Penn Computing
Computing Menu Computing A-Z
Computing Home Information Systems & Computing Penn

Decision Support Environment Implementation Model

The following describes the implementation model depicted in Figure 1.


Our current operational environment, which includes both centrally and locally managed systems, will serve as the primary source of data for the DSE. In this current operational environment, data are stored primarily by system in formats that often inhibit easy access and promote data redundancy.


In addition to University operational data, data from external sources as needed will be extracted for use in the DSE. As with University operational data, external data will be defined in the Repository (D) and processed through the filter (see C below) before entering DSE.


The data will go through a filter process which will select, extract, and transform the data, then aggregate them to the level needed by end users. The need to transform the data comes as a result of the seperate systematic storage of the operational data. For example, "Departmental" is represented by different sets of codes within the operational systems.


The Repository is the mechanism for defining information about University data, the applications which use them, and associated business rules and models. It will provide storage and access for these type of information and will act as a directory to the data for end users as well as systems personnel. The filter will use the definitions, models, business and transformation rules to prepare the operational data from DSE.
There will be standards and procedures for populating the Repository so that the information is consistent and complete. An application for accessing and populating the Repository will be developed and made available to a wide population within the University.


Once the data have passed through the filter, they will be loaded into the DSE Query database. Here data exist at various levels of aggregation and granularity. At the lowest level, Current Detail, are those data which have come from day-to-day operations of the enterprise: business transaction data such as grant proposal or contract award data. Depending on access requirements, the DSE may contain two or three years of Current Detail data at any one time.

Historical Detail are data which are no longer current but are still important to decision making at the University, such as job history. Data are provided at this level for accessibility and integration which does not exist in our current operational system environment. Requirements for historical data will be determined for each subject area. Initial loading of the subject area to the DSE will include the necessary Historical Detail collection at scheduled intervals. Access requirements will dictate how long Historical Detail needs to remain available on line.

At the next level of aggregation, Lightly Summarized, data has been summarized upward from Current and Historical Detail. Examples of these data are monthly departmental budget summaries, census data, and departmental teaching analysis. Again, depending on access requirements, the DSE might contain five year's worth of Lightly Summarized data.

Highly Summarized contains data at the highest level of aggregation. These data would provide information summaries at the Responsibility Center or Enterprise level across several years. An example might be, grants in support of the AIDS program over the last five years, or a five year view of the University's "Fact Book". Currently, information in the summarized form is frequently provided by school or departmental programmers. These programmers could continue to create these data, using DSE's Current Detail, Historical Detail, and Lightly Summarized collection with their results. Once in the query database, end users with appropriate tools could access, retrieve, and report this information without the support of programmers.

Data in the DSE will be refreshed based upon users needs for currency in combination with the availability of data in the current operational systems. The data in the DSE's query database are restricted to query access only and are never updated directly. The DSE environment will provide appropriate technologies to support security requirements and provide privacy needs while still providing access to the data.

The initial pilot for the DSE will be deployed on a single platform and will contain data from systems which are are not currently slated for immediate redevelopment. Over time, however, the DSE will contain all relevant administrative data and may span multiple platforms in a configuration to support the data architecture and organized to optimize data access.


There needs to be a variety of Query and Analysis tools available to support the needs of different users of the DSE (see chart below).

User Profile Access Objectives
Position or Role Simple Query Complex Query Statistical Analysis & Modeling Decision Support/EIS
Required Level of Technical Skill
low or
low to
medium to
Departmental personnel X X

Institutional Researchers;
Super Users Group;
Decision Support Analysts;
Systems Support Staff


Executive Managers;
Strategic Planners


The DSE needs to provide flexible and powerful tools for the complex queries as well as statistical analysis and modeling. Some tools, like SAS, may be located on servers rather than the desktop. Desktop tools such as GQL will be needed for simple and complex queries by users with little technical expertise. These desktop tools are easy to customize, provide cross platform support, and allow access to data without required foreknowledge of physical data location. Formal EIS applications will be built to provide easy access to highly customized and summarized data for modeling, manipulating and reporting by University executives. All of these tools need to have some level of flexibility in the retrieval and presentation of data. In addition, data must be portable to other desktop toold such as spreadsheets and word processors.


Data may be moved from the DSE to the Desktop otr other local servers where it may be combined with locally created data to form a Local Data Collection. There will be minimum standards for desktop hardware and software supported by the DSE. As mentioned previously, some of these Local Data Collections, particulary those with Center-or enterprise-wide significance, may be transferred from the desktop via a special filter to the DSE's Highly Summarized collection where they will be available to a much wider audience.


In the future, the DSS will be a logical base of data, providing data from the Future Operational Environment as well as translated data from the Current Operational Systems (A) and the External World (B). In some cases to satisfy a information need, data may need to be queried from both the Query Database and the Future Operational Environment. Though the data in the Future Operational Environment will be based upon the same definition, rules, and models as the Query Database, there may be a need to translate some of the data in the Future Operational.

Figure 1

[ Data Administration ]


Information Systems and Computing
University of Pennsylvania
Comments & Questions

University of Pennsylvania Penn Computing University of Pennsylvania Information Systems & Computing (ISC)
Information Systems and Computing, University of Pennsylvania