FDDA - Modelling a network

You start modelling by creating a new project, either through the menu or the tool bar. Next, you should select the algorithm you'd like to demonstrate. You can do this by selecting "Demonstration > Algorithm...". Then, you can add nodes and connections.

You add nodes by selecting the node tool (the one with the circle). As with the other tools, it stays selected after you add the new node if you press the <shift> key while selecting it. Then click into the project window wher you want your new node. You add connections - either directed or undirected - by selecting the tool and dragging a line from one node to the other. With the erazor tool, you may delete nodes or connections. You can add intermediate points on a line by holding <shift> and dragging on the line or by double clicking on the line. You can move these points to change the shape of the line. They look like this:

line with points

If you have the grid enabled (through 'View > Show grid') all nodes and points are automatically moved so they are centered on grid points. This makes it easy to align them vertically or horizontally. You can change the width of the grid in the options dialog ('Edit > Options). The grid isn't shown on exported JPEG files.

You select nodes, connections or points by clicking on them. You can select several at the same time by pressing the <ctrl> key or by drawing a selection rectangle with the mouse. If a connection or a node is selected, you can change their properties using the window on the right side of the application.

Nodes always have at least four properties (they can have more if you added some programmatically). The first one is called 'active' and determines whether the method start() of the node shall be called once the demonstration begins. All algorithms should have at least one node which is active. The method is also called if a node that wasn't working gets 'repaired' while the demo is running, BTW.
The second property determines the type of the node. There are two fixed types always available apart from those you've programmed. These are:

  1. Passive nodes: These don't do anything. All their methods are empty. This is the standard node type.
  2. Repeaters: These are shown differently in the project window, as small black circles. They don't act on their own, but they forward each incoming message to all their other neighbours. They don't count for the statistics, and you may use them e.g. to model a bus structure, as in the image below.

The third property of a node is its label. This is what is shown inside the node, so it shouldn't be longer than a few characters. An ID is probably best.
The last property determines whether the node is working. The methods of nodes that are not working are not called by the framework. They are crossed out in the project window.

Connections have three properties (always). One determines whether the connection is directed; directed connections can only transport messages in the direction they point. Another determines whether the connection is working; this is similar to nodes. Only working connections transport messages. The last property determines the speed of the connection, i.e. how fast messages are transported by it. 1.0 is the normal speed, while 2.0 would be twice as fast, 0.5 would be half as fast, etc. Important: the length of the connection determines how long a message will take; in other words, all messages have the same speed if their connections have the same speed.

While the demonstration of an algorithm is running, most modelling functions are disabled. You can't add or remove nodes or connections, you can't even move them (since that changes the length of the connections). One thing you can do, though, is changing the 'working' property of nodes or connections and see how your alogrithm reacts to failures.

URL dieser Seite: http://www.joerg-ruedenauer.de/Software/fdda/modelling.html
Autor dieser Seite: Jörg Rüdenauer
Letzte Änderung am: 11.02.2004

L-Space now!     Valid XHTML 1.0!     Valid CSS!