Control-plane attack chain: summary¶
Why these three playbooks form a chain¶
These playbooks implement a true control-plane attack, not data-plane abuse.
From the BGP hijacking tree, most attacks are data-plane: they announce false routes (forged letters). Defenses like RPKI are designed to catch these by checking announcements against authoritative ROAs (the Guild Registry).
This attack chain corrupts the Registry itself. By phase 3, when we hijack traffic, RPKI validation endorses our hijack as legitimate. We’re not bypassing security. We’re corrupting the foundations security depends on.
The three phases¶
Phase 1: Registry reconnaissance and initial ROA creation¶
Objective: Establish ourselves as normal RPKI participant
Key actions:
Map victim’s RPKI deployment status
Create legitimate ROA for our own resources
Document baseline state before manipulation
Success metric: We appear as routine RPKI-aware network operator
Control-plane relevance: Reconnaissance of validation infrastructure, establishing trust
Phase 2: ROA scope expansion and validation environment mapping¶
Objective: Poison the validation infrastructure
Key actions:
Create fraudulent ROA authorizing our AS for victim’s prefix
Map which regions enforce RPKI validation
Deploy monitoring for ROA status changes
Success metric: Fraudulent ROA published and stable for 48+ hours
Control-plane relevance: This is the control-plane attack. We’ve edited the authoritative registry. From this point forward, validators will endorse our announcements as legitimate.
Phase 3: Prefix hijacking with RPKI validation cover¶
Objective: Exploit poisoned validation infrastructure to hijack traffic
Key actions:
Announce more-specific prefix from our AS
Verify traffic interception
Forward traffic to maintain service (reduce detection)
Monitor for detection indicators
Execute controlled withdrawal
Success metric: Traffic intercepted whilst RPKI validation returns VALID
Control-plane relevance: Payoff. Our hijack appears legitimate to validation systems because we corrupted them in phase 2.
Why this is different from normal BGP hijacks¶
Data-plane hijack (typical BGP attack):
Attacker: "I originate 203.0.113.0/24"
RPKI Validator: "Checking ROAs... INVALID. Drop this announcement."
Defense: Works as designed.
Control-plane attack (this playbook chain):
Phase 1-2: Attacker creates fraudulent ROA
Phase 3: "I originate 203.0.113.128/25"
RPKI Validator: "Checking ROAs... VALID. Accept this announcement."
Defense: Fails because validation system itself is compromised.
The distinction from the control-plane vs data-plane page:
Most BGP incidents are forged letters carried by an overly trusting postal service. Control-plane attacks rewrite the Guild Registry itself. From that moment on, everyone is “following the rules”. The rules are simply wrong.
That’s exactly what we’ve done. Validators are following the rules (check announcements against ROAs). The ROAs are simply wrong (because we poisoned them).
Timeline and dependencies¶
Week 1-2: Phase 1
├─ Day 1: Reconnaissance (Action 1.1)
├─ Day 2: Create our legitimate ROA (Action 1.2)
├─ Day 2-3: Wait for publication (30-90 minutes)
├─ Day 3-7: Baseline documentation (Action 1.3)
└─ Day 7: Waiting period (establish presence)
Week 3: Phase 2 (HIGH RISK)
├─ Day 1: Create fraudulent ROA (Action 2.1) ← CRITICAL, may fail
├─ Day 1: Wait for publication (2-4 hours)
├─ Day 2: Map validation deployment (Action 2.2)
├─ Day 3-4: Deploy monitoring (Action 2.3)
└─ Day 5-7: Waiting period (ensure ROA stable)
Week 4: Phase 3 (VISIBLE)
├─ Hour 0: Announce hijack prefix (Action 3.1)
├─ Hour 0-1: Verify interception (Action 3.2)
├─ Hour 1: Establish forwarding (Action 3.3)
├─ Hours 1-X: Monitor hijack (Action 3.4)
└─ Hour X: Controlled withdrawal (Action 3.5)
Total timeline: 3-4 weeks from start to completion.
Critical path: Phase 2 Action 2.1 (fraudulent ROA creation). If this fails, entire chain aborts.
Prerequisites¶
This attack chain requires:
Technical:
BGP peering session (for announcement in phase 3)
AS number allocation (legitimately obtained)
IP address allocation (for creating legitimate ROA in phase 1)
Forwarding infrastructure (for polite hijacking in phase 3)
Access:
Compromised RIR account credentials (for creating fraudulent ROA in phase 2)
OR insider access to victim’s RPKI infrastructure
OR exploit in RIR validation logic
Operational:
Multi-week operational security (attack spans 3-4 weeks)
24/7 monitoring capability during phase 3
Abort procedures prepared (for rapid withdrawal if detected)
The credential compromise prerequisite is often the hardest part. This playbook assumes it’s already accomplished. In reality, obtaining RIR account access may require separate attack operation (phishing, credential stuffing, social engineering).
