DSS Architecture and IS/IT Infrastructure
Many academics discuss building Decision Support Systems in terms of four major components - the user interface, a database, models and analytical tools, and the DSS architecture and network (see Figure 6.1). One can label these components collectively as the overall architecture of a DSS. This traditional view of DSS components remains useful because it identifies commonalties between different types of DSS, but it provides only an initial perspective for understanding DSS architectures.
Figure 6.1. DSS components.
As noted previously, a major component in the design of a DSS is the user interface. The tools for building the user interface are sometimes termed DSS generators, query and reporting tools, and front-end development packages. DSS user interfaces can be distributed to clients in a "thick-client" architecture or delivered over a network using Web pages or Java applets in a "thin-client" architecture. A thin-client architecture where a user interacts using a web browser has many advantages, but until recently the sophistication of the user interface was limited compared to a thick-client architecture where a program resides on a DSS userís computer.
A DSS database is a collection of data organized for easy access and analysis. Large databases in enterprise-wide DSS are often called data warehouses or data marts. Document or unstructured data is stored differently than structured data. Web servers provide a powerful platform for unstructured data and documents. The architecture for a structured DSS database for a Data-Driven DSS often involves multiple servers, specialized hardware and in some cases both multidimensional and relational database software. The extraction, transformation, loading and indexing of structured DSS data is a black art, and there are as many data engineering strategies as there are data warehouses.
The DSS architecture and network component refers to how hardware is organized, how software and data are distributed in the system and how components of the DSS are integrated and physically connected. A major issue today is whether DSS should only be available using thin-client technology on a company intranet or available on the Global Internet. Scalability is also an important DSS issue. Scalability refers to the ability to "scale" hardware and software to support larger or smaller volumes of data and more or fewer users. Practical scalability is the ability to increase or decrease size or capability of a DSS in cost-effective increments.
The DSS framework discussed in Chapter 1 showed the different emphases that are placed on DSS components when specific types of DSS are actually constructed. Architecture, networking and security issues vary for Data-Driven, Document-Driven DSS, Model-Driven and Suggestion DSS. Multi-participant systems like Group and Inter-Organizational DSS rely heavily on network technologies. The architecture of a Data-Driven DSS emphasizes database performance and scalability. Most Model-Driven DSS architectures store the model software on a server and distribute the user interface software to clients. Networking issues create challenges for many types of DSS but especially for a geographically distributed, multi-participant DSS. Table 6.1 identifies some of the architecture requirements for different categories of DSS.
Table 6.1. DSS Framework and Architecture Issues.
An architecture for any information system is a formal definition of its elements or parts. A DSS/IS/IT architecture can be diagrammed in terms of four layers: the business process map, the systems architecture, the technical architecture, and a product delivery architecture. The business process architecture shows how tasks are completed. The systems architecture shows the software components. The technical architecture focuses on hardware, protocols and networking. The product delivery architecture focuses on outputs of the system.