Dropping back to category 4 would not save us enough to be worth risking a rewiring job in 5 years. On new construction, we're asking for a 4x6 foot closet on each floor.
*********************************************************************************************
Ron McCleary:
In our physical network design, we generally have been guided by five principles:
1. FLEXIBLE/CONFIGURABLE. Change is the one constant in a networking environment. We might want to move the whole computer center from one building to another. We might move a bunch offices from one place to another. When such things happen, it should be possible to simply reconfigure the network to handle the changes.
As we wire buildings, we have the twisted pair on each floor go to a wiring closet to patch panels. In only a few minutes, we can patch any pair into either our LAN or LAT system. The modular plugs at the wall can be quickly changed between 10BaseT and LAT. All the fiber into a building terminates in patch panels, and if we need a run through several buildings, we simply patch the necessary links together in the intervening buildings.
2. STANDARDS. We don't want to be tied to one or a few vendors. And, we don't want to be tied to obscure technology. The only way to have such freedom is to specify that all components meet the standards and to stay away from proprietary protocols and hardware. We have written into all purchases and RFBs that all fiber will be FDDI-compliant. (We aren't running it yet. We may leapfrog over it, but whatever is coming down the road will probably run on FDDI-compliant fiber.) All ethernet cabling and equipment must meet IEEE 802.3 standards. Fiber equipment must be FOIROL-compliant.
3. HIERARCHICAL. I assume that we are going to have some combination of fiber, coax, and twisted pair around for quite a while; maybe the coax will loose some of its popularity. Although the rates on copper may sometimes equal or exceed those on fiber, I assume those instances are anomalies and that generally fiber will carry more than copper. I assume that yet-to-be-discovered technology for fiber and copper will generally be backwards compatible to run over FDDI-compliant fibber and 10BaseT-compliant copper. With those assumptions, it makes sense to put fiber in places where the traffic is going to be greatest, say between buildings. On each floor, we will want 10BaseT copper.
As you can see, I don't like running all the twisted pair in a building to a single location. It is aesthetically unattractive to me to see all that twisted pair wiring running vertically. I like a 3-level hierarchy with fiber on top and twisted pair on bottom. I am not as clear about what to do about the middle level of the hierarchical system. We are still using thin ethernet until we can see whether fiber or twisted pair should handle the middle level. We figure that it doesn't cost us much to install, and we can and cheaply abandon it for either fiber or copper in a few years once the choice is clear to us.
4. EXCESSIVE AND UBIQUITOUS. Maybe this is just another side of the "flexible and configurable" aspect mentioned above. Regardless, we want lots of twisted pair on each floor of each building. We want fiber to every building, even if there is no need for computer communication there now. I look forward to the day when 90% or more of the request for new connections or changes to existing connection can be met without my technical staff pulling cabling. When we were planning the twisted pair for a remodeled building last year, if there was no plans to put computers in a room, we put in two sets of twisted pair. If there were plans for a computer connection, we put in even more sets. I don't think we wired the restrooms, but we probably should have! When we were planning our fiber system, we needed about 2-4 pairs to a single building to meet existing needs. We put in 12 pairs to all buildings around our quadrangle. We have already started using some of those other pairs.
5. CHEAP, DUMB, AND DISPOSABLE. Actually, these characteristics apply only to some parts of the network. I figure that we will have to live with the fiber and the copper for 10-20 years, so I am willing to put money into massive quantities of the best quality copper and fiber. The exception, as I mentioned above, is that middle level of the vertical runs between floors. We are using thin ethernet there with the plan of abandoning it for fiber or 10BaseT in a few years.
On the other hand, I figure the life of the devices on the ends of the fiber or copper will be about 2-5 years. There are two reasons for that. First, newer technology will make it impractical to repair anything over 3-5 years old; it'll make more sense to throw the old away and replace it with new. Second, we may want to abandon older protocols for newer faster ones as the demand for network capacity grows. I don't want to put a lot of money in devices unless there are very compelling reasons. I like cheap, dumb devices. Maybe I should say I like simple, inexpensive, well-built devices. Maybe no-frills is another way of saying it. Sales reps keep trying to sell me hubs and routers with all kinds of extra diagnostic and management tools. We do buy a few of those items where the usage dictates but not generally. I think that SNMP is going to be the thing to watch. In a few years, the additional cost for having it on a device will be small, and at that time I will probably consider SNMP a no-frills option.
DISCUSSION SECTION
With the rapidly increasing use of graphics and multimedia materials in teaching and learning, there will be pressure to move images from "anyplace" to "anyplace" on campus. Faculty and students are asking for on-demand personal computer access to the campus network from the library, classrooms, laboratories, studios, meeting rooms, residential areas, etc. This can place significant connectivity and transmission burdens on a campus network. The roundtable participants were asked if their network planning involved such "high end" strategies as fiber to the desktop and docking stations.
**********************************************************************************************************
Question: Should colleges be installing fiber directly to the desktop as some network designers have recommended?
**********************************************************************************************************
It depends on what you want at the end of that connection in the near future. If it is going to require high capacity connection, then fiber makes sense. It is pretty much an economic decision, in my opinion. If you don't need the high capacity in the next three years (I usually use 3 years for a decision-making frame), install twisted pair. In three years, the cost of the fiber equipment will drop dramatically and/or the speeds on twisted pair will increase. Either way, you're better off waiting.
**********************************************************************************************************
We are not doing so. We don't see the need, with Category 5 copper. That's supposed to handle 100 Mbps. Who can say what the economics will be in 5-10 years, but right now the adapter cards to handle that kind of bandwidth cost way too much. I know technology keeps changing rapidly, but how many small schools are planning in the next 5 years to go beyond ethernet or Token Ring for anything but the backbone? And, with ethernet switching, there is a way to get greater _effective_ bandwidth to servers or power users. If someone has a justification for that expense in today's tight budget times, I am certainly open to change my mind.
**********************************************************************************************************
I'll vote NO. Copper capacities are adequate now...we'll see if fiber is ever needed.
**********************************************************************************************************
Question: Are there any special design characteristics (e.g., docking stations for student computing devices) which should be considered for instructional spaces such as classroom, labs, and studios?
**********************************************************************************************************
My intuition says that it is too soon to know what is going to settle out of the technology for connecting student-owned machines. If your institution issues machines to each incoming student, then you can have good control over the kind of connection.
**********************************************************************************************************
I believe that soon most students will have a portable micro at college like we had a typewriter. These will run on batteries when they are used in most classrooms. Where we have computer classrooms now, it may be better in the future to just have network connections. It sure would be nice if we did not need to maintain current micros for student use, and they brought their own. Standards will be important to make this work.
**********************************************************************************************************
Docking stations tend to be machine-specific and almost as expensive as a real computer. We encourage diskette transfers and dial-up transfers. Provision of network connections makes sense. Those of us running TCP/IP need to be careful however, to avoid address conflicts when someone plugs into the net.
**********************************************************************************************************
Question: Are there any special design characteristics which should be considered to allow for the distribution of multimedia materials over the campus network?
**********************************************************************************************************
Speed! If you're serious about delivering multimedia, especially real-time video, over the network, you're talking about considerable capacity. And, you may have to design in capacity for much more than the average load. You don't want to get 10 minutes into a 20 minute session in a real-time video presentation and have it start acting strange because the overall load on the network reduced your affective bandwidth. Also, I think that there may tend to be more fileservers in a multimedia system just to distribute the heavy load over more CPUs.
*********************************************************************************************************
We pulled coax in when we put the inter-building fiber in the ground. The extra cost was minimal, and we'll be using that this year. Multimedia on the network is a little further off for us. I expect the demand to come when faculty start seeing multimedia applications on the Internet. Similar, low-cost, multi-way conferences among colleagues at different institutions might be a much more attractive thing. Right now, optical disk players are big, but they aren't connected to the network.
**********************************************************************************************************
Yes, we will need 100 MB. These applications do need to be on the network to share disk space. The files are much too large to install on stand alone micros. If they can be isolated from the rest of the network, that is good too. Imaging systems will need the same consideration.
**********************************************************************************************************
Divide, segment, isolate. Using a collapsed backbone inside a concentrator makes for easy segmentation, easy administration, and (as the need arises) easy implementation of new switching technology to get high bandwidth.
SECTION 5: BUILDING THE IDEAL CAMPUS NETWORK
Planning and building campus networks often proceed together in an iterative fashion. Even if we share a common vision of the "ideal" network environment, each campus network will have to be tailored to that campus. Here are some suggestions about ways to proceed in building your ideal network.
•Aim for the ideal
Design the network to reach everywhere, including all offices, classrooms, labs, studios, student residence areas, and auxiliary enterprises (food service, bookstore). Don't believe it if someone tells you that a particular building will never need to be connected to the network.
Students will need access to the network from the places they live as well as from classrooms, laboratories, studios, the library, and any place else that students work and study. Plan to wire the student residences for data, voice, and video. This is one of the major areas of network development on campuses today. Not only is the concept important, but there are revenue generating opportunities as well. For example, some colleges remarket long-distance telephone service and cable TV service to students.
•Get the power
Some person needs to be designated as the final authority on all network development in order to coordinate planning and construction, avoid competition among stakeholders, gain access to all campus buildings and spaces, define standards, and name everything which has to be named (e.g., rooms, wire closets, jacks, and sometimes even buildings). In most cases, the necessary authority should reside in a position which reports at a strategic level in the institution, which usually means either to the chief executive officer or someone else who reports to the CEO. The responsible person needs to have a broad institutional perspective, be a good communicator and consensus builder, and be a pragmatic visionary who can lead the campus in developing a network for the future as well as the present.
•Build alliances
Create campus alliances with important stakeholders such as the maintenance department, the telephone department, and the library. Use these alliances to help in planning and construction.
•Develop and enforce standards
It is very important to have clear institutional standards for all the components of the physical network, as well as the software and hardware which will provide services on the network. A campus without standards, make decisions on the basis of convenience instead of institutional strategy, will face costly and wasteful "entrepreneurial anarchy." Diversity is a great concept in many aspects of collegiate life, but technological diversity is expensive in terms of time, money, and effort. Standards make it possible to maximize use of human, financial, and physical resources.
Without standards, it is much more difficult to expand and upgrade the network, to train and deploy technical staff, and to support users. These support issues have a direct impact on the quality of the technology environment. At the construction stage, standards must be defined in order to proceed with the basic steps: write RFP's, compare bids, make purchases, test wiring.
•Be rigid about flexibility
Make the design as flexible as possible. Assume that all routers, hubs, bridges, and servers will have to replaced sooner than you think. Put in plenty of extra conduit and extra wire.
•Too much is never enough
Use lots of fiber between buildings, and Category Five Unshielded Twisted Pair copper wire within buildings. Put in lots of extra conduit wherever and whenever possible.
•What's fast enough today won't be tomorrow
Expect to use Ethernet speeds (10 megabits per second) now, FDDI/CDDI speeds (100 megabits per second) in the near future, and something faster (e.g., ATM) eventually. Using adequate fiber and high-quality UTP wire should make it possible to move up to any speeds likely to be necessary in the foreseeable future.
