CyberNeuro-RT
Most network defense works from a list of things somebody has already seen. That leaves a gap: the attack nobody has catalogued yet looks like ordinary traffic. CyberNeuro-RT closes that gap by learning your network instead of memorising someone else's.
It finds what it was never told to look for
Signature tools can only recognise an attack that has already been described to them. CyberNeuro-RT builds a picture of what normal looks like on your network and reports the traffic that does not fit, whether or not anyone has named it yet. The picture keeps updating, so it does not go stale as your network changes.
It runs where the traffic actually is
The same engine deploys in a cloud core, in a datacentre, or on a low-power chip attached to a sensor at a remote site. An edge deployment draws a small fraction of the power a conventional setup needs, which is what makes it practical somewhere with no rack and no cooling. One product to operate, not a separate edge line.
It respects how much time your analysts have
A queue nobody can finish is a queue that hides the real thing. Findings are ranked by how likely and how serious they are, tuned hard against false alarms, and presented so an analyst can see why something was flagged rather than just that it was. It sits behind the tools you already run and is built to reduce the chance an intrusion succeeds, not to promise none ever will.
The record behind it.
Developed with Lockheed Martin's MFC Division and Pennsylvania State University, under partial funding from the U.S. Department of Energy.
| Partners | Lockheed Martin Co. MFC Division · Pennsylvania State University |
|---|---|
| Funding | Partial funding from the U.S. Department of Energy |
| Deployment | Cloud core, datacentre, or low-power hardware at the edge |
| Position | Second layer, alongside existing perimeter tooling |
Tell us the program
and the problem.
We reply from San Jose, usually within two working days.