Attack tree (DNS)¶
The ways to break DNS run from the decades-old to the not-yet-arrived: cache poisoning older than DNSSEC, amplified floods that turn open resolvers into weapons, phishing pipelines now scripted end to end, and signatures captured today against the day they can be forged.
1. Exploit Protocol Weaknesses [OR]
1.1 Cache Poisoning [OR]
1.1.1 Exploit weak TXID entropy in DoH/DoT/DoQ resolvers [OR]
1.1.1.1 Birthday attack on 16-bit TXID space
1.1.1.2 Timing attacks on resolver response handling
1.1.1.3 Fragment-based poisoning attacks
1.1.2 Side-channel attacks on encrypted DNS [OR]
1.1.2.1 TLS padding oracle attacks on DoT
1.1.2.2 QUIC protocol timing analysis on DoQ
1.1.2.3 HTTP/2 stream correlation attacks on DoH
1.1.3 DNSSEC Exploitation [OR]
1.1.3.1 NSEC/NSEC3 walking for zone enumeration
1.1.3.2 RRSIG timing attacks for key recovery
1.1.3.3 Algorithm downgrade attacks (to SHA-1 or an unsupported algorithm)
Prerequisite: AND (Attacker can intercept traffic AND resolver lacks full DNSSEC validation)
1.2 DDoS Amplification [OR]
1.2.1 Abuse misconfigured encrypted DNS resolvers [OR]
1.2.1.1 DoQ reflection with large TXT records
1.2.1.2 DoH POST request amplification
1.2.1.3 DoT session resumption attacks
1.2.2 DNSSEC-based amplification [OR]
1.2.2.1 NSEC3 response amplification
1.2.2.2 Large RRSIG reflection attacks
1.2.2.3 DNAME chain exploitation
Prerequisite: AND (Open resolver available AND vulnerable payload size > 1000 bytes)
1.3 Protocol-Specific Vulnerabilities [OR]
1.3.1 QUIC Protocol Exploitation (DoQ) [OR]
1.3.1.1 Connection migration hijacking
1.3.1.2 Stream priority manipulation
1.3.1.3 QUIC spin bit side-channel
1.3.2 HTTP/2 Exploitation (DoH) [OR]
1.3.2.1 HPACK header compression attacks
1.3.2.2 Server push cache poisoning
1.3.2.3 Stream dependency manipulation
1.3.3 TLS Session Attacks (DoT) [OR]
1.3.3.1 Session ticket stealing
1.3.3.2 Pre-shared key exhaustion
1.3.3.3 Certificate transparency log poisoning
2. Attack Encrypted DNS [OR]
2.1 Privacy Leaks [OR]
2.1.1 Metadata Correlation [AND]
2.1.1.1 IP + timestamp correlation across multiple resolvers
2.1.1.2 Query size and timing analysis
2.1.1.3 Server name indication (SNI) monitoring
2.1.2 ML-based fingerprinting [OR]
2.1.2.1 Neural network traffic analysis
2.1.2.2 Query pattern recognition
2.1.2.3 Encrypted traffic classification
2.1.3 Protocol Identification [OR]
2.1.3.1 DoH/DoT/DoQ protocol fingerprinting
2.1.3.2 Application-level protocol detection
2.1.3.3 Middlebox cooperation for traffic analysis
2.2 Downgrade Attacks [OR]
2.2.1 Force fallback to plaintext DNS [OR]
2.2.1.1 TCP RST injection on port 853 (DoT)
2.2.1.2 HTTP/2 GOAWAY frame injection (DoH)
2.2.1.3 QUIC connection close spoofing (DoQ)
2.2.2 Encryption Bypass [OR]
2.2.2.1 Disable ECH (Encrypted Client Hello) in DoH
2.2.2.2 TLS version downgrade attacks
2.2.2.3 QUIC version negotiation manipulation
2.2.3 Middlebox Interference [OR]
2.2.3.1 ISP-level protocol blocking
2.2.3.2 Enterprise firewall policy enforcement
2.2.3.3 Government-mandated protocol filtering
2.3 Certificate Attacks [OR]
2.3.1 CA Compromise [OR]
2.3.1.1 Rogue certificate issuance
2.3.1.2 Intermediate CA exploitation
2.3.1.3 Certificate transparency log poisoning
2.3.2 Client Validation Bypass [OR]
2.3.2.1 Self-signed certificate acceptance
2.3.2.2 Certificate pinning bypass
2.3.2.3 Trust store manipulation
3. Cloud/SaaS Exploits [OR]
3.1 Kubernetes DNS Compromise [OR]
3.1.1 CoreDNS / Etcd cache poisoning [OR]
3.1.1.1 API server compromise
3.1.1.2 ConfigMap manipulation
3.1.1.3 Plugin vulnerability exploitation
3.1.2 NetworkPolicy Bypass [OR]
3.1.2.1 Privileged pod escape
3.1.2.2 Node-level network access