•Think comprehensively
Include network support facilities such as wiring closets, security systems, power conditioning, and lightning protection in the design. Expect to wire everywhere. Anticipate new buildings, campus reconstruction, and building renovations.
•Take the long view
New technologies such as the World-Wide Web, wireless communications, networked multimedia systems, etc. will have an impact on the way we operate. Try to see how these things will affect your campus networking plan. For example, if you have a sizable commuter population, how will the campus network be available to them? What role will wireless communication play?
•Don't forget to remember history
Record the details. Document and map everything. It is amazingly hard to remember where underground conduit runs a year or two after it has been buried without very detailed maps. It is difficult but worth trying to get the contractors to provide "as built" diagrams after a construction job is finished. Keep copies of all RFP's, plans, diagrams, and blueprints. And keep them in a logical, central place, not in the head of a single individual. It may be worthwhile to invest in a CAD package and scanner to help record the construction details.
•Watch it grow
Network use is one kind of activity which never seems to decay or decline. Even as some types of usage mature and level out, other network-based applications arise and consume new chunks of the network resources. So in a sense the network is both a fixed part of the campus infrastructure and a consumable resource which has to be continually replenished. Growing the network to keep up with the demands for more connections, more speed, and more bandwidth requires a continual supply of funds, not necessarily huge but at least predictable and regular. Justifying regular funding may require cost/benefit arguments based on network use, thus necessitating some set of metrics which measure such things as live network connections and network load/traffic, as well as things like number of people using the network, number of network connects to the on-line catalog, number of hits on the college Web site, etc. It is very useful to measure any kind of network-related activity which might help administrators understand the role and impact of the network on campus life.
DISCUSSION SECTION
The ideal campus network was described above as comprising five dimensions. The roundtable participants were asked to compare the way their own campus networks are being built with the ideal characteristics by answering a set of questions on this issue.
1. Physical Dimension
******************************************************** Questions: The author
claims the ideal campus network should be universal and reach every place on
campus where "knowledge workers" live and work. How close is your campus to
this ideal? Is the physical infrastructure in place to support every user from
every point on campus? How soon to do you expect to have the physical infrastructure
completely in place?
********************************************************
2. Protocol Dimension
Questions: Are you supporting multiple protocols? Are you moving towards one or two in particular?
3. Management Dimension
Questions: What are your biggest concerns about network management? Are you satisfied with the overall management of your network? How familiar is your staff with tools such as SNMP and RMON? Are any of you using high-end network management software such as HP Openview for Unix, Spectrum, Sun Net Manager?
4. Application Dimension
Questions: Does everyone on campus have access to the full suite of Internet applications? Do you have a campus wide information system (CWIS) using gopher or the Web? Is your CWIS widely used?
5. Cultural Dimension
Questions: Are you at the point where the payoff for the networking infrastructure is being realized in terms of academic and administration applications? Are faculty at your institution using the *network* in new and creative ways to enhance teaching and learning? Is your department and institution involved in improving administrative operations and reengineering administrative systems? Do you have in place a structure for ongoing assessment of administrative efficiency?
**********************************************************************************************************
Bill Doemel:
1. Physical Dimension.
August 1, 1995.
2. Protocol Dimension
We now support Appletalk, IPX, TCP/IP, LAT, and Decnet. We are moving to eliminate Decnet and LAT.
3. Management Dimension
We have a growing network and are preparing for major growth this next year. In preparation for that growth, we standardized on hubs--Cabletron--and on routers--Cisco. We are in the process of implementing Cabletron's network management tool. We also have a substantial investment in Cisco Routers and our network manager is currently being trained in the use of Cisco equipment.
4. Application Dimension
Yes, but not all are trained in its use. We are using Netscape as the primary access tool.
5. Cultural Dimension
We have had our campus network in place for five years and are now moving to extend the network to the living units--fraternities and dorms (we are an all-male liberal arts college). During that time, use of the network has increased substantially so that more than 60% of our students are in classes that actively use the network. The biggest impact has been the small network communities that we create. For example, a class in history has a directory or folder established for their class. Within that directory, we create a private subdirectory for each member of the class and for the professor. We also create a directory, called the commons to which all users have access. All of the members of this community belong to a group and can thus send e-mail to each all members of the group using the send to group function of Pegasus mail. Users deposit their documents in their folders. The Professor, having access to all folders, reads the documents, marks them up and may use them in class to illustrate a point. We also use these class communities to create team folders so that a team working on a project will have a folder in which to share their work. These network communities have been set up for some 30-40 classes this semester and are actively used by students and faculty. These special accounts have been used by faculty for these past five years, but the biggest increase in use occurred during the past two years--mostly because of word of mouth. The faculty users have become the biggest supporters of this technology.
For us access to the Internet is new, added last fall. Nevertheless, use has mushroomed with 80% or more of our 56 Kb bandwidth being consumed between 2:00 PM and 10:00 PM. Use has moved from casual recreational to course related. Indeed in one course, the instructor established links with students on several campuses to open discussions to a diversity of views. Our language faculty are aggressively incorporating Internet tools into their instruction of language so that students have the opportunity to exchange ideas using the native language. We are preparing to increase this use substantially as we move into a new language lab facility that has 18 computers and 2 tape machines -- quite a change.
Our Economics department is aggressively exploring ways to improve the teaching of economics using computers and were one of the first departments to establish a home page on our network. They have shared their success with other colleagues and those colleagues are moving to incorporate computers into their classes. The human toll of this adventure is substantial, however. For faculty, it means that to introduce this technology, they must learn it and develop teaching materials on their own time. The faculty is now urging the administration to allow released time for educational innovation in the same way that they release time for professional development.
**********************************************************************************************************
David Smallen:
1. Physical Dimension.
Our complete campus network is scheduled for completion September 1995. We are in the process of wiring the entire campus with voice/data/video to 2500 locations in over 50 buildings including all residence halls (one jack per student).
2. Protocol Dimension
Support TCP/IP, IPX, LAT, Localtalk. Will move towards IP and IPX.
3. Management Dimension
We will be probably using Openview to manage the entire network. We have standardized on 3Com smart hubs in buildings and Alantec Switching Powerhubs at the central points. Since all jacks will be live we can manage the entire shooting match from one location. The Alantec hubs allow one network probe to be moved to any segment. We expect to be using some type of RMON device. We have no substantial experience with any of this yet.
4. Application Dimension
Everyone has full Internet access. Our CWIS currently uses gopher but we also have a Web server up. Minimal use thus far as everyone appears to be too busy to supply timely information. This will change as the network becomes more pervasive.
5. Cultural Dimension
We are not yet at the point of receiving the payoff for our investments. I'd say more likely by the fall of 1996 we will have a much better opportunity to assess the benefits.
**********************************************************************************************************
Ken Pecka:
1. Physical Dimension.
We are planning the expansion of the campus network to every dorm room, office and classroom. The project is scheduled for completion in mid-August,1995.
2. Protocol Dimension
We currently support Appletalk and TCP/IP. We are moving to TCP/IP exclusively.
3. Management Dimension
We currently do very little "network management" in the way of monitoring and fine tuning the network. It is a major concern as we consider the potential growth of the campus network. We will be hiring a full-time network manager in July (if not sooner). We currently own HP Openview but make very little use of it.
4. Application Dimension
We are in the process of providing users with Internet applications. At this time it is a partial set of tools including E-mail, Gopher, WWW, telnet, and FTP. We have established a CWIS committee to design and direct the development of the CWIS.
5. Cultural Dimension
With our current network (reaching only 45 percent of the faculty and no dorm access) the payoffs have been limited. However, there has been an increasing demand for "access" to information resources by both faculty and students. The completion of the network this summer will position us take full advantage of academic and administrative applications. We do have a limited number of faculty who are actively utilizing the network to enhance the curriculum. At this time we have spent very little time involved in reengineering support or direction. The majority of our efforts have been in operations and maintenance as well as planning for the network completion.
DISCUSSION SECTION
When colleges and universities renovate or build new buildings, it is very important to incorporate a networking infrastructure which will, if possible, be adequate for 20-30 years. It is difficult to know exactly how to design a building with sufficient flexibility so it will be technologically adaptable for two decades. With the rapid advance in wireless technologies, some institutions may find it feasible to build wireless strategies into their network planning. The roundtable discussants were asked their opinions on this topic.
****************************************************************************************************
Question: How do you think WIRELESS networking options (e.g., microwave, laser, infrared, radio wave) will/should affect the design of small college networks?
*********************************************************************************************************
John Balling
One application for wireless communications is for temporary networks. I’m thinking of something like a temporary laboratory or a group of workstations moved into a gym for arena registration. We are actively considering wireless for these kinds of applications. There may also be a need to put a network into a building that really cannot be wired--possibly because of its historical value. There may be situations in which getting a wired network connection to a building or set of buildings is cost prohibitive. Certain types of wireless may allow field researchers or field monitoring devices to send data back to a host computer on campus.
Those are just a few ideas that come to mind. At this point, I don't see wireless taking the place of wired links for general campus networking--at least for quite some time. From the bit I know of wireless, the technologies are still changing rapidly and they don't seem to have reached to the point where they could support an entire campus. I also wonder about the real bandwidth of most types of wireless and network security.
***************************************************************************************************
Morris Galloway Jr.
In addition to the two reasons already mentioned -- historic buildings that cannot be wired, and temporary networks -- we might add crossing public right of way when permission to string wire cannot be obtained, redundant or backup routes, coverage in the library (as opposed to wiring all study desks) and the same for classrooms in which students are expected to use notebook computers.
***************************************************************************************************
Ron McCleary
I think wireless technology is pretty mature with regard to concentrated, longer distance communication. For the LAN, I would like to see the technology mature a while longer before making commitments.
I think that copper and glass are always going to be the preferable way. In 50 years, regardless of the standards and protocols, the largest, clean bandwidth will be on fiber, followed by copper, followed by wireless. There may be periods when copper will leapfrog fiber or wireless leapfrog copper, but given a little time, the order will be reestablished. Wireless makes sense when it isn't economical and practical to wire.
*************************************************************************
Ken Pecka
When I think of "network design" my mind tends to focus on the physical network characteristics. The thought of "wireless" options seems to leave me with a variety of options, not restricted to "physical" parameters. Therefore, I guess wireless options will always be possible regardless of the physical design of the network and can always be considered as we expand and grow our networks.
I agree that our primary designs should focus on the use of fiber and copper as distribution media. And to date, the cost effectiveness of the wireless options is limited in most cases. A design based on fiber and copper will not limit the use of wireless network option in the future. While a design based on wireless options may limit future fiber and/or copper options in the future (ie. a building designed without conduit and wall ports because a wireless network is planned).
So, I think I have convinced myself that wireless network options should be given "limited" consideration in the "design" of our networks.
DISCUSSION SECTION
Development and implementation of standards are major issues at educational institutions. The tension between centralization and decentralization of management responsibilities for networks, along with issues of technology support requirements, budget and funding structures, perceptions about the meaning of academic freedom, etc. all complicate the standards issue. The roundtable participants were asked their opinions on the standards issue.
