To face demands of competitivity, reactivity and efficiency, enterprises need constantly evolving information systems.

Technology now provides a wide variety of possibilities for building complex information system architectures. The increasing need for communication between applications, particularly using Internet, requires a more comprehensive and detailed consideration of how software resources are organized.

By representing this information in a detailed or general diagram, modeling helps to design and describe the architecture of information systems, facilitating any upgrades and enhancements.

There are many reasons for modeling an information system architecture. The modeler can describe the current system, focusing on the software architecture and how it is integrated within the organization, or on the technical infrastructure and IS system hardware. The modeler can also analyze which upgrades are required in an information system, whether this involves installing a new application, a new Internet site or a software package such as SAP, or optimizing a Client/Server architecture by finding how best to distribute data and processing.

The architect can also precisely represent the interactions between the applications of the enterprise and those of its partners, or between autonomous entities within the same enterprise as for example on installation of a Workflow engine.

In contrast to simple graphical representations or documents, modeling allows progressive building of a complete and consistent repository of information that is easy to maintain.

To describe the architecture of a complex information system, the first step often consists of creating geographical overviews that show the primary sites of the company and how they communicate.

Depending on the level of detail, the modeling may show the entire information system in general, or may map more specific subsets.

An information system is built in stages. Generally, enterprises have developed or purchased their applications when needed, when the technology or financial resources became available, etc.

Their accounting activities having been computerized, companies installed specialized tools focusing on their line of business. Then management tools arrived on the scene and we now see rapid growth in applications benefiting from Internet that combine the consultation and update technologies of geographically distant databases.

Now there is rapid growth in communication applications. All these applications developed gradually and sometimes independently of each other, but they now need to inter-communicate, share data, synchronize processing and have uniform interfaces.

Faced with these challenges, the enterprise that wants to control and manage the expansion of its information system has to have an inventory of its applications, of Internet/Extranet sites and the associated databases, and be able to locate where they are installed. It must also be able to define the functional and organizational scope of its applications (functions and users affected, information exchanged), and any interdependencies.

In this context, mapping of applications, Internet sites and databases becomes extremely beneficial to the enterprise.

The analysis of current hardware and software that mapping entails will reveal weak points in the information system: shortfalls, competing applications, data entered more than once, lack of uniformity. The enterprise can then decide how to fix these problems and what activities it can institute to improve organization and productivity.

The resulting maps are also a powerful tool when the information system is to be upgraded: they can be used to analyze the impact of a new application or software package on its future environment, formally specify EDI flows, or optimize a Client/Server architecture.

Enterprise information systems develop with wide differences over time to suit projects. Under these conditions, it becomes increasingly difficult to integrate new applications into existing information systems and to handle technological evolutions. Information system city planning consists of defining the different components of an information system and defining their assembly terms so as to facilitate integration of these new applications.

In the same way as for a city, you can design the enterprise information system as for a city planning project and define a durable splitting of the information system into independent districts.

The advantages of breaking down an application into components can be felt throughout the life cycle of the application. This approach is essential in the specification phase. The different requirements to be covered are then studied in sub-projects.

Defining the functional components maximizes the chance of successfully integrating a new application into the enterprise. During this crucial phase in the application life cycle, it is helpful to assign the different application functionalities to various org-units within the enterprise, in order to study their impact on the organization and set up any required training.

Also, representing application functional components facilitates their reuse.

Interactions defined at an international level by organizations such as the OAG (Open Applications Group) or IFX (Interactive Financial Exchange), as well as the description of associated XML schemas (available with MEGA System Blueprint) enable automation of exchanges between the applications of different enterprises. The use of interactions and standard messages enables these enterprises to communicate without prior knowledge of each other. This enables operation of market places, which are expanding rapidly. The same mechanisms can be used within an enterprise, as for example on installation of certain workflow engines, or to automate exchanges between an enterprise and its subsidiaries.

For more details, see "Describing Interactions", page 49.

Client/Server architectures are now found everywhere. This approach offers many possibilities in terms of reusing basic software components, portability and extensibility, so that information systems can be efficient and flexible no matter how complex the requirements.

However, not all application architectures are equivalent in terms of the information transported, time, risks, and costs.

Modeling helps answer these questions by defining the Client/Server architectures and breaking down the applications into client or server modules. Then the modeler can carefully consider where to locate the data and how to optimize its distribution and processing, based on a qualitative evaluation of the volume of information generated by the applications.

This means several different scenarios can be examined while still in the design phase, simulating alternative solutions of where databases are installed and replicated, or how application resources are shared.

Enterprise information systems often combine legacy hardware and new equipment. There are often several generations of the same type of hardware coexisting in an information system, which only serves to increase complexity.

Modeling the technical aspects of the information system architecture often contributes significantly to the knowledge and management capabilities of system architects and managers.

The diagrams produced identify and represent existing hardware (servers, workstations, etc.) and the networks linking them, with descriptions of their characteristics. These diagrams indicate the machines hosting databases and applications.