Deception counters targeted attacks by combining several closely linked techniques: it entices attackers, fools them with false data, watches what they do and checks who they communicate with. This page explains the main components and how they fit together.
Technology Overview
At the core is a decoy network, a system of fake assets that appear to an attacker like any other machine or service. Around it sit lures that lead attackers towards the decoys, monitoring that analyses traffic leaving the network, and an incident engine that decides when the evidence is strong enough to raise an alarm.
Decoy Hosts and Services
Decoys present services that attackers look for: shared folders, database servers, web applications, remote shells, mail servers, network devices and management consoles. Low-interaction decoys emulate just enough of a service to record a connection attempt. High-interaction decoys run real operating systems and applications, so an attacker can log in and explore, while every keystroke is recorded in a contained environment.
As part of this staged interaction, a decoy can offer fabricated data that looks like sensitive information. While the attacker works on the bait, defenders study their goals and tactics, and the attacker is tied up away from real assets. Good decoys mirror the naming, operating systems and services of the surrounding network, and are refreshed as the environment changes.
Lures, Breadcrumbs and Honeytokens
Attackers rarely scan a network at random; they follow credentials, saved connections and documents they find on compromised machines. Deception platforms therefore plant breadcrumbs on real endpoints: cached credentials, mapped drives, browser bookmarks and configuration entries that point to decoys. Honeytokens such as fake cloud keys or canary database records extend the same idea to data. We cover them in a separate post on honeytokens and canary credentials.
Outbound Traffic Monitoring
Traffic leaving the network carries tell-tale signs of compromise. Automated malware traffic tends to be regular, small and persistent, while human browsing comes in bursts. By separating the two and focusing analysis on traffic that cannot be explained as normal user or system activity, monitoring can find hosts that are beaconing to command-and-control servers without overwhelming analysts.
Incident Triggering and Forensics
An incident is raised when the evidence is strong: an interaction with a decoy, the use of a planted credential, or a combination of suspicious traffic and decoy activity from the same host. Because decoys have no legitimate users, these alerts are far more reliable than most. The recorded sessions then show responders which machine was compromised, what the attacker tried and where they went.
Integration
Deception works best as part of a wider security stack. Alerts are sent to the SIEM or incident response platform, endpoint tools can isolate the compromised host automatically, and findings feed back into detection rules for the rest of the environment. Frameworks such as MITRE Engage help teams plan which deceptive techniques to use against which attacker behaviours.