3.1.2.3 Cross-namespace traffic interception
3.1.3 Service Mesh Exploitation [OR]
3.1.3.1 Istio/Linkerd DNS redirection
3.1.3.2 mTLS certificate theft
3.1.3.3 Sidecar proxy manipulation
3.2 Serverless Abuse [OR]
3.2.1 DNS Tunnelling [OR]
3.2.1.1 Lambda TXT record exfiltration
3.2.1.2 Cloud Functions DNS over HTTPS
3.2.1.3 Azure Functions private resolver abuse
3.2.2 Resource Exhaustion [OR]
3.2.2.1 DNS query burst attacks
3.2.2.2 Recursion depth exploitation
3.2.2.3 Cache saturation attacks
3.2.3 Cloud Integration Attacks [OR]
3.2.3.1 AWS Route 53 resolver hijacking
3.2.3.2 Google Cloud DNS API abuse
3.2.3.3 Azure Private DNS zone poisoning
3.3 Container Registry Attacks [OR]
3.3.1 Image Pull Manipulation [OR]
3.3.1.1 DNS spoofing for registry redirection
3.3.1.2 MITM attacks on image downloads
3.3.1.3 Cache poisoning for malicious images
3.3.2 Supply Chain Compromise [OR]
3.3.2.1 Malicious library injection via DNS
3.3.2.2 Dependency confusion attacks
3.3.2.3 Package manager DNS hijacking
4. Supply Chain Attacks [OR]
4.1 Registrar Hijacking [OR]
4.1.1 API Key Compromise [OR]
4.1.1.1 Cloudflare token theft
4.1.1.2 AWS Route 53 key leakage
4.1.1.3 Google Domains API abuse
4.1.2 Social Engineering [OR]
4.1.2.1 Registrar support impersonation
4.1.2.2 Post-GDPR WHOIS information gaps
4.1.2.3 Phone number porting attacks
4.1.3 Registry System Attacks [OR]
4.1.3.1 EPP protocol exploitation
4.1.3.2 Registry lock bypass
4.1.3.3 Transfer process manipulation
4.2 Subdomain Takeover [AND]
4.2.1 Dangling Resource Identification [OR]
4.2.1.1 CNAME mapping to unused cloud resources
4.2.1.2 NS record pointing to decommissioned servers
4.2.1.3 MX record targeting disabled services
4.2.2 Malicious Content Deployment [OR]
4.2.2.1 GitHub Pages site cloning
4.2.2.2 S3 bucket takeover
4.2.2.3 Azure Blob Storage hijacking
4.2.3 Persistence Mechanisms [OR]
4.2.3.1 SSL certificate procurement
4.2.3.2 DNS record obfuscation
4.2.3.3 Monitoring evasion techniques
4.3 CDN Compromise [OR]
4.3.1 DNS-Based CDN Manipulation [OR]
4.3.1.1 Edge server cache poisoning
4.3.1.2 Origin DNS spoofing
4.3.1.3 GeoDNS manipulation
4.3.2 Certificate Manipulation [OR]
4.3.2.1 SAN certificate abuse
4.3.2.2 CDN SSL termination bypass
4.3.2.3 Multi-CDN configuration conflicts
5. AI/ML-Augmented Attacks [OR]
5.1 Evasion Techniques [OR]
5.1.1 Reputation System Poisoning [OR]
5.1.1.1 DNS query pattern manipulation
5.1.1.2 Behavioural model contamination
5.1.1.3 Feedback loop exploitation
5.1.2 Query Obfuscation [OR]
5.1.2.1 GAN-generated benign-looking queries
5.1.2.2 CDN traffic mimicry
5.1.2.3 Legitimate domain spoofing
5.1.3 Adaptive Attacks [OR]
5.1.3.1 Reinforcement learning for evasion
5.1.3.2 Genetic algorithm optimisation
5.1.3.3 Transfer learning across networks
5.2 Phishing Automation [AND]
5.2.1 Domain Generation [OR]
5.2.1.1 LLM-generated homograph domains
5.2.1.2 Context-aware typosquatting
5.2.1.3 Cultural adaptation algorithms
5.2.2 Infrastructure Management [OR]
5.2.2.1 Dynamic DNS fast-flux networks
5.2.2.2 Automated certificate procurement
5.2.2.3 Multi-CDN abuse for resilience
5.2.3 Target Identification [OR]
5.2.3.1 NLP-based brand monitoring
5.2.3.2 Social media sentiment analysis
5.2.3.3 Employee behaviour prediction
6. Post-Quantum Threats [OR]
6.1 Cryptographic Harvesting [AND]
6.1.1 DNSSEC signature capture for later forgery [OR]
6.1.1.1 Record ECDSA-P256 signatures to forge once P-256 breaks
6.1.1.2 Record RSA-2048 DNSKEYs to forge once RSA breaks
6.1.1.3 NSEC3 chain enumeration
6.1.2 Quantum Decryption Preparation [OR]
6.1.2.1 Long-term encrypted data storage
6.1.2.2 Future decryption capability planning
