Msnm-S - Multivariate Statistical Network Monitoring-Sensor
MSNM-S (Multivariate Statistical Network Monitoring-Sensor) shows the practical suitability of the approaches constitute inwards PCA-MSNM together with inwards Hierarchical PCA-MSNM works. The rootage i acquaint the MSNM approach together with novel multivariate statistical methodology for network anomaly detection piece the 2d i proposes the previous i inwards a hierarchical together with structured network systems. The principal thought behind these works, is the purpose of multivariate statistical techniques to generate useful information inwards the shape of 2 statistics. Such a calorie-free information comes from lower to higher levels inwards a network hierarchy. This way, the root sensor (for example, a edge router) received all the statistical information beingness able to compute its ain statistics (Q,D). By inspecting this statistics, a safety analyst tin determine if anomalous effect are happening when around of the statistic values are inwards a higher house for certain command limits.
[A pre-print unloose of the move is avilable at https://arxiv.org/abs/1907.13612]
MSNM-S is conceived to hold upward extremely scalable together with aseptic because simply 2 parameters are sent amid levels or devices inwards the monitored network or system. Additionally, the MSNMSensor is able to deal multiple together with heterogeneous type information sources at each monitored devices cheers to the FCParser (Feature equally a Counter Parser) characteristic technology scientific discipline approach.
Installation
Requirements
MSNSensor runs alongside python 2.7 together with has been successfully tested on Ubuntu from 16.04 version together with above. Also, the next dependencies has to hold upward installed.
How to install
Creating a python execution environs is, in all likelihood the ameliorate agency to run the application. So I recommend you lot to produce i earlier doing the requeriments installation. Anaconda environs tin assistance you lot and, if you lot produce upward one's hear to purpose it, run the next commands:
How to run an example
Please run across instructions at examples or download the pre-configured VM at MSNM-S-UBUNTU. We recommend you lot to purpose the VM. Remember to push clit the repository to teach the MSNM-S projection updated. In the following, you lot tin run across the necessary steps to run the pre-configure experiment inwards the VM:
Running the MSNM-Ss (backend)
Open a final window together with activate netflow daemon together with collector.
Open a novel final window.
Authors together with license
MSNM Sensor - GNU GPL - Roberto Magán-Carrión, José Camacho together with Gabriel Maciá-Fernández
[A pre-print unloose of the move is avilable at https://arxiv.org/abs/1907.13612]
MSNM-S is conceived to hold upward extremely scalable together with aseptic because simply 2 parameters are sent amid levels or devices inwards the monitored network or system. Additionally, the MSNMSensor is able to deal multiple together with heterogeneous type information sources at each monitored devices cheers to the FCParser (Feature equally a Counter Parser) characteristic technology scientific discipline approach.
Installation
Requirements
MSNSensor runs alongside python 2.7 together with has been successfully tested on Ubuntu from 16.04 version together with above. Also, the next dependencies has to hold upward installed.
- numpy >= 1.14
- scipy >=1.0
- pyyaml >= 3.12
- IPy >= 0.83
- pandas >= 0.22
- watchdog >= 0.8.3
- FCParser (Feature equally a Counter Parser) == 1.0.0
How to install
Creating a python execution environs is, in all likelihood the ameliorate agency to run the application. So I recommend you lot to produce i earlier doing the requeriments installation. Anaconda environs tin assistance you lot and, if you lot produce upward one's hear to purpose it, run the next commands:
$ conda produce -n py27 python=2.7 $ conda activate py27(py27) $ pip install -r requirements.txtHow to run an example
Please run across instructions at examples or download the pre-configured VM at MSNM-S-UBUNTU. We recommend you lot to purpose the VM. Remember to push clit the repository to teach the MSNM-S projection updated. In the following, you lot tin run across the necessary steps to run the pre-configure experiment inwards the VM:
Running the MSNM-Ss (backend)
Open a final window together with activate netflow daemon together with collector.
$ cd /msnm-sensor/scripts/netflow/ $ sudo ./activateNetflow.sh (pass: msnm1234)$ cd /msnm-sensor/scripts/ $ conda activate py27 $ ./start_experiment.sh ../examples/scenario_4/ $ ps -ef | grep msnmsensor (just to banking concern friction match if all the 4 MSNM-Ss are running) $ tail -500f /msnm-sensor/examples/scenario_4/borderRouter/logs/msnm.log (another agency to run across how the MSNM-S is working. Replace the advert of the MSNM-S if you lot desire to run across the others.)Open a novel final window.
$ cd /msnm-sensor/dashboard/ $ conda activate msnm-dashboard $ ln -s ../examples examples $ python manage.py runserverAuthors together with license
MSNM Sensor - GNU GPL - Roberto Magán-Carrión, José Camacho together with Gabriel Maciá-Fernández
