Hello Guys! I’m sharing Alejandro’s proposal for extended discussion, its worth for brainstorming, as this implementation is focused on OpenFlow, used broadly on SDN for communication between the controller and switches.
Part of the pathfinding job that serves as support of the code development and mentioned on the code repo can be found here: http://ieeexplore.ieee.org/document/5735752/?reload=true The noise patterns based on the findings from the paper can be caught definitely by ML, and specially if patterns present a volumetric behavior. So lets get discussion rolling. Thoughts? Thanks, Cesar On March 16, 2017 at 12:21:19 PM, () wrote: From: Alejandro Alcalde [mailto:[email protected]] Sent: Wednesday, March 8, 2017 7:41 AM Subject: [Collaboration proposal] - MafDet Hi, I was developing a system very similar to yours, but only for detecting DDoS. I want to collaborate in Apache Spot, right now I am reading the documentation to learn how Spot works. But since I've been developing my system for a few months, may be some of the code or ideas can be interesting to spot. The project is at https://github.com/algui91/MafDet Basically it aims to detect DDoS using SDNs and the Ryu controller for collecting OpenFlow Statistics. That part is totally completed, what is left is The MachineLearning, I've planned to use SOM for the anomaly detection. Let me know what you think. Best regards -- Alejandro Alcalde - elbauldelprogramador.com My public key: https://keyserver.ubuntu.com/pks/lookup?op=get&search=0x333395FF3326824B
