/cae-performance-data-monitoring-tool

Primary LanguageC++MIT LicenseMIT

CAE: Performance Data Monitoring Tool

This tool was developed as part of the Softwaretechnik lecture at FH Aachen University for the CAE GmbH.
The lecture is part of the "Applied Mathematics and Computer Science" bachelor's degree.
The working group for this project consisted of:

Jonas Böcker
Simeon Burk
Tom Ferdinandt
Julius Gummersbach
Josias Kasemann
Taylor Seyler

Installation

(This installation was tested on windows using the mingw compiler)

Install Websocketpp:

  1. Clone https://github.com/zaphoyd/websocketpp repository into the same directory as the project folder

Install Boost:

  1. Download the zip file from https://www.boost.org/users/history/version_1_81_0.html
  2. Unpack it in a folder called boostlibrary in the same directory as the project folder.

Install nlohmann/json:

  1. Clone https://github.com/nlohmann/json repository into the same directory as the project folder

Install all dependencies from resources/requirements.txt with pip

Usage

  1. Start the dummy-server script with python3 resources/server.py
  2. Build and run the "run" configuration of the outermost CMakeLists.txt file
  3. Open the core/main/webpage/index.html file in a browser (tested with edge and chrome)

About this Tool

This tool is used to display data about threads that ran in a simulation instance.
On startup, the tool connects to a websocket server and receives a JSON object containing the data.
The data is then validated and stored in the graph- and data-manager.
When the user selects a thread to be displayed on the webpage, the webpage requests the graph and other data from the controller. In response the controller then creates an svg of the graph on demand and provides the data to the webpage.
The webpage then displays this svg and a table containing the modules of the selected thread.
Additionally, the user may choose to select two threads and perform a mathematical operation on them. The result of that operation is stored as a new graph and can itself be used in further operations. This feature can be used to compare the performance of multiple threads.

Project Timeline

During the SWT Project all MATSE students form groups to work on a software project for one of the companies that take part in the MATSE trainee program

Our group took on a project offer from the CAE GmbH, a company that mainly develops training systems for the german air force, navy and army.
The goal was to develop a tool that would receive data from a simulation at runtime and display it as graphs in a to be determined GUI.
These graphs were to be updated at runtime as new data from the simulation comes in.
It was also required that the user could select two of the graphs and perform a mathematic operation on them (e.g. add, subtract, multiply, divide) and store the result as a new graph. This lead to the quite interesting circumstance that the graphs were to form a mathematical group.

As we were all attending a lecture on C++ at the time, we decided to develop the software in C++ to apply what we had learned in the lecture in practice.
We started off by brainstorming a general design and structure for our software. You can find the class diagram here.
Implementing most of our initial design took us about 50% of the scheduled project time.

Throughout the whole project it was difficult to get example data for testing our project, partially because the CAE GmbH was not allowed to share the original, confidential simulation data with us.

We had initially planned to write a GUI in C++ using Qt or a similar framework, but we encountered some licensing problems. Because of that we decided to build a small webpage instead. This offered us the opportunity to learn more about hmtl, css and javascript as well.

We were almost done with most of the software when we were told that the application would in fact not receive live updates at runtime. Instead, it would receive all the data in one big batch at startup. This lead to a complete redesign of the software, which costed us a lot of time. Because of that we were not able to implement all the smaller features we had planned, like data exporting capabilities.

We quickly came up with a new design for the software. We were reusing and simplifying some of our initial ideas, as the new requirements were of course much simpler to implement. You can find the new class diagram here.
On January 22nd 2023 we presented the project at the MATSE Softwaretechnik fair. The tool will now be used and maintained by the CAE GmbH.

Lessons Learned

Apart from the obvious lessons learned about C++ and developing a website, we also gained a lot of experience in the process of developing a software and working in a team. The big misunderstanding about the requirements showed us how important it is to make sure everybody is on the same page before investing a lot of time into an idea.