beta version

AI CyberSecurity

AI based sniffing anomaly detector

Problem
Data is transported on physical layers as TCP packets. Each of this layer could have a throughput of 100gbit/s. Being able to process these data could show hacker attacks or network malfunctions.
State of art network algorithms are based on reasoning engine such as Microsoft Z3. These solutions are NP-hard: the computational power needed is very high
Hundreds of layers in parallel make this challenge even harder. These data need to be processed in nearly real time to get actionable insights: this leads to the need to develop a solution from the Physical layer to the Application one.
Denial Of Service - Q4 2017 vs Q1 2018 +47%
Denial-of-service attack (DoS attack) is a cyber-attack in which the hacker seeks to make a machine or network resource unavailable to its intended users by flooding the targeted machine or resource with superfluous requests in an attempt to overload systems.

Solution
  1. Step 1: Feature Extraction
  2. A hardware layer provides the capability to reduce the throughput keeping all the information available. This layer persists the stream locally.
  3. Step 2: On field Anomaly Detection
  4. FPGA based layer applies Machine Learning algorithms on a physical appliance to detect anomalies as soon as possible.
  5. Step 3: Cloud Training
  6. All data are also copied on the Cloud to store data, create analytics and advanced analytics dashboard and enhance Machine Learning algorithms effectiveness.