**************************************************************************************************
Questions: What would you say about standards to colleges which are just getting started in networking? What do you think about the importance of network standards? How loose or rigorous should they be? Who should define/enforce them? Do you have clearly articulated standards on your campus? How important do you think standards are to the future development of your network?
**********************************************************************************************************
Morris Galloway Jr.:
We have settled on SNMP as a management protocol standard. Individual devices don't have to support SNMP, but all hubs do. Right now we are trying out several TCP/IP products for DOS/Windows. We will choose one or two that have all the features we need, and stick to them. We'll probably also pick one or two Ethernet cards to support.
******************************************************************************************************
Jim Nash:
The more we do with the network, the more I feel the need to adopt standards. Particularly now that budgets are tight. Allowing multiple standards or having none at all was a luxury of the past. I believe that we have an obligation to our institutional to adopt standards, and resist deviations from them. There are times when exceptions can be justified. Personal choice is not a good reason.
We have 4 times as many PCS as Macs on campus. The Macs are a constant problem for us on the network. I am sure that this is a result of our greater knowledge of PCS, and lack of knowledge about Macs. I have heard other colleges say just the opposite that Macs are no problem and PCS are a pain. When we have multiple standards, we do not seem to give equal service. It is better to have one and do it well, then to have more and do it poorly.
We have a standard PC (Gateway) and Mac. The models change each year, but we only deal with two micro vendors. We have standard lasers (HP), network cards and concentrators (Cabletron), software (WordPerfect, Borland, Novell, MS-DOS, Windows, System 7, NCSA TCP/IP, Multinet, VMS), and servers (DEC). There are many other products being used on campus that are not part of the standards. These are the products that take a disproportionate amount of time because we are not familiar with them. We do refuse to support some products. It is difficult to get even the standard products all working together on the network.
You need a committee or group to set these standards, and support them. If you try to do it on your own, you will surely meet resistance. Even with broad based support, some will not follow.
*************************************************************************************************
Thomas A. Warger:
I have found that standards have become our guideposts, where in the past we looked to major brands. Now we know that if the network equipment we are investigating complies with the standards suite in our strategy, then we know that it "qualifies." Based on our experience, it is very important that adherence to standards be strict because network troubleshooting is hard enough without running into avoidable incompatibility.
My belief is that standards have to be set by the central computing outfit on campus. But at the same time, those standards should be reviewed with the appropriate consultative bodies to test out whether the statement of standards is clear and sensible. In my experience, nobody outside computing services on campus will know or care much about things like TCP/IP or SNMP. But they will holler if you tell them their departmental LAN is going to be cut out of the campus network because you no longer support (Starlan, TOPS, whatever).
Our campus standards are not well articulated. They exists as an evolving set of decisions with our central computing support organization. It has not seemed necessary to publish all the details. For the foreseeable future, we expect standards to be the criteria by which we understand and plan the expansion of our network and its evolution through technologies we don't even know about yet. Blue sky dream: network electronics with the smarts to recognize and react to other network components that subscribe to outdated standards or competing standards or variants of standards.
*********************************************************************************************************
Paul Bishop:
This is probably one of the more interesting issues that we have addressed or at least it's one that has attracted my attention on a regular basis. I see the issue along a thin grey line that runs from hardware and protocols to application software, with pros and cons at every point. In MHO standards are most valuable at the hardware and protocol level. We've standardized on 10BaseT and LocalTalk within buildings for network connections and AppleTalk and TCP/IP for the supported protocols. However the issue of setting standards for application programs is a much thornier issue. On one hand it would be nice to only support one word processing program, one communications program, one graphics program, etc. However my experience has been that in the diverse academic environment one size does not fit all.... especially software. Even though I've fought software standards as too restrictive, a number of de facto standards have emerged over the years and are now causing headaches. Microsoft Word and Excel have become campus standards by simply being there with needed features seven years ago with no apparent competition. Both have become slow and bloated over the past few years, with too many upgrades and incompatible file formats. Yet the mere mention of other packages that might provide the features we need in a more palatable package is met with gasps of horror from faculty and staff. Sort of reminiscent of the inertia of social systems and paradigms as described by Kuhn. In many ways the luxury of working at a smaller campus is the ability to support a greater diversity of applications and retain that ability to quickly recognize and select a better product when one appears. I will confess that this attitude might reflect the high percentage of Macintosh computers on my campus. In many instances I find myself helping an individual with a software package I've never seen before, merely helping that individual to articulate their needs into something related to what appears on the menu choices within the application.
*****************************************************************************************************
Bob Hodge:
Standards for networking are an evolutionary step from standards for other types of high tech. Especially in an academic, free to be environment, "standards" have a very negative connotation. I suspect that every new technology or invention goes through a phase of competing/conflicting standards. I read where Edison may have acted as less than a gentleman to promote DC power over AC. Eventually, that got resolved so we don't have academic committees discussing types of power, or television standards, or telephone systems. Networks are in their infancy, so there will be long years of discussion; our grandchildren will laugh at these primitive systems which today seem to be so state of the art.
We have agreed together, in our university, that related systems should be highly integrated to maximize their usefulness. That implies common networks as well as platforms, databases, etc. or at least simple pathways between them. Standards, in this situation, do not assure that everybody does everything the same way as much as it requires everyone to assume individual responsibility to provide the simple pathways. With decentralized computing must come decentralized responsibility.
The operating environment of just a few years ago included an administrative system which was highly integrated (when it all runs on the same machine, who cares about network standards?), an autonomous library which didn't have much interest in anything other than paper, an audio/visual center whose high tech components were primarily from the broadcast industry, and an computer science department who set off on their own to assure that it could be state of the are. Today, they are all running on the same network in various states of integration and with varying degrees of satisfaction. But, they do all acknowledge that without the common network, none except the administrative system could be successful by themselves.
So, we agree on integrated systems and the need to share networks from a practical sense. Living together is a big deal at our residence based university so "house rules" are common. Network standards are "house rules", many of which are left to the end user to abide by rather than the IT folks to police. They do know that if their connection messes up the network the solution is simple and swift - they will be unplugged.
Where do they come from? The department who is most knowledgeable, supervised by a person who is responsible for listening to the needs of all participants to assure the maximum benefit for the university as a whole (not necessarily the collective benefit of each department). Call the department the IT department and the senior administrator the chief information officer who plays a role of entrepreneur, decision maker, and I believe in the case of networking, benevolent dictator. (We don't seem to question the controller about accounting standards, so why does it seem to require a committee to talk about wiring standards or protocols?) Not all will agree with me on this point.
I believe the breadth of standards is a balance between the ideal of free-to-be academics and the practical limitations of what can be supported. As such, it appears to be an administrative issue. The context is important. How loose or rigorous should standards be? As broad as can be reasonably supported. What should they be? That is a campus issue based on the needs of the users, the history of the campus in terms of existing technology, experience of those responsible for making it work, etc. I don't see a "best" set of standards for small colleges (e.g.. we use Pathworks to provide LAN type services because it is a natural extension to our existing equipment and experience with Ethernet/Decnet; for those who are starting new at other colleges, I tell them to go with Novell because it is ubiquitous and a lot of stuff runs on it.
How do we address those who want something else? First, the network should be broad enough to support different speeds, protocols, etc.. I think we now support four protocols on the same Ethernet which covers the vast majority of requests. Then, I get the end user to take a piece of paper and put two dots on it. They find that they can connect the dots with one line. But, add just one more dot, and it takes two more lines. One more dot takes three lines. Users want the dots; IT assumes the burden of the lines. We usually come to an understanding pretty quickly that more lines are difficult to draw, especially if the user has to pay for them as well as their dots of functionality.
This all takes on new dimensions as we are addressing not only traditional data on the network but digitized voice and data as well, and doing it for areas and departments which have previously been autonomous or at least not involved with shared technology. I suspect that we will have a number of discussions until we fully understand all that is needed and the technologies stabilize to support what we want to accomplish.
I see the issue more as a cultural one than a technology one. Integration is the goal, technical standards are some objectives by which we attain the goal.
***************************************************************************************************
Leo D. Geoffrion
At Skidmore we just completed an extensive self-evaluation process. In one of the more interesting role reversals, we found our faculty pushing for standards even more vigorously than the Computer Center people. I think users have seen the dark side of personal autonomy and recognize the convenience of having all systems behave consistently and predictably.
Personally, I've been less rigid about standards than most -- primarily because I've found that standards are largely a myth. Nearly every standard out there -- at least the ones that concern users -- has a dozen variants, extensions, or other hacks that call into question whether it can ever be called a single standard.
On the international scene, I'm disgusted by the continuous "war by standards". Whenever one vendor threatens to dominate a market, the others team up and invent a competing standard whose goal is more to muddy the marketplace than to provide superior functionality.
We hold to a few basic network standards, such as ethernet, Netware, TCP/IP, SNMP, .... Even within these, there are often great internal debates, such as which Netware drivers to use (ODI, NDIS, PACKET, ...) and which editions of Netware should be installed. It's important, however, to remember that very few of our users care what networking protocol is used. The important standards are those that affect how the user works.
In some places, universality is critical. The best example is telephony and electronic mail. Both are worthless if they can't connect to all other systems. Hence, we "own" those applications and enforce standards vigorously. Even here, our clean standards will likely crack in the coming years as more vendors bundle their proprietary mail systems into the operating system or the software suites. It will be increasingly difficult to tell users to abandon "free software" that comes pre-installed on their new computer.
In other places, universality is less important. For example, most users would rather fight than switch to a different word processor or numerical analysis package. We've largely stopped trying to argue for standards and a. inform users which software packages we'll support. b. provide translators for the nonstandard packages - if available.
The bottom line: There is no simple answer to the standards question. I find it a constant process of negotiation -- both internal to the College, and external to the larger computer industry.
*****************************************************************************************************
John Langeland:
We are finding the "faculty pushing for standards" in the context of student or "instructional" computing. There is an interesting interplay at work between networking standards and computing the more general notion of computing standards.
All of us have been witnesses and participants in a variety of debates and decisions about campus computing standards. Not all that long ago the focus was on mainframe and mini operating systems and software; currently we all participate in the Mac/Windows/OS2/DOS/NT/UNIX (in its many variants), and at the same time debate standards issues about applications software for wordprocessing, spreadsheets, databases and communications software.
But standards issues about networking per se have not really been hard, at least not for us. Part of the ease with which we've found our way to our current hardware and software standards can probably be attributed to the fact very few members of the campus community have any real interest in the details of how data communications works. They have great interest in reliability and functionality, but like electrical power, the telephone system, and the plumbing that keeps the toilets in working order the details don't matter. Networking is out of sight and mind, and as long as it provides the required functionality most of our users don't think about them. As a result, the emotional debates and habits that sometimes enliven discussions about operating systems and applications software haven't played a significant part in setting standards.