Detection and countermeasures¶
Where defenders can catch this:
Phase 1:
Detection likelihood: ~0%
Appears like routine RPKI deployment
Phase 2:
Detection likelihood: ~30-50%
Fraudulent ROA creation may trigger RIR validation alerts
RPKI audit logs show ROA for prefix we don’t control
Requires defender actually reviewing audit logs (rare)
Phase 3:
Detection likelihood: ~40-70% (varies by victim’s monitoring)
BGP announcement visible in public monitors
Traffic path changes visible in NetFlow
But RPKI validation returns VALID, reducing alert priority
Why RPKI validation fails to detect:
RPKI is designed to prevent unauthorized announcements by checking against authoritative ROAs. This attack works because:
We created authoritative ROA (phase 2)
Validators trust that ROA (it’s signed by RIR)
Our announcement matches ROA (AS64513 for 203.0.113.0/24 maxLength /25)
Validation returns VALID
The defense mechanism is working correctly. It’s checking the right source. The source is simply poisoned.
Effective countermeasures:
RPKI audit logging and review
Monitor for ROA creation/modification
Alert on ROAs for prefixes not allocated to account
Requires active log review (not just collection)
Multi-path validation
Don’t rely solely on RPKI
Correlate RPKI with IRR, WHOIS, manual verification
Flag announcements that validate via RPKI but fail other checks
Traffic baseline monitoring
Monitor traffic source AS distribution
Alert on sudden shifts (traffic now coming from unexpected AS)
Requires NetFlow analysis and baselining
BGP community-based validation
Use signed BGP communities for additional validation layer
Requires widespread deployment (not currently available)
The most effective countermeasure is RPKI audit trail monitoring, which would catch phase 2 before phase 3 executes. Most organizations don’t have this because RPKI itself is new and auditing the validation infrastructure is not yet standard practice.
Operational security considerations¶
Attribution risk:
Low during phase 1 (appears legitimate) High during phase 2 (fraudulent ROA creation leaves audit trail) Medium during phase 3 (announcement visible but plausibly deniable as “testing”)
Post-operation, the strongest evidence is:
RIR audit logs showing fraudulent ROA creation from specific account
Timeline correlation (ROA created, announcement made, withdrawal executed)
If investigation reaches this depth, cover story of “operator error” becomes less plausible
Cover stories:
Phase 2 detection: “Spreadsheet error during bulk ROA update, accidentally included wrong prefix” Phase 3 detection: “Testing RPKI deployment, mistakenly announced prefix from test configuration”
Both are plausible because RPKI is complex and these errors genuinely happen. Cover stories become less believable if:
Multiple phases detected and correlated
Timeline shows coordination (too precise to be accidental)
Forwarding infrastructure discovered (demonstrates intent, not accident)
Cleanup:
Post-operation, revoke fraudulent ROA promptly. Long-term persistence is strongest evidence of deliberate attack versus mistake.
Why this chain was selected for Red Lantern¶
These three playbooks demonstrate:
Structural vulnerability in trust infrastructure
RPKI is meant to secure BGP
RPKI itself can be compromised
Defense becomes attack vector
Multi-phase attack requiring operational security
Not single-action attack
Requires weeks of preparation
Tests detection across multiple phases
Realistic attack path with budget reality gaps
Credential compromise (happens)
Incomplete RPKI auditing (common)
Limited validation deployment (40% globally)
Clear distinction between control-plane and data-plane
Explicitly demonstrates Registry manipulation
Shows why validation alone insufficient
Illustrates need for validation system auditing
This is valuable for Wazuh detection engineering because:
Most BGP detection rules focus on announcements (data-plane)
Few rules exist for RPKI infrastructure monitoring (control-plane)
This attack shows why both layers need monitoring
Limitations and caveats¶
This attack chain will fail if:
RIR has robust ROA validation (catches fraudulent ROA in phase 2)
Victim has RPKI audit monitoring (detects fraudulent ROA before phase 3)
Global RPKI validation enforcement >80% (our announcement dropped everywhere)
Victim has real-time BGP monitoring with alerting (detected in phase 3 within minutes)
Success requires victim to have:
Partial or absent RPKI auditing (common)
Incomplete validation deployment globally (currently true, ~40% deployment)
Limited BGP monitoring (common outside major transit providers)
As RPKI deployment increases and auditing practices mature, this attack path becomes less viable. Currently (2024-2025), it remains realistic threat.
References¶
BGP hijacking attack tree - structural attack taxonomy
Control-plane vs data-plane distinction - conceptual framework
RPKI documentation - validation infrastructure
Routinator RPKI validator - validation tool
BGPmon public monitoring - announcement visibility
Success criteria for complete chain¶
The three-phase chain succeeds when:
Fraudulent ROA is created and remains published (phase 2)
Traffic is intercepted with RPKI validation returning VALID (phase 3)
Operational objective achieved (traffic visibility/interception for required duration)
Clean withdrawal executed without detection (phase 3)
Post-operation cleanup completed (fraudulent ROA revoked)
The chain demonstrates control-plane attack if validation systems endorsed our hijack as legitimate at time of execution, even if later investigation reveals the attack.
From the purple page:
A control-plane attack, properly defined, does not lie within the rules. It rewrites the rules themselves.
That’s what we’ve accomplished. We rewrote the RPKI rules (phase 2), then operated within those corrupted rules (phase 3). Everyone followed the rules. The rules were simply wrong.