Development of a program (written in Haskell) that visualizes a live animation of a network simulation (dynamic flow of packets)

In this Master thesis project a dynamic animation of Network-on-Chip simulation runs shall be developed. There are three types of information that shall be visualized dynamically during a simulation.

- The flow of individual packets shall be visible;
- The current load on links, switches and on buffers shall be encoded
by visual effects and colours.
- The change of the load, called the \emph{load gradient} on links,
switches and buffers shall be
encoded by visual effects and colors as well. It shall be separate and clearly distinct from the current load.

The user shall be able to dynamically control the speed of the animation
and a few visualization parameters. The actual simulation shall not be
dynamically controlled by the user, e.g. the generation of packets and the behavior of resources is determined as a batch process at the beginning of the simulation. Only the visualization shall be controllable during the simulation. In particular the user shall be able to do the following.

- Step through the simulation cycle by cycle;
- Change the step width counted in number of cycles;
- Set and change the animation granularity, which is the period in
number of cycles in which the visual image is updated. E.g. a granularity of 1 means the image is updated every simulation cycle; a granularity of 100 means the image is updated once in 100 cycles.
- Set and change the window for load measurement. A window of 1 means
the load is defined as the load in the current cycle. A window of 10
means the load is defined as the average load of the last 10 cycles.
- Set and change the window for the load gradient. A window of 1 means
the gradient is defined as the difference of the load between the
previous and the current cycle. A window of 10 means the gradient is
defined as the difference of the load of two windows, each of size load measurement window. The current window is the current load measurement window. The previous window is the load measurement windows load gradient cycles in the past. For instance, if the load measurement window is 1 and the gradient window is 10, then the gradient is defined as the load difference of the current cycle to the load 10 cycles ago.
- Set and change what information is made visible. At least the
following options should exist.
+ Packets are visible or not;
+ Packets are shown with their id string;
+ Load on switches is visible or not;
+ Load on links is visible or not;
+ Load on interface buffers is visible or not;
+ Load gradient on switches is visible or not;
+ Load gradient on links is visible or not;
+ Load gradient on interface buffers is visible or not;

The animation shall be developed in the functional langauguage Haskell based on the Yampa library,that allows visualization and animation.
For more information see: http://web.it.kth.se/~axel/MSc-Theses/node1.html