Given the prevalence of Ethernet-based, fiber-backbone, UTP house wiring with multiple protocols, I suspect that most academic institutions have been able to develop a cost-effective strategy (and usually the same cost-effective strategy) to provide highly inter-operating data communications paths. This local strategy and the standards associated with it were, for us at least, not so much a decision to standardize as it was a decision to commit to a certain level of data communications functionality. Our current "standard" is Ethernet speeds, but we know that FDDI and/or ATM data rates are not too far in the future. Within our Ethernet speed standard we feel free to shop for hardware and software that will inter-operate on this network. And *surprise* -- everything does (well almost!). TCP/IP with some DECNET and AppleTalk and something for your PC LAN all cohabit and more-or-less interoperate as required.
But what I've found intriguing is that as networking has become increasingly functional and ubiquitous, concerns about software that doesn't inter-operate has generated calls for increased standards for computing platforms. One faculty member recently articulated this by saying "I'd never use one for my own research, but what I wish would happen is that every student would compute on a Macintosh connected to the network using the same set of software. Then I could focus on substantive matters rather than spending time teaching a few students how to use a Macintosh." There have also been statements like "every incoming freshman should know how to use e-mail from the first day of classes" so that I can use e-mail for discussion purposes and course business.
So I see a new dynamic in which the ability of students and faculty to use networking in course work creates a new set of interests in common, easy-to-use, inter-operating applications. I personally find this a refreshing change from the celebration of chaos that attends the "whatever and however you want to compute" mentality that some view as appropriate in academic environments.
The focus on networked functionality generates and builds upon networking standards like TCP/IP and Z39.50 and also upon software like Mosaic and Gopher. And these networking software standards with a client-server software strategy have finessed many of the problems inherent in a heterogenous computing environment. Whether it's UNIX or Mac or PC you can go net surfing with some style because Mosaic and gopher and FTP etc works well in all of these environments.
But having mastered networking, some users then discover that they can't share a wordprocessing document or a spreadsheet or a database file with their colleague down the hall or across campus because of lack of standards about applications software used on campus. And then they ask the question why we don't have better standards in place on our campus for software used for instruction and administration.
*************************************************************************
David L. Smallen:
Standards are essential in an environment in which you want to provide a high level (quality) of support to users of IT, without increasing central staff size. Our experience has been that at the lowest levels of networking connections (the protocols, the wiring, etc., the campus is quite willing to delegate the responsibility for decisions to others - in fact care little, and know little about this area). When you get to the application level things become a little more muddled. As we all know, things are application driven. One example. We went with a Pathworks/Ultrix solution for networking a DOS/Windows Lab because it was consistent with our knowledge base and appeared to offer similar functionality to Novell for what we wanted to do. While this was probably a valid argument, we have had to deal with second guessing that one program or another would run better if we were using Novell.
We have centrally agreed that TCP/IP, Novell, and AppleTalk, will be the standards at the network protocol level. It would be nice to have just one, but I don't think that is possible since some of our applications require one or the other. I think we can support this suite, as long as people are willing to work towards becoming self-sufficient at the higher (application) levels.
For similar reasons, I don't think it is realistic to support only DOS or Macintosh exclusively. At present applications in a variety of disciplines only run on one or the other platform. The best approach we have found is to think of discipline-specific applications from general purpose applications such as word processing, data base, communications, spreadsheets, etc.. The central organization can only be knowledgeable about a few of these general purpose applications so we have (mostly historically) chosen one dominant general purpose application on each supported platform (e.g. Writenow, Excel, Filemaker on the Mac; WordPerfect, Lotus, PC-File on DOS machines). We recognize that individuals might choose to use something else but they couldn't expect us to be knowledgeable - although we certainly try and often succeed in helping them because most problems can be solved by analogy. This is just a practical reality. Notice, I purposely did not say "we support this, and not support that". We have found this to be counterproductive and portrays the central organization as "gatekeepers" rather than helpers - and often cuts off discussions at too early a stage. Most people can be helped in a reasonable amount of time by knowledgeable, service-oriented IT staff, and it helps the IT staff to do their job better if they are not viewed as obstacles. Of course, they have to know when to say they have reached the end of their knowledge base.
So, in summary, I think standards are important, but standards should not be thought of (or perceived) as barriers.
********************************************************************************************************
Ron McCleary:
Have a general plan for the network you want in 5-10 years, then be sure that each action is a step in creating that network. Fiber and copper are relatively cheap in comparison to labor, so put lots of it in.
The devices on the end of the wire and fiber (e.g. hubs, routers) are relatively expensive and go obsolete fast. Look for solutions that intuitively feel good for 3 years. You can always replace the stuff in 3 years with better, newer, and faster stuff and move still working old stuff to other, low-use areas.
How do standards fit into this strategy? They are what makes the strategy work. The strategy recognizes that a network is a continually growing and changing entity. The only way that the growth and change can happen with the least turmoil is to follow standards so that each new component works with what is already there. Standards are the only way to insure "plug and play" with regard to the network.
What do you think about the importance of network standards? Very important. I especially like the ones that are not very tied to specific vendor; that is, I like 10BaseT over Token ring.
How loose or rigorous should they be? I think they should be rigorously adhered to. You end up buying stuff from so many different vendors and putting it together yourself. Who wants to buy tinker toys from several different vendors only to discover they don't fit together after you spend the money.
Who should define/enforce them? I assume this should be done by the Computer Services or whatever you call the department responsible for installing and maintaining networks. If that is not done by a single department, there are probably going to be problems getting things to work together later.
Do you have clearly articulated standards on your campus? We have standards, but they're not "clearly articulated." All network installation and maintenance is done by just a few people in my department. We have been able to stick with standard just because of our size and because installation and support is centralized.
How important do you think standards are to the future development of your network? I think they are very important. We'll always be buying new stuff for our network. We don't want to get 3-5 years into networking only do discover that we can't connect stuff the way we want because we bought something that didn't fit our standards.
*****************************************************************************************************
Bill Doemel:
What would you say about standards to colleges which are just getting started in networking? Keep it simple. We standardized on hubs (Cabletron) , NIC cards (Asante for Macs and Intel for PCS), routers (Cisco), and standards for 10-base-T wiring and termination (Level 5). We support Macintosh computers and Gateway (you can purchase replacement parts from a number of sources).
What do you think about the importance of network standards? For example, by having one hub manufacturer and their software, management of the network is much simpler. We have a hot spare for the system. We also have spares for the other systems.
How loose or rigorous should they be? Standards may change from year to year simply because of manufacturer changes. Before we selected Cabletron, for example, we had Asante, and several other hubs on the campus. Also, companies change. For example, for 4 years, we purchased only Zenith computers then Zenith service and cost/performance dropped and so we did an evaluation on systems.
Who should define/enforce them? We have found that a central group that includes reps from all over campus -- academic and administrative provides the political will to agree to standards and then enforce standards. Departments can't go out and buy equipment -- only CS can purchase equipment.
Do you have clearly articulated standards on your campus? Yes.
How important do you think standards are to the future development of your network? To minimize staff costs and problems, standards are very important for us. We have 7 people who manage about 300 nodes. We are moving to extend the network to dorms and fraternities - about 900 potential new nodes. Standards are important if you are going to maintain such a system.
DISCUSSION SECTION
It is very useful to record as much data as possible about computer use and then boil that data down into a few key statistics which can be used with Trustees and senior administrators to describe the growth in computing, and hence, the need for more resources. This was fairly simple in a terminal-host environment. But metrics in a network environment are much harder to define. And yet we will continue to need meaningful metrics to define and measure use of the network and related resources. The roundtable participants were asked their opinions about the use of networking metrics.
*********************************************************************************************************
Question: What do you use for networking metrics at your campus? Have you discovered any understandable, explainable, easily interpretable metrics to tell the story of network use/growth?
*********************************************************************************************************
Dave Rotman:
We use two metrics now, though we plan ("as time permits") to do more work in this area.
1. E-mail utilization
WordPerfect Office tells us how many messages are routed between servers. We are averaging 8,000-10,000 such messages per day.
2. User opinion
In a December survey, 95% of the faculty and 66% of the students reported using the network daily.
*************************************************************************
Leo Geoffrian:
This is a very apt question -- one for which our ideas are still evolving.
Here are some of the things we're looking at. I've added a few of the more interesting findings within parentheses.
1. Electronic Mail
a. The percentage of the community that have E-mail boxes on our central mail server
b. The percentage of people who have E-mail that is unread for more than 2 weeks (e.g. idle mailboxes that no one is reading). (75% of the mailboxes are read at least once every 2 weeks)
c. Number of mail messages sent/received per day.
2. Network registrations:
On our net, all devices must have an IP address that is obtained via bootp. Thus, we can count the number of entries in the bootp tables.
3. Word processing:
We asked some of the faculty who assign essays to estimate the percentage of papers submitted that were obviously from a word processor. (nearly 100% of the papers)
4. Printer counts:
We have a batch job that obtains a count from each printer of the number of pages printed in the previous week. We are installing similar code on our Netware servers to maintain page counts off those printers as well. -- only a subset of printers support this feature. Fortunately, they are our most popular printers.
5. Head counts:
The resident assistants in one dorm went from room to room to count how many rooms contain at least one computer. (The count surprised us -- 60% -- much higher than we had guessed).
Over the past 5 years, the number of interactive logins on our campus has decreased steadily -- and we're even proud of the downward trend!
We interpret it as evidence of the shift from traditional time-sharing to client/server applications such as gopher, Eudora (POP mail), Usenet NEWS, as well as pure microcomputer nets, such as running stat packages inside of Windows from the Netware servers. For example, we operate a gopher server that has become very popular among our students and faculty. The accounting info within the server shows that we have about 6,000 gopher connections per month with about 4800 documents transferred from the server to a client. Virtually none of this traffic counts as a login on the systems yet it's evidence of substantial network utilization.
*************************************************************************
Carl Heideman:
At Hope, we use as many of the following metrics as we have time to obtain:
Pages printed: We either use system accounting data (on the Vaxes) or get page counts from the printers (Novell LANs).
Applications used: We count the number of times WordPerfect, Lotus, etc gets invoked on our LANS. We have written a simple program in-house to do this.
Disk Space: We chart the availability of disk space on all LANs and VAXes.
System Benchmarks: We have a standard set of "programs" we run and time to judge system performance. Examples on the LANs:
Time it takes to boot a system that uses a boot PROM
Time to invoke Windows
Time to perform a large WordPerfect mail merge, including the time it takes to invoke WordPerfect and load the files from the server.
We run these benchmarks many times a month in batch and average the results.
We run a similar set on the VAXes, so that we can gauge relative performance
of the LANs to the VAXes.
Modem usage: Total and average daily dial-ins, plus a report of maximum concurrent usage.
*************************************************************************
John Langeland:
1. One very useful metric at Trinity this past year has been the notion of a "active connected machine." The major interest is (of course) for machines owned by students but active on the network. We had to experiment with this a bit to come up with a good measure because all dormitory connections to the network are active when the students arrive. We give them instructions and then (when it works like it should!) they plug and play. This meant we had to find a way to measure connections directly. (We did it with weekend-long samples using Interpol to get Mac connections; we know the handful of PC users on the net.) It's been very helpful to talk about the "900 student-owned machines active on the network"; it's made it possible to refocus attention from our computer labs to the machines in student rooms. We hope to develop some new measures that include all machines -- not just AppleTalk -- yet this semester so we can figure out how many of our own machines are active on the net!
