Penn State University
Penn State University, with 24 campuses and 80,000 students, wanted to make its huge storehouse of information readily available to students, faculty, and staff. First-generation web applications were mainly designed as stand-alone systems, and simply combining legacy mainframe systems with client-server functions quickly crashed key business applications during heavy usage.
Penn State undertook a massive internet development project designed to give students and staff direct access to its wealth of institutional information. Dubbed eLion, the new system would deliver a custom search engine, a hypertext listing of suggested academic and advising references, open access to student information systems, an artificial intelligence-based advising service, and a variety of support services for faculty, staff and students.
“Accommodating the web is meaningless if you can’t preserve existing application and database investments. We had been building mainframe systems for 20 years. As we create client-server and intranet applications, our philosophy is to transition and improve the capabilities of those mainframe systems, not replace them.” – Peter deVries, Director of Advanced Technology, Penn State University
Unfortunately, PSU’s first solution, a two-tiered system, was unsuccessful. The mainframe was so overburdened that key business applications halted during heavy usage. PSU’s IT professionals realized they needed to off-load much of the processing requirements from the legacy system if things were to get back on track. Many of the data requests were being made by students on their own behalf, such as course information, grades, and transcripts. Ken Blythe, Penn State’s Senior Director for Aministrative Systems, recalled the staff’s deliberations:
“So we wondered, what if the university could set up a system that would let the students help themselves to this information in a secure fashion? It would save on administrative expenses, speed information response and take a tremendous load off the university staff.”
A three-tiered system thus began to take shape. The mainframe would remain as the data server. Web servers would host the primary application logic. And the presentation layer would be a combination of traditional desktop clients and thin clients running web browsers.
Cincom Smalltalk’s object-oriented design is inherently suited to the web’s distributed makeup. Rather than storing data and business functions separately, Smalltalk objects comprise both procedures and data, so that they can be combined more flexibly than data or procedures taken separately, and objects can be readily shared among applications.
“Cincom Smalltalk is the first product that puts a full-blown, object-oriented development environment behind the web page. It enabled us to put our client-server application on the web with very little effort,” – Ken Blythe
“Cincom Smalltalk’s object-oriented architecture enabled rapid development and provided us with the opportunity to continually expand the system through component reuse.” – Peter deVries
- To read the full story, click here.