6.1.2.3 Harvest-then-decrypt campaigns
6.1.3 Transition Period Exploitation [OR]
6.1.3.1 Algorithm confusion attacks
6.1.3.2 Hybrid scheme weaknesses
6.1.3.3 Backward compatibility exploitation
7. Data Exfiltration Techniques [OR]
7.1 DNS Tunnelling [AND]
7.1.1 Protocol Selection [OR]
7.1.1.1 Traditional DNS (TXT, NULL records)
7.1.1.2 DoH/DoT/DoQ encrypted tunnelling
7.1.1.3 ICMP-based DNS manipulation
7.1.2 Evasion Methods [OR]
7.1.2.1 Query rate limiting bypass
7.1.2.2 Legitimate traffic blending
7.1.2.3 Multiple resolver rotation
7.1.3 Data Encoding [OR]
7.1.3.1 Base32/64 encoding variations
7.1.3.2 Compression with error correction
7.1.3.3 Fragmentation and reassembly
7.2 Covert Channels [OR]
7.2.1 Timing-Based Exfiltration [OR]
7.2.1.1 Query response timing modulation
7.2.1.2 DNS refresh interval exploitation
7.2.1.3 TTL value manipulation
7.2.2 Storage Channels [OR]
7.2.2.1 DNS cache poisoning with data
7.2.2.2 NSEC3 gap exploitation
7.2.2.3 DNSSEC signature embedding
7.2.3 Behavioural Patterns [OR]
7.2.3.1 Query sequence encoding
7.2.3.2 Resolver selection patterns
7.2.3.3 Domain name generation algorithms
7.3 Exfiltration Infrastructure [OR]
7.3.1 Command and Control [OR]
7.3.1.1 Dynamic domain generation
7.3.1.2 DNS-based payload delivery
7.3.1.3 Dead drop resolvers
7.3.2 Data Processing [OR]
7.3.2.1 Distributed exfiltration aggregation
7.3.2.2 On-the-fly decoding services
7.3.2.3 Cloud function data processing
7.3.3 Persistence Mechanisms [OR]
7.3.3.1 Multiple exfiltration pathways
7.3.3.2 Fallback communication channels
7.3.3.3 Anti-forensic techniques
Nitty gritty risk table¶
Attack Path |
Technical Complexity |
Resources Required |
Risk Level |
Notes |
|---|---|---|---|---|
1.1.1.1 Birthday attack on 16-bit TXID space |
Medium |
Low |
Medium |
Feasible with traffic access and many queries; mitigated by DNSSEC and source port randomisation. |
1.1.1.2 Timing attacks on resolver response handling |
High |
Medium |
High |
Requires precise measurements and traffic observation to bias/guess transactions. |
1.1.1.3 Fragment-based poisoning attacks |
High |
Medium |
High |
Exploits IP fragmentation behaviours; success depends on network middleboxes and resolver settings. |
1.1.2.1 TLS padding oracle attacks on DoT |
Very High |
Medium |
High |
Complex cryptographic side-channel; requires tailored interaction with target DoT stack. |
1.1.2.2 QUIC protocol timing analysis on DoQ |
High |
Medium |
High |
Side-channel on QUIC recovery/timing; needs careful lab setup or privileged network vantage. |
1.1.2.3 HTTP/2 stream correlation attacks on DoH |
High |
Medium |
High |
Associates query streams via timing/priority; depends on client/server implementation quirks. |
1.1.3.1 NSEC/NSEC3 walking for zone enumeration |
Medium |
Low |
Medium |
Information disclosure rather than integrity break; rate limits and opt-out reduce impact. |
1.1.3.2 RRSIG timing attacks for key recovery |
Very High |
High |
Medium |
Academic-style side-channel; practical exploitation is difficult on hardened stacks. |
1.1.3.3 Algorithm downgrade (to SHA-1 or unsupported alg) |
High |
Medium |
High |
Downgrades to a forgeable algorithm the resolver still accepts. |
1.2.1.1 DoQ reflection with large TXT records |
Medium |
Medium |
High |
Leverages path amplification via QUIC; requires open/misconfigured resolvers. |
1.2.1.2 DoH POST request amplification |
Medium |
Medium |
High |
Uses HTTP request semantics for bandwidth multiplication; CDN/proxy behaviour matters. |
1.2.1.3 DoT session resumption attacks |
High |
Medium |
Medium |
Abuses TLS resumption tickets to cut cost per query; mitigations include rate and ticket controls. |
1.2.2.1 NSEC3 response amplification |