2. Trinity has also found statistics from our e-mail system to be valuable. The most useful numbers so far have been related to the number of active e-mail users on campus, by category. We've calculated the number and percentage of students in each class, as well as faculty and staff who use e-mail regularly. For example, because we can say that nearly 100 percent of the faculty are regular e-mail users, electronic "junk mail" has begun to replace paper "junk mail" for internal postings and notices.
3. We've been working on measures more directly related to instructional uses of the network. Electronic discussion software (we've been using PacerForum) has been very popular; we have 450 students in courses this semester that are using this software. Similarly, we provide a network server that provides file access for PC and Macintosh users for document exchange for classes. This service has 27 classes/sections actively using this document exchange service this semester. While many other classes are using other networked resources, this one requires that we set up folders and security; as a result we know which courses are actively using this networked resource.
4. As our network has grown, I've begun to think about whether meaningful metrics can be extracted from our network traffic. We run an array of network file servers that contain software and services; I've got my networking types exploring means that would allow us to count the number of sessions (or perhaps the number of workstations) that accessed a specific server. Since these servers have generally well-defined roles, we may be able to develop measures like "number of students who used the networked print server" or "number of students who accessed the shared class software server." We don't have any network-traffic based metrics yet, but we've started to think about them.
This approach might be extended as well to begin to understand what use people are making of resources through the Internet; a profile of packet counts by IP address would probably provide some insight into the uses made by students and others of external resources. As less computing happens on environments under our direct control, we will probably need to get serious about indirect measures of the kinds of uses to which networked resources are placed.
The development of metrics for access of network resources (particularly for instructional purposes) is essential if we hope to understand and improve the uses made of this technology. While it's useful in the local context to talk about the number of students who use a particular service; the development of standardized metrics that can be applied in multiple institutions would be of more general interest. I suspect that measures based directly on traffic are the ones that will be most likely to provide more general, standardized measures.
*************************************************************************
Morris Galloway Jr.:
We are only in the planning stages of moving gopher, WWW, and e-mail off the UNIX box and onto PCS and Macs. Right now, we get the statistics on these things from our host. We can, of course, log connections to our root gopher server, and filter them into on-campus and off-campus. That's where we plan to start.
*************************************************************************
Gary Schlickeiser:
We have several metrics we use but I am not sure that any of them really tell the story, including:
1. A Network Funding Plan
Funding is a central issue in planning and building networks. Most colleges build the basic physical infrastructure with some source of capital funding such as gifts, grants, loans, or bond issues. Beyond that, there are a number of other areas where operating budgets or reserve funding will be needed to sustain the network. So part of the network plan should be a financial plan for the construction, support, and maintenance of the campus network and associated services. The basic components of this plan would include:
•Capital funds, for building the basic network infrastructure
•Support funds, for the on-going support of network components and services (e.g., personnel costs, maintenance contracts, Internet line charges, fees for regional providers of Internet connections)
•Maintenance funds, to replace damaged, worn out, or functionally obsolete equipment
2. The Cost of Building a Campus Network
Every campus network is different, so it is very difficult to provide a general formula or model to accurately predict the building cost. For example, a campus with a network of steam tunnels will find it much cheaper to install conduit and fiber than one which has to dig up or punch under city, county or state roads. Evaluating costs gets particularly difficult as the campus network becomes more extensive and complex, since nearly all information technology items can be considered part of the network.
3. Support and Maintenance Costs for Networks
In terms of non-personnel costs, funds will be required for on-going support of items such as leased telephone lines for Internet connections, maintenance contracts on some network components (e.g., routers, servers, some software), as well as repair and replacement of networking equipment. In order to compare network support costs, roundtable participants from ten small colleges provided financial information pertinent to this area. The data included how much they spend yearly on operating budget items like maintenance contracts, Internet/BITnet connections, network software costs, etc. Their answers varied widely, and didn't have any particular correlation with the institutions’ E&G budgets, number of students, or number of faculty. Nevertheless, here is the data:
Network Operating Costs Reported by Ten Small Colleges
| Item | Low | Median | High |
|---|---|---|---|
| Total cost | $15,000 | $38,600 | $160,000 |
| Cost per student | $8 | $28 | $128 |
| Cost per faculty member | $100 | $375 | $1300 |
At the risk of over-extrapolation from a very small data set, it seems likely that small college networking newcomers should expect to spend about $40,000 per year (or about $30 per student, or about $400 per faculty member) to manage a campus network during the early stages of development.
As with all areas of technology, the networking equipment will eventually wear out or become functionally obsolete. The institution will need some strategy to fund replacement equipment. Keeping an equipment inventory, with the expected replacement cost and the expected lifetime for each item, is necessary so yearly costs can be predicted. Since this is a new area for most institutions, it is hard to predict lifetimes. If you assume, for simplicity, that a component will last N years, then each year (100/N) percent of the replacement cost should be deposited in a depreciation reserve fund. It is very important to give the network management staff access to this reserve account on an "as needed" basis.
As with desktop computing equipment such as personal computers and printers, there is little agreement among college administrators about the expected lifetime of electronic components, in part because of the difficulty of defining "functionally obsolete." Some kinds of network equipment may perform their designated function for many years, while others need to upgraded almost yearly to keep up with demands for higher speed, greater bandwidth, or increased security. Institutions need to understand that many kinds of information technology items are just like other essential "consumables" (e.g., water, electricity, toilet paper) which are funded and used continually.
4. Staffing issues related to networking:
What are the staffing implications of networking? What are the new responsibilities which are connected with the development and management of a full-campus network? Here are some of the areas where staff time will be required.
•User Support
Faculty, staff and students will need all the types of user support that computing requires (consulting, training, documentation). User support in a microcomputer network environment may be substantially tougher than in a time-sharing environment, because users can tailor their personal machines.
•Software Support
Network operating systems will need regular management, maintenance, and upgrading. Applications are increasing dramatically in both quantity and variety.
•Technical Support
Someone will have to run cable, make and maintain network connections, trouble-shoot problems, install routers, etc. This kind of service could be outsourced or it could be done by a college employee.
•External Relations
Someone will have to work with vendors and other external partners, represent the institution in connecting to the Internet, talk to parents and alumni who want access to the campus network, and perhaps help with external fund raising.
•Policy Issues
Campus networks provide access to campus mail and campus data, activities which typically generate lots of questions about proper use, ownership, authorization, control, accuracy, security - a whole host of issues.
How many people does this take? Data collected from seventeen small colleges which have fairly well developed campus network programs indicated that the number of FTEs devoted to network support ranged from a low of .5 FTE to a high of 5.3 FTEs, with a median of 1.5 FTEs. Thus colleges which are developing a campus network should expect that all the associated activities will require the work of nearly two full-time people, although this work is typically spread across a number of positions ranging from top-level administrators to technicians.
DISCUSSION SECTION
Institutions engaged in developing a campus network are advised to develop a "financial plan for the construction, support, and maintenance of the campus network and associated services" which includes capital funds, support funds for ongoing costs, and maintenance funds to replace worn out and obsolete equipment. The roundtable participants were asked some questions about their networking funding and staffing arrangements.
*********************************************************************************************************
Questions: To what extent is funding for Information Technology (IT) at your institution centralized? What is the scope of the IT budget? Do departments make individual purchasing decisions and pay for information services to the IT organization? How much control over funding should the IT department have?
**********************************************************************************************************
David Rotman:
All student computing is funded centrally. Some faculty and staff computers are funded centrally, with the trend to move towards one central pool for all computer purchases. Our issue is not so much one of control but one of consistency. We would like to get all of our computers onto a replacement schedule, but many budgetary units (departments) are too small to have a routine budget entry for computer replacement. By centralizing the purchases, we can level out the bumps.
**********************************************************************************************************
Leo Geoffrian:
This is an evolving issue at Skidmore.
In the past: each unit submitted a budget, with the Computer Center responsible for all public facilities and the computer network.
In the present: The Computer Center, library, and Media Services coordinated its budgets. We submitted separate operating budgets but a single capital budget. Simultaneously, the Financial Services people have pulled all the computing capital requests together and are reviewing them as an integrated whole. This is the first year for this integration.
In the future: A planning group is looking at how this all should be done in future years. I don't yet know what they will recommend.
**********************************************************************************************************
Morris Galloway Jr.:
We have recently moved to recentralize purchasing. There are two centers: academic and administrative. All purchases of computer equipment must be approved by the relevant center and by the business office. Budget money has been centralized for academics and will be for administrative purchases in the next fiscal year. We prefer "coordination" rather than "control." However, absent compelling justification, we do not give anew Pentium or Power PC to someone who is doing only word processing. We would rather buy new equipment for labs and users who really need the power, and rotate older equipment as appropriate. Of course, there is a cost in staff time to move files and equipment. We still believe central coordination is worth the effort.
**********************************************************************************************************
David Smallen:
All decisions about equipment are made as part of the regular budget process.
What is the scope of the IT budget? It covers all computing equipment.
Do departments make individual purchasing decisions and pay for information services to the IT organization? Departments make their own requests for equipment, but generally seek advice from the IT organization. IT services are not charged back.
How much control over funding should the IT department have? Control is not the right word. Things will work best if there is a better understanding by senior administration about the costs/benefits of controlling the diversity of the IT environment.
**********************************************************************************************************
Ray Phillips:
Funding for IT (computer hardware, software, networks, and many supplies such as laser printer cartridges) is highly centralized at Colby. The only kinds of expenditures that use funds outside of ITS (Information Technology Services) are external grants (NSF, etc) or specially targeted internal research funds (i.e., faculty research grants or funds associated with endowed chairs). We have worked hard to establish a partnership with all departments to ensure that by coordinating the purchase and allocation of IT resources, we add value in delivering reliable overall service, repairs, upgrades, and replacements. No department repays ITS for products or services. Our responsibility is to deliver the products, services, and infrastructure to support communication. Even when purchases are made from other funding sources, the PO is placed through ITS. This is not control, it is providing a valuable service that mitigates problems that users will otherwise experience in the future. There has been enough experience on campus with the alternative to make the vast majority of members of the community convinced that our service is valuable and to actively seek it out when initiating purchases.
**********************************************************************************************************
Question: Are students charged separately for IT services or is it rolled into tuition?
**********************************************************************************************************
David Rotman:
General campus computing is funded out of tuition revenues. Computers in the residence halls are provided out of room charges.
**********************************************************************************************************
Leo Geoffrian:
As a matter of philosophy, the College tries hard to avoid item charges. The sense is that tuition is expensive enough that it's bad business to tack on additional fees for various services.
**********************************************************************************************************
Morris Galloway Jr.:
Rolled into tuition.
**********************************************************************************************************
David Smallen:
Rolled into tuition
**********************************************************************************************************
Ray Phillips:
Students are not charged for IT. These resources are part of what they can expect to have access to when they come to Colby. There are open access clusters and there are more restricted departmental clusters in laboratories. LaserWriters are provided in all clusters for student use and there is no charge for normal use. Students are encouraged to print drafts as well as final copy to aid in the writing process and for other academic projects, but printing of multiple copies is prohibited by policy (and enforced typically by peer pressure).
