ALSDA - Analyzing Latency of the Semantic Data Aggregator

This repository contains part of the code developed during the master's Thesis System for Netflow traffic enhancement in aim of cryptojacking detection, the separation of this repository is as follows:

• drivers. Contains the code written in Java for the input (input-driver) and output (Netlfow2CDS) drivers.
• aggregation. Two applications developed in Apache Flink with Java API for unidirectional to bidirectional flow conversion (Netflow2Bidirectional) and Netflowv9 aggregation (goflowAggregator)
• docker. Contains three folders to create the images for:
  • The Cryptomining Detection System (CDS) in crypto-detector
  • An event generator web server developed in python that lets you control the traffic rate of events written into a certain topic for testing purposes event-generator.
  • An image of goflow2 modified to obtain more fields goflow2-collector
• yang-models. Contains the netflowv9 (netflow-v9.yang) and aggregated (netflow-v9-agg.yang) yang models.
• utils/flink_tester. Contains the applications FlinkLatencyA and FlinkLatencyB used to measure the latency of any componente of the YANG-NAS system.
• tests. Contains all the scripts developed to test the latency, memory and CPU of the systems, with the corresponding scripts for creating the graphs.
  • controlled_DATA. Contains the data obtained when performed latency measures with controlled rate.
  • DATA. Contains the data obtained when performed latency measures WITHOUT controlled rate, also for the measures of CPU and MEMORY.
  • CODE. Contains the script to correlate latency events obtained from FlinkLatencyA and FlinkLatencyB, parsing prometheus JSON, creating graphs and more.
• mlsda_deployment. Contains a script that automates the deployment and measurement of latency for an application.