Medium |
Medium |
High |
DNSSEC negative responses can be large; filtering and minimisation reduce effect. |
1.2.2.2 Large RRSIG reflection attacks |
Medium |
Medium |
High |
Exploits oversised signed responses; best mitigated by response size limits and egress filtering. |
1.2.2.3 DNAME chain exploitation |
Medium |
Medium |
Medium |
Chained indirections inflate responses; effective mainly with weak ACLs. |
1.3.1.1 Connection migration hijacking |
Very High |
High |
High |
Requires QUIC internals knowledge and network control to spoof migration paths. |
1.3.1.2 Stream priority manipulation |
High |
Medium |
Medium |
Exploits scheduler to starve/shape DNS streams; mainly service degradation. |
1.3.1.3 QUIC spin bit side-channel |
Medium |
Low |
Medium |
Traffic analysis vector; limited by optional spin bit and padding. |
1.3.2.1 HPACK header compression attacks |
High |
Medium |
High |
Compression side-channels (e.g., BREACH-style) adapted to DoH; requires precise control/measurement. |
1.3.2.2 Server push cache poisoning |
High |
Medium |
High |
Abuses HTTP/2 push semantics to seed caches with attacker-chosen artefacts. |
1.3.2.3 Stream dependency manipulation |
Medium |
Low |
Medium |
QoS manipulation to infer/perturb query patterns. |
1.3.3.1 Session ticket stealing |
High |
Medium |
High |
Steals TLS/DoT tickets to resume as victim; needs endpoint compromise or MITM on storage. |
1.3.3.2 Pre-shared key exhaustion |
Medium |
Medium |
Medium |
Forces rotation/exhaustion of PSKs; primarily DoS on session setup. |
1.3.3.3 Certificate transparency log poisoning |
Very High |
High |
Medium |
Requires ecosystem-level manipulation; detection and auditing make success difficult. |
2.1.1.1 IP + timestamp correlation across multiple resolvers |
Medium |
Low |
Medium |
Cross-correlation deanonymises clients using multi-resolver setups. |
2.1.1.2 Query size and timing analysis |
Medium |
Low |
Medium |
Infers domains from packet sizes/timings even when encrypted. |
2.1.1.3 Server name indication (SNI) monitoring |
Low |
Low |
Medium |
Residual metadata (e.g., SNI when ECH disabled) leaks destinations. |
2.1.2.1 Neural network traffic analysis |
High |
Medium |
High |
ML models classify encrypted flows; requires labelled data and training pipeline. |
2.1.2.2 Query pattern recognition |
Medium |
Low |
Medium |
Identifies applications/users by request rhythms across sessions. |
2.1.2.3 Encrypted traffic classification |
High |
Medium |
High |
Generic encrypted flow fingerprinting; scales with telemetry access. |
2.1.3.1 DoH/DoT/DoQ protocol fingerprinting |
Low |
Low |
Medium |
Distinguishes protocols via handshake/behaviour traits for policy enforcement or blocking. |
2.1.3.2 Application-level protocol detection |
Medium |
Low |
Medium |
Infers client apps from traffic patterns/URLs used by DoH endpoints. |
2.1.3.3 Middlebox cooperation for traffic analysis |
Medium |
Medium |
High |
Correlated vantage points (ISP/CDN) increase deanonymisation power. |
2.2.1.1 TCP RST injection on port 853 (DoT) |
Medium |
Low |
High |
Active interference to force plaintext fallback; mitigated by hard-fail policies. |
2.2.1.2 HTTP/2 GOAWAY frame injection (DoH) |
High |
Medium |
High |
Requires HTTP/2 manipulation capabilities; targets client fallback logic. |
2.2.1.3 QUIC connection close spoofing (DoQ) |
High |
Medium |
High |
Spoofs transport errors to trigger downgrade; path validation can help. |
2.2.2.1 Disable ECH (Encrypted Client Hello) in DoH |
Medium |
Low |
Medium |
Strips or blocks ECH to expose SNI; depends on network control. |
2.2.2.2 TLS version downgrade attacks |
High |
Medium |
High |