*********************************************************************************************************
Questions: Do you have a depreciation reserve to replace worn out and obsolete equipment? If so, what is the N years that you expect a component will last?
**********************************************************************************************************
David Rotman:
We don't have a reserve, but are close to having sufficient annual budget to cover all replacements.
We struggle with the lifespan issue. Our current thought is that a 486 DX 2/66 would have a life of 3-4 years and a Pentium would have a life of 4-5 years. One major unknown is the impact of Windows 95 and other software which might (eventually) not run on the 486 processor.
**********************************************************************************************************
Leo Geoffrian:
Such a reserve exists for the "central" services (i.e. the units managed by the Computer Center). The funding is adequate for a 5-year replacement cycle. The College is trying to establish a similar reserve for departmental and office computers but it's not clear where the funds will be found.
**********************************************************************************************************
Morris Galloway Jr.:
No, but we are trying to get one - for all equipment, not just computers. We plan for a 6 year useful life on computers.
*********************************************************************************************************
David Smallen:
Not yet.
**********************************************************************************************************
Ray Phillips:
We are budgeting for the replacement of about 23 percent of the College-owned desktop computers (almost entirely Macintosh) each year. We do not create a depreciation reserve but have established through our standard budget a capital component that will ensure a replacement cycle of about 4.5 years.
We are still seriously considering a 3 year replacement cycle. An objective is to put a reasonably configured Macintosh on a faculty or staff member's desk with the expectation that the machine will be replaced in 3 years (some faculty members will need more frequent replacement). When the machine is replaced, the old machine will be sold (it will be a one user machine). With 750 Macs on campus owned by the College, with our current 4-year replacement plan, we have major logistical problems. If we buy about 170 new computers a year, shift 170 of the existing computers around, and sell 170 oldest computers a year, we have to handle over 500 computer systems each summer and support about 300 faculty and staff going through a hardware transition (moving files between hard drives and reconfiguring for resumption of operations). If we can buy 250 computers a year and only have to support about 200 individuals going through a hardware transition, with their computers being sold and not redistributed, they should experience in this 3 year replacement plan more satisfactory upgrades and less frequent disruption of their work. A few individuals will, obviously need more frequent replacements and some redistribution will probably occur.
We have even looked at a 2-year replacement cycle, a strategy that relies on some significant resale value for the old computers. For us the biggest problem is finding an adequate channel for the disposal of 375 to 450 2-year old computers for which we expect to get a good return. It significantly leverages the academic discounts but because those discounts are now not as significant as they once were relative to what can be experienced through standard retail channels, this will be a very difficult strategy to implement. It is not clear that, for most individual faculty and staff member, replacement of the hardware more than once every 3 years is currently justified. At some point we may find that faculty and staff are moving ahead with information technology so rapidly that a 2-year replacement plan becomes necessary and we will look for a transition to that replacement strategy when it becomes identified as critical to the academic mission of the College.
**********************************************************************************************************
Questions: What do you do to try to match funding resources to need? How well do you understand the needs of each and every department? Do you perform a needs analysis?
**********************************************************************************************************
David Rotman:
We rely on persuasive argumentation from the requesting departments.
**********************************************************************************************************
Leo Geoffrian:
Most of this is evolving rapidly. In the past, we did mostly ad-hoc analysis. Eighteen months ago, the College did a major self-study of its information resources and proposed moving to more formal funding and review processes in order to assure adequate support for the resources. The College is now trying to fill in the details of what this all means.
**********************************************************************************************************
Morris Galloway Jr.:
With fewer than 80 faculty and 70 administrative computer users, needs analysis can be pretty informal. We do meet with users and discuss how they are using their equipment.
**********************************************************************************************************
David Smallen:
Talk with each department annually.
How well do you understand the needs of each and every department? Generally as well as they are willing to discuss them.
Do you perform a needs analysis? Generally nothing that formal.
**********************************************************************************************************
Ray Phillips:
We solicit hardware requests from each department and provide consultation in the development of prioritized lists of replacements and new workstations. The comprehensive requests for all departments are reviewed and allocations made by a group that consists of academic and administrative vice presidents, the faculty chair of the information technology committee, and members of ITS. In general, we have a sense of what kinds of hardware will be added to the College inventory to best manage support and repair costs, as well as plan for future capabilities like CD-ROM drive access and the ability of notebook computers to drive an external color monitor or projection system. Thus, we are providing some kinds of upgrades that are not even being requested. We buy very few different models each year but the number of each model purchased is linked to the budget and the pattern of requests. The allocation process applies the hardware resources in a way that meets the highest priority needs, so it is only indirectly matching funding resources to need. We think we understand each and every department's needs fairly well. We also keep in touch during the year and can usually anticipate pretty well what will be requested annually. The biggest challenge is limiting the expansion in number of workstations, which then must each be supported, maintained, and replaced. We are expecting to add about 20 stations each year for the next few years and reach a stable plateau of 900 workstations. [I once thought a plateau of 500 machines was reasonable but we soared past that.] Requests for new stations require the most convincing justification.
**********************************************************************************************************
Questions: Given your pattern of network development and growth, what do you foresee in terms of staffing changes over the next 3-5 years? Similarly, how do you think your expenditures for "networking" will change in that time period?
*********************************************************************************************************
Leo D. Geoffrion:
We expect staffing and expenditures to rise sharply as we move to bring all of the dorms onto the network (That should triple the size of our net!). After that, it should level off. We know that new networking technologies will eventually lead to a need to upgrade the net, but I expect that this will be slow and gradual, with much backward compatibility. For example, we may someday upgrade the backbone, but that does not make all of the dorm wiring obsolete instantly...
*********************************************************************************************************
David L. Smallen
Since Hamilton is at an early stage of networking, especially high speed networking, I expect our operating expenditures to grow by $150,000 on networking over that period. This will be for staffing (2 people) and maintenance services. I expect that to be the major growth area.
**************************************************************************************************
Ken Pecka
As we plan to complete our campus network to include dorms, we expect staffing to change in two areas: 1) network operations and support (in the way of technicians and network management), and 2) user support services (in the way of ongoing training). We expect that this would equate to 2 FTE by the third year.
Since we have a significant number of buildings and dorms not currently on the network, our network expenditures will be substantial. These expenditures will go beyond the network installation and can be summarized in three categories:
1) network infrastructure - completion of network backbone and building wiring;
2) network components - hubs, routers/bridges, additional disk space, new file servers, etc.;
3) network node (office computers) - upgrades and replacements for the computers in the offices across campus.
*****************************************************************************************************
Bob Hodge:
We have fallen prey to a non-textbook approach of guessing what the market will be and planning to be there when asked. It has actually worked quite well as I am positioned to be the right meetings to listen to visions and evolving programs. Part of my activity is translating those into more understandable objectives rather than asking the end users to do that. Because I am responsible for both long range planning in the university and IT, my meeting times do double duty; its kind of a straight shot from planning to the IT folks.
Supposedly, the staffing needed to manage an integrated network will be less than managing three separate networks. Yet, the growth in usage will far exceed the productivity gains of an integrated system. The number of users will increase which will mean more connections. Most importantly, the hours of active use within a day will increase such that we will need to consider duplicating our management expertise to keep the network fully functional 24 hours a day, seven days a week.
Looking into the mirror dimly, I am guessing that we will double our operational expenses for networking in five years, and will double the funding to connect new users in three years. Hence, my new involvement in fund raising, taking visions to objectives and interpreting them for donors. I see that as an evolving need for CIOs in small colleges.
******************************************************************************************************
Al Essa:
Austin College is planning to install fiber optics this summer to connect all our buildings. We hope to phase in residence hall wiring over the next two years. Our network currently has 200 nodes but should jump to approximately 1000 nodes. I am currently working on an IT plan, including staffing requirements. We are committed to supporting (though not equally) Macintoshes, IBM-Compatibles, and UNIX workstations. Our goal is to provide comprehensive network services from every point on campus.
Staffing: My best guess at the moment is that we will need at least 3 FTEs dedicated to technical and network support for 200 (staff & faculty) and 1200 (students). One of the FTEs should be a network guru with strong expertise in protocols, routers, system administration (Unix, Novell, Apple). The other two FTEs would cover the full range of technical support from hardware repair to network diagnostics.
Expenditures: We do not and will not charge users for network services. Certainly, the amount allocated for networking will need to increase as the network expands. Our budget for IT is currently 2.5 percent of E&G. My guess is that we will need to bring it up to 4 percent. I am optimistic that we will get the increases, especially if we can demonstrate that IT can play a constructive role in enhancing the academic computing environment and in introducing much needed efficiencies in administrative departments and business processes.
**************************************************************************************************
Paul Bishop:
Given the current trend in "right sizing" and the general tightening of budgets we are planning zero growth in staff and hopefully only zero growth in budget dollars for networking.
So what do we do? For starters we are trying to reduce the amount of physical work necessary to maintain the network. As we migrate from twisted pair LocalTalk wiring in buildings to 10baseT Ethernet we are buying smart hubs that allow us to manage them from the comfort of offices in the Computing Center. In addition we are pre-connecting dormitory rooms during the Summer for students requiring network connections in the Fall. To economize we configure the network to support the number of actual users, not the total number of possible users. That is, we have network wiring in every dorm room, office space, and work areas, but only provide sufficient ports to support the number of actual users. As the number of users in a building increases we then allocate funds to purchase another hub to support the additional ports. So far we have been able to keep one step ahead of the demand. Given the financial forecasts for the next 3 years I suspect we will continue to operate with this model.
**********************************************************************************************
Carl Heideman:
I do not foresee any staffing increases. What I do foresee is more members of our department to have roles in supporting aspects of networking. Where our network was once managed entirely by our lead telephone tech and VAX system manager, more than half our staff now has its hands in things. I'm sure this will continue. Currently, our programming staff is learning how they can help support the growth of the network.
We've had a data PBX for years that has cost a fair amount of money for upgrades and service contracts. We are completely self- maintained on our network, and have so far been able to pay for repairs and upgrades (once the network was installed) almost entirely out of budget money that used to be spent on the PBX.
**********************************************************************************************************
Questions - Organizational structure: Describe your organizational structure. Do you have a single management organization which is responsible for all aspects of network development and support? Is the division between academic and administrative computing still viable?
**********************************************************************************************************
David Smallen:
Hamilton has always had one organization that provided computing services to the campus community. That organization is also responsible for voice and data communications. In addition, a very flat organizational structure is used since most problem-solving involves multiple individuals.
**********************************************************************************************************
Bill Doemel:
Computing Services at Wabash College supports telephony, data networking and computing on the campus. Being small, we have a flat organization, all of us being able to gather for a weekly meeting about a single table. On our campus, the library is a separate organization and includes media services. Both the Librarian and the Director of Computer Services officially report to the Dean of the College.
As Director of Computing, I also keep close ties with the Treasurer and the President of the College. We have chosen to keep the combined Computing Services structure associated with the Dean to assure that our primary responsibility is the support of teaching and learning. At the same time, as Director of Computer Services, I maintain a close relationship with the Treasurer of the College and the President. In effect, I report to these three individuals.
