Regenerating Networked Systems’ Monitoring Traces with Neural Networks (NOMS'22)
Algorithm for correcting sessions of users of large-scale networked systems based on deep learning.
Input parameters:
Arguments(run_NOMS22.py):
-h, --help Show this help message and exit
--append, -a Append output logging file with analysis results (default=False)
--trials, -r Number of trials (default=1)
--start_trials, -s Start trials (default=0)
--skip_train, -t Skip training of the machine learning model training?
--campaign, -c Campaign [demo, lstm, no-lstm, deterministic](default=demo)
--verbosity, -v Verbosity logging level (INFO=20 DEBUG=10)
--------------------------------------------------------------
Arguments(main.py):
-h, --help Show this help message and exit
--original_swarm_file File of ground truth.
--training_swarm_file File of training samples
--corrected_swarm_file File of correction
--validation_swarm_file File of validation
--failed_swarm_file File of failed swarm
--analyse_file Analyse results with statistics
--dense_layers Number of dense layers (e.g. 1, 2, 3)
--neurons NEURONS Number neurons per layer
--cells CELLS Numbers cells(neurons) LSTM
--num_sample_training Number samples for training
--num_epochs Number epochs training
--analyse_file_mode Open mode (e.g. 'w' or 'a')
--model_architecture_file Full model architecture file
--model_weights_file Full model weights file
--size_window_left Left window size
--size_window_right Right window size
--threshold i.e. alpha (e.g. 0.5 - 0.95)
--pif PIF Pif (only for statistics)
--dataset DATASET Dataset (only for statistics)
--seed SEED Seed (only for statistics)
--lstm_mode Activate LSTM mode
--no-lstm_mode Deactivate LSTM mode
--skip_train, -t Skip training of the machine learning model
--deterministic_mode Set deterministic correction mode
--skip_correct, -c Skip correction of the dataset
--skip_analyse, -a Skip analysis of the results
--verbosity, -v Verbosity logging level (INFO=20 DEBUG=10)
--mode MODE Mode
--------------------------------------------------------------
Full traces available at: https://github.com/ComputerNetworks-UFRGS/TraceCollection/tree/master/01_traces
Run (all F_prob experiments):
python3 run_nom22.py -c lstm
Run (only one F_prob scenario)
python3 main.py
Run (all F_mon experiments):
python3 run_mif.py -c lstm
Run (only one F_mon scenario)
python3 main_mif.py
Requirements:
matplotlib 3.4.1
tensorflow 2.4.1
tqdm 4.60.0
numpy 1.18.5
keras 2.4.3
setuptools 45.2.0
h5py 2.10.0
ACKNOWLEDGMENTS
This study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. We also received funding from Rio Grande do Sul Research Foundation (FAPERGS) - Grant ARD 10/2020 and Nvidia – Academic Hardware Grant