In FortiDeceptor deployments, deception technology is valuable because legitimate users and applications should have little reason to interact with a well-placed decoy. That can make decoy activity unusually high-signal compared with many conventional detections. The signal, however, depends on deployment quality. Decoys that look unrealistic, sit in irrelevant segments, expose the wrong services, or generate operational noise lose their advantage.
FortiDeceptor creates decoy virtual machines and lures across enterprise, cloud, and specialized environments. For Fortinet security operations, deception is an instrumented detection layer: choose where attackers are likely to move, present believable targets, capture activity safely, and route incidents into a response workflow.
Effective attack-surface discovery begins with the real environment: segments, operating systems, services, identities, administrative pathways, and high-value assets. That inventory also makes decoy placement believable because the deception can mirror what an intruder expects to find nearby.
Place decoys where lateral movement would matter
FortiDeceptor best practices emphasize coverage across network segments and locations. A decoy in an isolated lab does not provide much visibility into compromise of production identity, server, user, or OT networks. Deployment should follow likely attacker movement paths and the locations where unauthorized discovery would be meaningful.
Match decoy operating systems and services to the segment. Windows servers, Linux systems, RDP, SMB, SSH, web services, industrial protocols, and application decoys should reflect what an attacker expects to find nearby. Too much uniformity makes the deception obvious.
Coverage should include both crown-jewel adjacency and common transit areas. A decoy near critical systems can detect late-stage movement, while decoys in user and server segments can expose reconnaissance earlier.
Placement should be risk-driven rather than evenly distributed by device count. Segments containing privileged administration, sensitive servers, identity infrastructure, or paths to critical applications may justify denser or more specialized deception than low-risk networks.
Remote offices and cloud networks should be included when attackers can move through them. A deception design focused only on the main data center can miss compromise that begins in SaaS-connected, remote-access, or cloud-hosted environments.
Realism requires more than an IP address
A convincing decoy has plausible hostname, service banners, open services, user artifacts, file shares, and network placement. FortiDeceptor supports deception operating systems, services, and lures that can be tailored to the environment. The goal is not theatrical realism; it is enough consistency that automated tooling and a human intruder do not immediately reject the target.
Use fake data structures that resemble local naming and business patterns without containing real secrets. Lures can point attackers toward decoys, but they should not create operational confusion for legitimate users or administrators.
Review realism after infrastructure changes. If the real environment moves to newer operating systems, naming conventions, or cloud services while decoys remain static for years, the deception becomes easier to identify.
Realism also includes network behavior. Decoys should respond on ports and services that make sense for their claimed role. A host named like a database server that exposes only generic web services creates an inconsistency a skilled attacker may recognize.
Keep decoy fingerprints current after upgrades. If banners, protocol versions, or OS characteristics lag far behind the surrounding environment, they can become obvious markers rather than convincing targets.
Safe lists and outbound controls protect the environment
Security teams and administrators sometimes need to access decoys for maintenance or testing. Safe-list controls help prevent those activities from being treated as attacker behavior, but safe lists must remain narrow. Broad exclusions can create blind spots an attacker may inherit.
Outbound traffic deserves special attention. A compromised decoy should not become a pivot platform for attacking real systems or the internet. FortiDeceptor deployment options include controls to block outbound traffic on supported decoys. Validate those controls rather than assuming the decoy is harmless because it is fake.
Domain-joined or customized decoys require even more care. Credentials, trust relationships, DNS configuration, and management access should be designed so the decoy appears believable without creating a new privileged path.
Safe lists should be reviewed against scanner infrastructure and management tooling. If vulnerability scanners are allowed to interact with decoys, their traffic may be intentionally suppressed; if those scanners are compromised later, the same exception can hide attacker activity.
Outbound restrictions should be tested with the specific services enabled on the decoy. Some believable interactions require limited dependencies such as DNS or directory communication, but broad egress undermines containment.
Alerts should preserve the attacker’s sequence
The advantage of deception is not only that an alert fired. It is that the interaction can reveal tools, commands, source systems, credentials attempted, protocols used, and lateral movement behavior. Preserve the sequence so analysts can understand what the attacker did before and after the decoy interaction.
