[<-- Previous] [Home] [Contents] [Next -->]

A reliable baseline information infrastructure

We have identified one of the key characteristics of the current crisis as the ever-increasing disparity between supply and demand. We assume that in the near future, there will be no significant increase in support resources. The most important contribution we can make to reducing inefficiencies is to establish a highly reliable information infrastructure.

The infrastructure should be defined by user needs, be intuitive and consistent, and require minimal effort to learn and use. It must enable us to do simple things simply and do repetitive things efficiently. Most important, the infrastructure should be available when needed, be engineered not to break, be vigilantly monitored, be quickly repaired when things go wrong, and have minimal resource requirements. It should create a base for:

• the development of solutions for more complex problems
• the exploration of new applications
• the exploration of new technologies

We think that current technology is capable of providing a reliable baseline infrastructure to meet about 80 percent of the needs of 80 percent of the users. The software for this environment includes one of the office suites, a robust e-mail package, and a Web browser. The hardware for this environment must run the software with reasonable performance, provide a consistent user interface, and be readily available.

A tidy technology solution would allow a campus to support a single hardware platform and a single software suite. Such an environment would meet the 80 percent criteria, but it is arguably impractical and undesirable for most campuses to implement. From the perspective of what the user sees, however, the differences between Macintosh and Wintel machines have become relatively minor. Similarly, all major applications programs are converging on a similar graphic interface, a common set of functions and functionality, and good import/export capabilities. We can achieve our goals in ways acceptable to most users if we are careful and disciplined in how we design and configure our environments.

Most campuses have specified, or have at least tried to specify, a standard computer environment. These standards have lessened the support problem to a small degree, but they certainly have not solved it. The failure of these standards to resolve the support problems reflects a segmented approach to the support crisis. When developing and implementing a standard information infrastructure, we should consider the following critical success factors.

The infrastructure must be designed

Our current environments have evolved through a process that has yielded satisfactory results but is far too inefficient for the 1990s. An infrastructure design must start with an information architecture that describes the information needs, databases, and information processes common to the entire institution.

The architecture expresses the information and information processes that are most important to the institution in terms that we can translate into technology. It helps us maintain a holistic perspective, so that we do not over-solve some problems and overlook others. Even more important, it prevents us from investing too heavily in technology that is interesting but that has little relevance to real problems. The architecture also provides a set of guidelines for interconnecting environments that address the unique needs of departments and individuals to environments that address institution-wide concerns. For example, a researcher who can add 14th-century characters to a keyboard map and include a 14th-century dictionary in a word processor need not support a unique computing environment to conduct scholarly work.

Architecture designs guide our decisions about which technologies are needed and what they need to do. No longer is it sufficient to install a piece of software on the server and then fix the problems when the customers point them out, or to change the network configuration and reconnect people who were disconnected. To achieve the level of reliability expected of the contemporary environment, we must fix problems without breaking our systems. To do that, our systems must be well designed and monitored vigilantly.

IT staff must manage environments, not fight fires

Providing a reliable infrastructure that continues to meet institutional needs implies changes for our IT organizations. In the past, we rewarded our firefighters. It was okay to break something as long as you could fix it before anyone noticed. We rarely had the time to do things right, but we always seemed to find the time to do them again. Early users of information technology understood this culture and learned to deal with it. Today's mainstream user will not tolerate unreliability. Many IT support providers have the appropriate attitude and aptitude to find a problem and fix it once it occurs. But now we need staff who can understand a complex system well enough to ferret out its weak points and fix them before service is affected. Such highly knowledgeable staff will function as infrastructure managers, responsible for fireproofing rather than firefighting-a role shift that is key to success in the new information technology support model.

The infrastructure must be hierarchical

Rather than being flat, the information infrastructure must be hierarchical: an individual's environment is built on top of the departmental environment, which in turn lives atop the institutional environment. (See Figure 1 for a graphic illustration of this model.) The department can create its own environment because it does not have to spend all of its time supporting the basic technology, and the individual faculty member, administrator, or researcher then builds upon both bases. The architecture and design of the infrastructure ensure that these layers interconnect effectively and efficiently. This part of the model also provides a way for the campus to deal with the rapid changes in the technology as well as the ever-increasing expectations of our customers.

It is our nature to explore new applications and push the envelope of technology. If these efforts occur in a common environment, they can be more easily made available to others in the institution. Thus we provide for a migration of new and useful applications from the individual, to the department, to the institution. The electronic versions of the 14th-century manuscripts become available to everyone in the Spanish department. The efforts of the Spanish department to develop Spanish word processing becomes part of the foreign language word processing module in the institutional infrastructure. The mechanism that allows the infrastructure to evolve quickly, and in harmony with user needs, is another critical success factor for the standard environment.

The new architecture must be implemented with support from users

Implementing the standard information infrastructure is somewhat of a chicken-and-egg problem. Customers are reluctant to give up idiosyncratic environments when they do not see the advantage of a standard infrastructure, and it is difficult to implement a new architecture if the customers are not clamoring for it. We think there may also be some reluctance on the part of the IT organization to fully commit to an environment that it does not solely define, develop, and manage. Standard environments, however, can actually deliver the true potential of information technology. A few examples include:

• Return-to-service. When a desktop computer or server breaks, the most urgent need is to return its functionality to the user. If an OC cannot fix the computer using diagnostic tools within 15 minutes, s/he should be able to replace it from a swap-out pool. With standard environments, a small investment in such a pool provides a great advantage.

   

• Personnel turnover. If all its administrative information is in a standard environment, a department will be relatively unaffected by personnel turnover. The new secretary will know how to get the letters out and send notices to all faculty, and the new chair will know where to find last year's annual report and cut and paste from it to generate this year's report.

   

• Classroom presentation. When classroom presentation technology is standardized, faculty are confident that they can teach in any classroom.

   

• "Seat" capacity. If incoming students are given a description of systems that will work in the college or university's information environment, they can make good purchasing decisions and have less need to use the seats in public computing facilities. This enables an institution to minimize its acquisition of machines that quickly become obsolete.

   

• "Critical mass" support. With a common environment, one is likely to get a question answered by the person in the next office, in the hall, or at the lunch table. Departmental and central support staff thus have more time to add reliable and robust features to the environment.

[<-- Previous] [Home] [Contents] [Next -->]