SLIPS, the Stratosphere behavioural network intrusion detection system that reconstructs traffic into per host profiles with Zeek and Redis and raises evidence from behavioural and threat intelligence detection modules.
SLIPS (Stratosphere Linux IPS) is a free, open source behavioural network intrusion detection and prevention system from the Stratosphere Laboratory. Rather than matching packets against fixed signatures, it hands captured traffic to Zeek to reconstruct connection flows, builds a behavioural profile for every host it sees inside Redis, and raises evidence and alerts from a suite of detection modules covering port scans, long and unusual connections, threat intelligence blacklists, command and control channel behaviour and malicious TLS fingerprints. It can analyse a live interface, act as an inline sensor on traffic routed through it, or run offline over packet captures for forensics, and it writes its findings to structured alert files that downstream tooling can consume.
This image is honest about scope. SLIPS's core mechanism, profiling the traffic it is handed, works exactly as designed on a cloud virtual network. What the platform constrains is visibility: a virtual network interface receives only the frames addressed to it, so this appliance inspects its own machine's traffic out of the box, inspects everything you deliberately route through it once you enable IP forwarding and a user defined route, and reads captures taken elsewhere, but it cannot passively monitor a neighbouring machine's traffic the way a physical network tap or a SPAN port can. The guide says so plainly rather than leaving you to discover it.
cloudimg delivers SLIPS as an unmodified checkout of the upstream release, so the complete corresponding source ships on the image alongside a written offer, and the behavioural and threat intelligence detection core is proven at build time to raise real evidence from a known malicious capture before the image is ever published. Redis carries no password because SLIPS does not support one, so every Redis instance is bound to loopback and the only port open is SSH, using the key you supplied. There is no login, no web interface and no bootstrap credential anywhere in the image to leak. The capture interface is a fact about your machine, not ours, so the shipped configuration is fail closed by design and refuses to start until first boot has rendered it against this instance's own interface, enforced by two independent levers that are each tested to fail closed. The memory heavy machine learning models are tuned to run within the recommended size and can be re enabled on a larger machine. Every build ships fully patched, and no build time traffic, evidence or packet captures survive into your image. Backed by 24/7 expert support.