Modeler as an Authoring Tool
Traditional users of Modeler view their deliverable as the results of one or more simulation studies. However, other users may view their objective as the creation of a customized modeling environment. The environment consists of modified Riverbed analysis software that allows end users to create models at the network level, specify data collection, run simulations, and do analysis and animations.
The customization of the modeling environment seen by the end user is made possible by Modeler. Using Modeler, sophisticated models of nodes and processes can be created. These models can then be prepared as a library for use by the end user who is less familiar with the detailed operation of the system. To the end user, these models appear as off-the-shelf components; he is protected from the underlying complexity.
In addition, the user interface can be tailored to offer a truly customized environment to the end user. Modeler offers extremely open interfaces, allowing easy integration with complementary network planning and management tools. Key features supporting this type of application are listed in the following table.
Table 19: Key Features Supporting Development of Customized Modeling Environments
| Feature | Description |
|---|---|
| Model Security | An author can protect models so that only authorized users can view or use them. Models are encrypted and unusable when shipped. They can then be unlocked with three security levels, supporting different requirements to protect information. Only the author can provide keys to unlock models. |
| External Tool Support | Authors can augment Modeler’s user interface by adding action buttons, menus, and menu items, or by defining actions that can be performed when users click on objects. The processing associated with each action is specified using a scripting language, which includes support for invoking external programs, as well as importing data from those programs. |
| Asynchronous Commands | Modeler can implement commands that are asynchronously received from external programs via interprocess communication. These actions include opening tools, loading models, graphical positioning, and modification of objects. This allows models to be updated in real-time as the state of other applications change. |
| Attribute Interfaces | The key to productizing models is tailoring the interface that they present to the user. For node and process models, attributes and statistics constitute the majority of the visible interface. Attributes can be customized in derived models (without changing base models) by renaming them, pre-assigning their value and hiding them. Their properties can be controlled including allowing you to: restrict ranges; offer meaningful symbolic values; specify default values; and customize documentation. Similar capabilities are provided for the statistics offered by a model. |
| Application Program Interface | Modeler offers an Application Program Interface (API) called EMA (External Model Access). EMA allows your programs to access (create and/or parse) models. This means that you can automatically generate models to reflect information that is stored elsewhere (for example, in Network Management databases). You can also extract information from models and simulation output files to use them in external applications. Because EMA is a published interface, you can do robust integration of Modeler and other tools that will survive new software releases. |