Implementation of Artificial Intelligence-based Network and Security Monitor
Nurmi, Jaakko (2020)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2020121528489
https://urn.fi/URN:NBN:fi:amk-2020121528489
Tiivistelmä
This thesis main topic was to develop a prototype of a real-time, programmable and modular platform for network monitoring purposes by using agile software development methods. On the platform, an artificial intelligence-based data analysis processes for detecting a change in network behaviour and methods for automatic data enrichment were implemented.
The theory part contains a discussion about the key methods and techniques which was utilized in the development process and simplified operation principles of each developed process. Some developed processes were tested practically to evaluate the problems in the processes.
Modules for automatic processing and data analysis were also developed. These modules can be connected in case it is needed.
The most important data collection methods were benchmarked to detect problematic situations in the operation in different realistic situations. With the perception from the benchmark test, the problematic parts of the data collection were discovered and proposals for the solution were made which could be developed and tested in the next iterations of the development process.
Working Artificial intelligence-based detection and data enrichment methods were created. The results of the thesis allow multiple continuous research and development projects related to data collection and data analysis with statistical and artificial intelligence-based methods.
The theory part contains a discussion about the key methods and techniques which was utilized in the development process and simplified operation principles of each developed process. Some developed processes were tested practically to evaluate the problems in the processes.
Modules for automatic processing and data analysis were also developed. These modules can be connected in case it is needed.
The most important data collection methods were benchmarked to detect problematic situations in the operation in different realistic situations. With the perception from the benchmark test, the problematic parts of the data collection were discovered and proposals for the solution were made which could be developed and tested in the next iterations of the development process.
Working Artificial intelligence-based detection and data enrichment methods were created. The results of the thesis allow multiple continuous research and development projects related to data collection and data analysis with statistical and artificial intelligence-based methods.