**********************************************************************************************************
Morris Galloway Jr.:
Administrative services handles all networking for the campus as well as administrative computing. Academic computing handles labs, faculty computing, and classroom use of computing. The division works, but we cooperate closely on networking and software issues. One day (if building plans go as expected) we hope to be under one roof.
**********************************************************************************************************
David Rotman:
We have one computing organization, responsible for academic computing, administrative computing, and networking. We report to the academic VP.
**********************************************************************************************************
Robert Hodge:
One area is responsible for all kinds of networking - voice, data, and video. This is especially important those technologies converge. With us, video and voice will use the same switch; that kind of obviates the need for integrated management of integrated technologies.
I don't think the division between academic and administrative computing has ever been viable. Just as one can't determine whether a piece of wire in the ground is for data, voice, or video, it cannot be identified as academic or administrative either. There are certainly some different uses of computing between the two, but practically, I don't see a conceptual difference that is worthy of separating the two, let alone a financially practical reason for attempting it. There may be good reason to have some specialists for both groups, but not an organizational or budget separation.
**********************************************************************************************************
Questions - Relative staff size: How many FTEs support information technology (academic computing, administrative computing, network support)?
**********************************************************************************************************
David Smallen:
There are a total of 13 people supporting all the functions listed. We have 1650 FTE students.
**********************************************************************************************************
Bill Doemel:
We have a student body of 830 men, a faculty of about 75, and a staff of about 100. Computer Services now has 7 FTE: Director, Networking Support, Administrative Support (2), Macintosh Support, Computer Repair, Office Manager. After July 1, we assume control of the telephone system and will change one position, gain 2 positions from another office, and add one new position. The person supporting networks will also support the Phone system and will assume responsibility for two operator positions. We have hired a new person to support the DOS-Windows operating system.
We have contracted with a third party, Resicom, to provide support, user management, and long distance billing for our phone system.
We are part of a consortium of about 24 private colleges and seminaries in Indiana which maintains a shared DRA library database on a Digital Alpha machine in Indianapolis (maple.palni.edu). Each institution maintains their own OPAK on this system and users can search holdings within their library or within other member libraries. Communication to this system is through our Internet connection which is provided and maintained by IHETS, another consortium, including both private and public institutions.
**********************************************************************************************************
Morris Galloway Jr.:
Administrative computing: 1 management, 4 staff, 1 clerical including networking for the campus (We outsource some network problem resolution.) Academic computing: 0.5 FTE plus student workers Plus a librarian who handles the library system
**********************************************************************************************************
David Rotman:
We have 10.5 FTE employees in our department (plus about 300 hours/week of student help). Our college has about 2400 students.
**********************************************************************************************************
Robert Hodge:
MIS - 3, Network and hardware maintenance 2, telephone system 2, purchases and maintenance contracts 1, help desk/customer support/training 3, information center for PC consulting/training/installation 3, general support 3.5 which also cover our second campus, system management and operations 2. We use possibly four work study students at any one time for phone system management, programming, and PC work.
**********************************************************************************************************
Questions - Residential computing: Is your institution implementing a "residential computing" program (network connections in the residence halls). If so, what are the staffing challenges (experiences) for carrying out such a program?
**********************************************************************************************************
David Smallen:
We are in the process of wiring the entire campus, including one information outlet per student in residence halls. The wiring will be complete by fall of 1995. We have just gone through a reorganization to focus our efforts more solidly in the training and networking areas for next year. It is clear that we will never have enough staff to meet the demand.
**********************************************************************************************************
Bill Doemel:
We are extending our data network and telephone network to all dormitories and fraternities on the campus. To facilitate this process we have added two students who will work full-time for computer services during the summer preparing documentation and helping to configure the new network hubs. We have also hired a new person to support DOS/Windows operating systems absorbing that responsibility from the Network person. To support the residence halls, we are adding 12 students who will serve as consultants, one for each living unit.
**********************************************************************************************************
David Rotman:
We have college-owned computers in 850 residence hall rooms and will be adding 125 more this summer when two new res. halls are constructed. We have 3 full-time employees dedicated to PC service on campus (850 computers in dorms and 400 other computers), plus we use about 90 hours/week of student help on repairs. The residence hall computing has also increased our work load in answering the telephone, fixing e-mail problems, administering accounts, etc. Residence hall computing has had a significant impact on how we do business as a college--the effort on our part is certainly worth doing!
**********************************************************************************************************
Robert Hodge:
Two groups of faculty and two groups of students have separately suggested that student owned computers and data to the rooms is inevitable but none want it right now, simply because it is clear that the academic program is not prepared to make full use of all of that to make the investment worthwhile. The shortest time estimated for the need is 3 years; the longest is 7.
We have talked with some vendors who would come to campus to sell and configure the computers, allow us to outsource all maintenance, and would handle all financing. We see no reasonable way right now to accommodate that large increase in equipment.
**********************************************************************************************************
Questions - Staffing challenges: What is your single biggest challenge in the area of staffing?
**********************************************************************************************************
David Smallen:
Balancing expectations with what we are capable of delivering. I feel that this is the biggest challenge that all service organizations have to face in times of constrained budgets and rising expectations. And this balancing act must be actively supported by other signals/actions given by senior administrators, else it won't happen.
**********************************************************************************************************
Bill Doemel:
Our biggest challenge is meeting the ever growing needs of Administrative Computing. We are in the third of a four year implementation of Datatel software on our system. Our community of users agreed that they would assume more control of this new system reducing the need for staff at the Computing Center level. The reality has been that we have had to provide much of the support for some offices for a variety of reasons. As is the case with many small institutions staff sizes in all offices are small and there is relatively little turnover. While these staff are very loyal to the institution and very reliable in their jobs, they resist change and often are unprepared for the challenges brought by the "information age." Also at the upper administrative level, we need people who are both competent in their job and knowledgeable about computing.
Let me explain. As a microbiologist, I had to understand and be able to program computers to accomplish my scientific goals. I knew the data, knew what I wanted to do with that data, and therefore knew that I had to control the analyses that were done with a computer. The computer staff supported my use of the computer but I was in control. With the advent of the microcomputer and then networks, I was freed from the limitations imposed by a central computer and central computer organization.
With administrative computing, we want to move the applications from computer services to the offices. However, the people in those offices, while very competent in their present jobs, either don't want to accept or are unable to accept the responsibility for managing and controlling their data. They expect Computer Services to do that for them. In effect, that means that we must not only understand the programming but their jobs as well. Large institutions have addressed this problem by creating new positions and hiring people with computing experience. These new hires are then introduced to the work of that administrative department. In time, I expect that some of these people will then move into administrative roles.
Clearly, then this is a temporary situation. Through changes in education at both the undergraduate and graduate levels, computing applications are being used throughout the curriculum. Students of the humanities and social sciences are learning about their fields and about ways to use computers in their fields. As these people move into the job market, they will bring with them the requisite knowledge of computing needed to accomplish their jobs. Like scientists, they not only will expect distributed computing but will be prepared to use the computers. Until that time, I need to convince our administration and the people who control the hiring of staff that we must work to identify and attract people who bring both a knowledge of the job and knowledge of computing.
**********************************************************************************************************
Morris Galloway Jr.:
Staff training time to keep up with changes.
*********************************************************************************************************
David Rotman:
Trying to keep our staffing level high enough to do the work, but not asking for more than we really need.
**********************************************************************************************************
Robert Hodge:
As the new technologies come which are integrated, the breadth of our involvement is increasing very fast. Compressed digital television on the same pipe as data and phone means we are now in the television business. Overall, if it is computing or communications in any shape, the planning, purchases, implementation and ongoing administration comes from this single place.
All that to say that it is bringing our staff up to speed on some new non-traditional yet converging technologies is our biggest issue. It is not only a training issue, but an issue of the people having a desire and an aptitude to move out of cubicle style computing into new areas which make them much more visible.
SECTION 7: PLANNING FOR NEW ACTIVITIES AND UNEXPECTED CONSEQUENCES
A comprehensive campus network provides opportunities for colleges to do some things better and to do some totally new things. Campus planners should anticipate that a wide range of new activities will develop, things which are now common on campuses which have full networks but which were not possible before the existence of networks. Some of these activities are related to the on-campus dimension of the network, while others are the result of the campus network being connected to the Internet. In addition, unexpected consequences can result from the existence of a campus network as well as the process of creating a campus network. So campus planners need to anticipate both the development of new activities and surprising secondary effects.
New campus activities generated by a campus network
Following are some common service areas and activities which become possible when everyone on campus uses the network.
•Use of electronic mail as a universal communications medium, making campus communications much faster, more flexible, and considerably simpler
•Access to library resources, including the on-line catalog of library materials, from anywhere on campus
•Use of scarce, shared hardware resources (printers, plotters, scanners, CD-ROM servers) in new ways
•Cost effective acquisition of campus wide software licenses, which can promote the use of standards and simplify training and support requirements
•Use of academic and administrative data bases by students, faculty and staff, unrestricted by departmental lines
•Creation of a campus wide information system (nowadays, an Intranet) to provide an common information repository with a standard interface accessible to the whole campus
•Use of electronic discussion groups to supplement classroom teaching, simplify administrative work, and give students a forum for unlimited, on-line discussions
•Development of a system of on line service requests to simplify and improve campus administrative service processes such as requests for transcripts, health center appointments, maintenance department repair calls, etc.
These kinds of resources and services can enhance teaching and learning, allow for much better campus communication, provide expanded access to information resources, break down barriers between campus factions, and knit the campus together in a much more homogeneous communications environment.
Planning for Internet access:
The Internet, first through the wide-spread use of electronic mail and most recently through the amazing explosion of World-Wide Web use, is having a dramatic impact on teaching and learning on college campuses. No educational institution can claim to offer an "information age" education to its students without a campus environment which allows students and faculty to participate fully in this dramatic activity. All aspects of scholarly activity are being rapidly reshaped by the use of hypertextual, multi-media information deposited, organized, and shared on the Web, as well as the incredible volume of electronic communication taking place via electronic mail, news groups, and listserves.
Providing full access for all faculty, students, and staff to the networked world outside the campus is one of the most important reasons to build a campus network. Services and resources such as use of electronic mail, participation in discussion groups, access to bibliographic and full-text materials, data sharing, etc. exist in the global electronic community just as they do on campus. There are several points about planning for Internet use which campus network planners should note.
•Plan for universal access
Everyone (faculty, students, staff) can do their work better if they can use the Internet from their desktop. In fact, many of the benefits, such as the potential of the Web for instructional use or as an "Intranet" CWIS, can only be realized if everyone on campus has access to them.
•Anticipate future upgrades in speed and capacity
At the moment, most Internet activity involves relatively slow transmission of characters or text files. However, the rapid expansion of Web use for graphics, audio, and full-motion video will require substantial network bandwidth (carrying capacity). Campus network planners should anticipate an ever-growing demand for faster network speeds and greater information carrying capacity.
•Plan to support users with documentation and training
As with any area of technology, people need help learning how to navigate the Internet. The navigation metaphor is a good one, since the Internet is more like a three-dimensional, trackless ocean than a linear highway. Locating information in this state of cyber-anarchy is very difficult. In particular, faculty and students will need help in order to capitalize on the myriad educational opportunities the Internet affords.