Correlate decoy activity with other security telemetry and threat-intelligence context when it changes the interpretation. A source address already associated with malicious infrastructure may raise urgency, but a decoy touch from an internal workstation can be even more significant because it suggests compromise or unauthorized scanning.
Do not collapse all decoy interactions into one generic severity. A port scan, failed login, successful shell, credential use, file access, and command execution represent different stages and should trigger different investigation depth.
Preserve the source context of internal hits. DHCP address, authenticated user, switch port, endpoint identity, and preceding network events can turn a single decoy connection into a fast containment decision. Integrations should enrich the event automatically where possible.
Decoy incidents can also reveal credential misuse. If a fake credential or lure is touched, identify where that artifact was planted and which real endpoint likely exposed it to the attacker.
Reset behavior is part of incident containment
FortiDeceptor can reset decoys after incidents or administrative behavior. Resetting restores a clean state and prevents one attacker interaction from permanently changing what later attackers see. The reset interval should balance evidence preservation with operational cleanliness.
Define what data must be exported before reset: packet captures, command history, process activity, files, screenshots, hashes, source addresses, and timestamps. Automatic reset without evidence preservation can destroy the very intelligence the decoy was deployed to collect.
Test the reset path and confirm that the decoy returns with the expected identity, services, and lures. A decoy that silently fails to relaunch creates a coverage gap.
Reset intervals should be chosen by evidence needs. A rapidly resetting decoy returns to service quickly but may erase a longer attacker session. A slower reset preserves more interaction at the cost of leaving the environment in a changed state longer.
Test decoy recovery after software upgrades and configuration changes. A deployment that appears healthy in the management interface may still have broken services or stale lures that reduce realism.
Route high-confidence signals into a disciplined response
Because decoy interactions are often high-signal, teams may be tempted to automate aggressive containment. Start with enrichment, case creation, endpoint lookup, identity lookup, and evidence capture. Isolation or blocking should consider the possibility of security scanners, misrouted legitimate traffic, or administrator mistakes.
The response model should fit the organization’s security-operations architecture. Deception does not replace endpoint, identity, network, or SIEM evidence; it gives those systems a strong pivot point for deciding what to investigate next.
Define ownership for internal versus external sources. A decoy hit from an employee workstation may require endpoint containment and user investigation, while an internet source touching an externally exposed deception asset may primarily support threat intelligence.
Define severity based on interaction depth and source context. A single SYN to a decoy should not automatically receive the same priority as authenticated access, command execution, or lateral movement from a managed workstation. A tiered model preserves the high-signal advantage of deception.
Create analyst playbooks for common decoy interactions so evidence collection is consistent. Include endpoint lookup, user identity, nearby assets, recent authentication, DNS history, and whether the source touched other decoys. Consistent pivots reduce investigation time.
Measure signal quality and coverage
Track decoy coverage by segment, decoy health, lure deployment, interaction volume, percentage of alerts attributable to authorized scanning, time from interaction to analyst review, incidents confirmed from deception, and decoys that never receive any traffic or telemetry.
Use those measurements to refine the program inside the broader Fortinet environment. A decoy with no interactions may be perfectly placed in a quiet segment, or it may be invisible to the attack paths that matter. Review placement instead of assuming silence equals safety.
Effective deception creates a controlled asymmetry: attackers see believable opportunities while defenders see activity that should almost never occur. The policy succeeds when that asymmetry produces actionable evidence without introducing new risk.
Red-team and purple-team exercises are useful for validation. Ask testers whether decoys were discoverable, believable, and placed on realistic movement paths, then compare their feedback with telemetry. The goal is not to fool a tester forever; it is to create interactions defenders can detect early.
Track the ratio of high-confidence malicious or unauthorized interactions to benign touches. If legitimate scanners dominate, tune safe lists or placement rather than accepting chronic noise that teaches analysts to ignore deception alerts.
Retire decoys that no longer match the network. Replacing stale decoys is part of maintaining credibility because attackers can compare host age, exposed services, and naming conventions with neighboring systems.