Ziv Baida: Architecture Visualization

The last phase in my studies for a Masters degree in computer science was a research about architecture visualization.
The goal of the research project is investigating how architectures can be visualized to support the decision-making process of business managers. The project focuses on three levels of architectures: enterprise architecture, domain architecture and system architecture.
It is important to understand the following:

  1. The research does not refer to architectures of the "physical world" (architectures of buildings and cities), but to architectures, as known in the IT-world. Those refer to (information)-systems in their environment: the business.

  2. Visualization is used to support the decision-making process of business managers (CEO, CIO, managers of business units etc). We discuss only what serves the decision-making process of business managers.

  3. Visualization is stakeholder-specific. A CEO requires different information than an IT-manager, and the information has to be presented in a different way.

My MSc. Thesis

The thesis which summarizes the research is available here. To download the thesis, right-click here, and then choose the SAVE AS option (3.6 MB).
The thesis is available as a PDF file. You need the Adobe Acrobat Reader software to read it. If you do not yet have this software installed on your computer, you can download it for free from www.adobe.com.

Management Summary

Following is the Management Summary of my thesis.

The goal of this research project is investigating how architectures can be visualized to support the decision-making process of business managers. The project focuses on three levels of architectures: enterprise architecture, domain architecture and system architecture.
This document presents the planning of a communication activity, in which an architect wants to communicate with a business manager. Visualization is presented as a tool in this process. It is preceded by, among others, a personality analysis of business managers and a contents analysis in which the contents of the message to be conveyed are defined. The contents analysis is guided by the principle that it's not necessary to visualize the whole architecture for a certain stakeholder; it's enough to visualize those aspects that this stakeholder considers important. The aspects we identified as most important for business managers are: costs and benefits, effectiveness, facilitating change, schedule, feasibility and risk analysis and integration. The latter proved to be interpreted differently by business managers and by IT-experts. Business managers consider this term to refer to the integration of organizations, whereas IT-experts refer to the integration of information systems.
We present a set of fourteen guidelines, with which visualizations should comply.
A major conclusion of this research is that before the architect introduces some architecture to business managers, he must ensure that the business managers have a powerful enough mental model of architectures. The term 'mental model of architectures' refers to the internal representation that a person has of architectures. Research on the field of management theory yielded the conclusion that managers require powerful mental models of architectures in order to make decisions regarding architectures. From interviews with IT-experts we concluded that managers often lack this mental model. Creating a mental model of architectures is therefore required in order to ensure that a manager can make decisions about architectures. For this reason, we discuss visualizations for the creation of a mental model of architectures first; only then do we discuss architecture visualizations.
The following principles are defined as main criteria for good visualizations. These are important conclusions of our research.

  1. Visualizations should support verbal communication between architects and business managers.

  2. Visualizations must support switching between the architect's profession and the business manager's profession. The ability to switch back and forth between these professions is identified as a key factor in winning the managers' trust, and must therefore be supported by the visualization.

  3. Visualizations should demonstrate that the architect speaks the business manager's language, since we suggest visualization as a tool to bridge the gap between architects and business managers.

  4. Architecture visualizations must be dynamic. Dynamics are required to support main characteristics of architecture. It also helps keeping the attention of the audience focused on the screen.

  5. Architecture visualizations must be related to the mental model visualization. If business managers cannot relate both visualizations to each other, they cannot use their mental model to argue about architectures, and to make decisions. This would mean that their mental model of architectures is not powerful enough.

We created two visualizations: the first one establishes a mental model of architectures, the second one is an example of how architectures can be visualized, complying with the guidelines we provide. These two visualizations are presented and discussed in this document.

My Presentations

Two presentations are available:

  1. A presentation that includes two visualizations I implemented, based on the theory which I explain in my thesis. This presentation was given to business managers during the validation phase of my research. This presentation is discussed in my thesis.

  2. A presentation that I gave on February 19th 2002 in the presence of architects. This presentation discusses the results of my research, and also has a link to the first presentation. I gave a shorter version of this presentation during the workshop How 2 Sell Architecture, organized for account managers at Cap Gemini Ernst & Young on February 25th 2002.

Both presentations are very big (together more than 30 MB), and are therefore not available here. If you are interested in them, please contact Elsbeth Schultheiss from Cap Gemini Ernst & Young. She can be reached via the email address elsbeth.schultheiss@cgey.nl, starting March 6th 2002.


Return to the main page of my homepage.