•Expect a range of new policy issues related to Internet use
Many issues related to information access and use are being raised in the context of Internet access. Among the most challenging of these are issues related to fair use, intellectual property rights, plagiarism, and access to Internet materials which some people consider pornographic. Campuses, as well as the general public, will have to deal with these issues in a variety of contexts as society deals with the impact of the Internet.
•Plan on being an information provider as well as a consumer
Faculty, staff, and students at colleges might be just consumers of information from the Internet, or "lurkers" on discussion groups. But the egalitarianism of the Internet makes it possible for anyone to participate in discussions, make documents available via a Web page, or publish and distribute an electronic journal. Besides the challenges of providing a working environment which allows everyone on campus access to Web browsers and Web development tools, campus leaders are dealing with a range of new issues related to the kind of information which is being provided to the Internet world on personal Web pages via college servers.
Planning for unexpected results:
Networking has revolutionized the way we work, teach, learn, communicate, and think. However, our campus networking strategies often have unexpected, secondary effects which are not necessarily positive. These tend to be effects which develop in an evolutionary way and it may be hard to notice them. But they will happen and should be anticipated by campus network planners.
A secondary effect is simply an unintended consequence, something unplanned which occurs as a result of a planned activity. For example, one college installed a Maintenance department work order request system as part of an on-line service request menu. Basically, this system just harnessed the e-mail system with some batch files to allow college administrators and faculty to request repair and maintenance work without making a phone call or sending a paper form. The Maintenance staff managers then assigned the work to the trades people via e-mail and tracked the work orders in a database. The system was expected to save time for college staff and the Maintenance department, and it did. What was unexpected was that students almost immediately started using the system to report maintenance problems in the residence areas directly to the Maintenance department. Historically, all such problems were first reported to the Housing staff, who in turn sent the requests on to Maintenance. Once the students started using the network-based system, a whole step was eliminated, which speeded up the process significantly. And the Housing staff suddenly discovered that they had extra time. All in all, this was very positive secondary effect. Not all unintended consequences are so beneficial. Following are a few of the kinds of secondary effects to anticipate in planning a campus network.
•Academic challenges
Wide use of a campus network by students and faculty will raise challenges to traditional modes of instruction and assessment. Students and teachers will experiment with new modes of instruction and communication. Web-based classes, group projects carried out over the network, distance learning activities, round-the-clock e-mail discussions, and distribution of computer-based courseware across the campus all raise new issues about the nature and management of the educational process. Faculty will be challenged to find new ways to evaluate student papers created in a network environment where traditional approaches to recognizing intellectual privacy rights are very difficult to assess. College administrators will be challenged to find new ways to measure the performance and competence of faculty who experiment with new approaches to pedagogy over the Internet.
Many colleges do not have adequate classroom facilities to take advantage of the instructional opportunities available in a networked teaching environment. Classrooms, laboratories, studios, and study areas all need network access. Most colleges will need some classrooms which have networked computers at every student desk, along with display and control equipment to allow the instructor to provide a visual instructional focus in the room. Requests/demands for new instructional facilities which allow for use of new technologies. Building new types of classrooms will be a very expensive proposition on most campuses, but there will be strongly felt needs for such facilities when faculty and students begin to understand the value of teaching and learning on a networked campus.
•Continually Rising Expectations
It is almost axiomatic that demand will exceed supply when it comes to technology on campus. With computing technology constantly getting faster, smaller, and more powerful, and with new network applications developing constantly, faculty, staff, and students will expect and demand a continually "improving" technology context on campus. Network administrators can expect constant demands for more network speed and bandwidth as faculty and administrators want to move very large files - and don't want any delays. And members of the campus community expect that the network will be available with perfect reliability and constant access. Now that huge numbers of people have Internet e-mail access, colleges are often asked to provide such services on campus as well as off campus to parents, alumni, community members, high school students, and neighboring colleges.
•Changes in work styles and job classifications
Access to a full campus network quickly leads creative workers to find new ways to do their jobs by ignoring the old hierarchical styles of information flow and management. People simply go where the information is and find people to work with who will help them do their job. So the availability of a campus network makes possible work styles and problem solutions based on new people networks. To promote this type of creative work style, some staff members will need training and encouragement to develop more creative approaches to problems. Other workers will eagerly take on more comprehensive tasks outside their normal departmental areas, thus raising difficulty questions about how to classify such jobs in the kind of position classification systems used at colleges.
Access to the Internet from all staff work areas raises a number of issues related to employee supervision and assessment. Should employees be encouraged to "cruise" the Internet to find information, make contacts, and participate in discussion groups? How can managers evaluate the productive value of the time an employee spends learning to use the World-Wide Web?
Existence of a full campus network will lead to the need for new kinds of positions. Besides the obvious ones related to network management, the development of campus-wide databases will require someone to act as a database administrator, with sufficient campus-wide authority to support centralized management of information. It is unclear where such positions should reside in the campus employment structure, but the people holding such positions need to be comprehensive thinkers with a genuine collegiate perspective.
•Changes in communication patterns
A campus network will very likely have a strongly positive impact on campus communication patterns, including more, better, and new forms of communication among all campus constituencies. But it is also quite possible that the most immediate consequence of a campus network will be that the traditional communication patterns are completely upset. Suddenly, any student can send e-mail to the President without intervention from a secretary or any other traditional filter. And anyone, no matter how eccentric or altered with recreational chemicals, can send e-mail to everyone on campus, on any bizarre subject whatever, at any time of the day or night. Such incidents quickly raise the campus consciousness about the power of e-mail, and typically, give the campus a strong imperative to begin deal with important policy issues related to student behavior, privacy, harassment, security, residence life, authority, and administrative control in the networked environment.
SECTION 8: CONCLUDING THOUGHTS
The electronic roundtable which provided a significant portion of both the quantity and quality of information in this document was a unique way to discuss ideas and gather information on a very important topic. The roundtable participants shared a wide assortment of observations, personal experiences, ideas, suggestions, and opinions. Anyone planning a campus network should find much to think about from the material provided by these highly experienced campus leaders. For those stalwart readers who have not yet had enough, this concluding section is a summary of some of the points the author feels most important to stress from his experience in developing, managing, and observing networked organizational environments.
•A full-campus network is one of the most important features of a college environment. While a fundamental part of the infrastructure, networks are not static; they must change and grow continually, often at a much different pace than other parts of the infrastructure. A network is no longer an option for any college which claims to prepare students to be life-long learners and good citizens in the Information Age. In the very near future, every student will own a network-capable computer and will expect access to the networked information world. The World-Wide Web, one manifestation of the global information network, has the potential to dramatically change teaching and learning, much more so than just access to computing. So a campus network is absolutely essential.
•Building a campus network is a profound act of change on a campus. In the broad sense, creating a campus network is really a massive organizational development project. Colleges often don’t realize that developing a network will fundamentally alter the way the members of the campus community teach, learn, work, and play.
•People who build campus networks have to be both "scouts and wagonmasters." That is, they have to figure out the appropriate technological options for the institution, help the rest of the community understand the options and choose appropriate ones, and then perform the myriad challenging tasks of helping the campus move toward the chosen goals. Campus technology managers are fundamentally change agents, and as such, will not always be popular. It is not possible to do a good job and be liked by everyone.
•Colleges which aspire to have a high-quality technology environment need to set standards and manage networks from a true collegiate perspective, in the sense of choosing to maintain a set of technological options which are genuinely supportable with the financial and human capital available. It is surprising that in a time of increasingly constrained resources and continually rising support demands, many colleges still do not have the institutional will to make decisions which are best for the institution at the risk of offending a small group of vocal employees or students. For example, many colleges are trying to build full campus networks by hooking together nearly incompatible, departmental local area networks which were developed in an earlier era of entrepreneurial anarchy. It is disturbing to see how much time, effort, and money colleges spend trying to cobble together and support very dissimilar systems which exist because the original decisions were made on the basis of parochialism without assessing the real cost to the institution.
•Many colleges have not really come to grips with the funding implications of campus networks. A campus network, while a part of the infrastructure, also needs to be viewed as a "consumable" like library books, football uniforms, dormitory furniture, electricity, water, and toilet paper. This requires some on-going, consistent, predictable source of funding, not just some one-time costs to lay cable. Interestingly, many colleges which typically charge only one (often very large) comprehensive fee to students levy a special "technology fee" to finance computing and networking, thus verifying that technology is not viewed in the same light as other essential services or resources.
•Without a doubt, people are the key factor in whether or not a college will be able to develop an exemplary network environment. Strong leadership and support from senior executives is an absolutely key ingredient. Colleges need pragmatic visionaries to imagine and then implement the changes necessary to build a networked institution, and thus need a structure which allows leadership to exist and flourish. Sadly, colleges often cannot capitalize on the full potential of networking because they are unwilling or unable to provide staff to support it. Colleges which have high-quality staff have the best chance of having a high-quality information technology environment.
|
The purpose of this project is to use a mediated, electronic roundtable discussion to collect ideas about the development and management of campus networks at small colleges, with the expectation that the ideas generated by this roundtable will eventually be made available to CAUSE members in some appropriate format. Although all of us are familiar with Internet listservers, the electronic roundtable discussion has some features which may be a new mode of operation for us. The basic format for this roundtable is based on ideas developed by Ken Blythe, Pennsylvania State University. The following guidelines or "rules" for the discussion are based on Ken's recommendations about how roundtables of this sort can be most effective. The roundtable discussion will be carried out on a special listserver set up by the CAUSE staff. The discussion will proceed by rounds, with each round focused on a topic in a structured outline, possibly branching into related sub-topics if that seems fruitful. Each round, and the discussion in total, will continue for a fixed time period, after which the ideas generated will be synthesized and edited by the roundtable mediator. It is the responsibility of the mediator to program, monitor, and summarize the discussion. The mediator will decide the number, content, and length of each round of the discussion. The mediator may participate in the discussion, and may intervene to modify the direction of the discussion of that seems necessary. For each round, the mediator will provide a short declarative or interrogative statement to provoke discussion, and will set a time limit on that round. Participants should try to give responses which are short, consequential, and interesting. As the discussion proceeds, participants may send as many messages as they wish to address either the main topic of the round or the responses from other participants. The time limit on each round will be short, probably one or two weeks. However, if the discussion takes a particularly interesting twist, the mediator may amend both the topic and the time limit. The point of the time limit is to maintain the momentum of the discussion, not to stop it if it is proceeding well. The discussion will end when all the topics of the outline have been addressed. The mediator will then prepare a summary of the discussion and circulate it to the participants to obtain their confirmation that the summary is accurate. The ideas generated by the discussion may eventually be published by CAUSE. The roundtable discussion will be most effective if participants adhere to the basic courtesies of polite discourse. Participants should identify themselves by name and institution on all messages they initiate. Each message should have a clear, descriptive subject line. Foul language and flaming are prohibited. Each participant should feel free to enter the discussion at any time. Jargon, acronyms, and symbols which might be unclear should be defined. |