Sentinel

Detecting, analyzing, and mitigating anomalous network packets through a multi-dimensional latent space representation

Features

  • IoT failures result in loss of life
  • IoT devices operate in safety systems with size and power constraints
  • IoT economics drive device costs to a minimum, leaving security as an afterthought
  • IoT device lifetimes are measured in decades unlike conventional IT systems
  • IoT devices operate in environments where the users do not have the knowledge necessary for secure deployment

Deep learning packet inspection for
all loT traffic

  • One device monitors 30-XXXX IoT devices
  • Uses <15 W of power afterthought
  • Sub-millisecond latency/supports 5G
  • All packets are analyzed on the device (no cloud)
  • Finds anomalous network packets with >90% accuracy in real-time
  • Detects zero-day attacks in real-time
  • Enables simple visual identification of anomalous packets

USP:

  • On-premises 5G-capable intrusion detection device
  • Real-time deep packet inspection Real-time anomalous packet detection and alert
  • Output readable and visualizable by non-experts.
  • Output integrated with IoT network topology.
  • An inexpensive device with a small form factor and low power requirement
  • Deep packet inspection on every packet!
  • Does true Deep Packet Inspection. Looks at the payload and not just the meta-data.
  • On Site device, not in the cloud. Hardware based, which creates a zero latency scenario.
  • Uses an Ensemble Method, combining Supervised Approach with Unsupervised Clustering Approach
  • Can Monitor 1,000’s of IoT devices at once with zero latency.
  • Watches traffic coming into and out of the network and also watches inter-network traffic for anomalies.
  • Up to 1,000 Rad exposure.
  • No processor to slow the devices. It uses FPGA technology to run the system.
  • Combines ML/AI so the machine continues to learn and adapt to new potential attacks.