Classical downgrade if endpoints permit weak versions/ciphers. |
2.2.2.3 QUIC version negotiation manipulation |
High |
Medium |
High |
Abuses version negotiation to weaker behaviour/performance. |
2.2.3.1 ISP-level protocol blocking |
Low |
Medium |
High |
Coarse-grained interference at scale; policy/legal environment dependent. |
2.2.3.2 Enterprise firewall policy enforcement |
Low |
Low |
Medium |
Localised blocking/inspection; limited to enterprise boundaries. |
2.2.3.3 Government-mandated protocol filtering |
Low |
High |
High |
High-impact nation-scale filtering; requires regulatory authority. |
2.3.1.1 Rogue certificate issuance |
High |
High |
High |
Compromised/abused CA issues certs for DoH/DoT endpoints. |
2.3.1.2 Intermediate CA exploitation |
High |
High |
High |
Targeting subordinate CAs for lateral issuance capability. |
2.3.1.3 Certificate transparency log poisoning |
Very High |
High |
Medium |
Attempts to corrupt ecosystem observability; hard to sustain covertly. |
2.3.2.1 Self-signed certificate acceptance |
Low |
Low |
Medium |
Misconfigured clients accept self-signed/invalid certs. |
2.3.2.2 Certificate pinning bypass |
High |
Medium |
High |
Requires binary patching/hook or proxy control on client. |
2.3.2.3 Trust store manipulation |
Medium |
Medium |
High |
Installs rogue roots/intermediates on endpoints; persistent if unnoticed. |
3.1.1.1 API server compromise |
High |
High |
High |
Kubernetes control-plane breach enables CoreDNS poisoning. |
3.1.1.2 ConfigMap manipulation |
Medium |
Medium |
High |
Alters CoreDNS config/plugins to redirect/poison cluster DNS. |
3.1.1.3 Plugin vulnerability exploitation |
High |
Medium |
High |
Targets CoreDNS/ETCD plugins for code execution or poisoning. |
3.1.2.1 Privileged pod escape |
High |
Medium |
High |
Escape to node network namespace to intercept DNS flows. |
3.1.2.2 Node-level network access |
Medium |
Medium |
High |
DaemonSet/agent on nodes can observe/alter DNS traffic. |
3.1.2.3 Cross-namespace traffic interception |
High |
Medium |
Medium |
Exploits policy gaps for lateral DNS observation within cluster. |
3.1.3.1 Istio/Linkerd DNS redirection |
High |
Medium |
High |
Service mesh policies/proxies used to reroute DNS to attacker. |
3.1.3.2 mTLS certificate theft |
High |
High |
High |
Theft of mesh certs enables decryption/impersonation of DNS sidecars. |
3.1.3.3 Sidecar proxy manipulation |
Medium |
Medium |
High |
Misconfig or compromise of sidecars to tamper with DNS egress. |
3.2.1.1 Lambda TXT record exfiltration |
Medium |
Low |
Medium |
Uses serverless functions to encode data in DNS responses. |
3.2.1.2 Cloud Functions DNS over HTTPS |
Medium |
Low |
Medium |
DoH endpoints hosted on serverless for stealthy tunnelling. |
3.2.1.3 Azure Functions private resolver abuse |
Medium |
Medium |
Medium |
Misusing private resolver integrations for data movement. |
3.2.2.1 DNS query burst attacks |
Low |
Low |
Medium |
Spiky invocation patterns drive resolver autoscaling/limits. |
3.2.2.2 Recursion depth exploitation |
Medium |
Low |
Medium |
Forces deep chains to inflate compute and latency. |
3.2.2.3 Cache saturation attacks |
Medium |
Medium |
Medium |
Fills caches with low-TTL entries to evict hot data. |
3.2.3.1 AWS Route 53 resolver hijacking |
High |
High |
High |
IAM/API misuse to change resolver rules/forwarders. |
3.2.3.2 Google Cloud DNS API abuse |
High |
High |
High |
Abuse of project/service accounts to alter DNS configuration. |
3.2.3.3 Azure Private DNS zone poisoning |
High |
High |
High |
Compromise of Azure DNS resources to redirect internal names. |
3.3.1.1 DNS spoofing for registry redirection |
Medium |
Medium |
High |
Manipulates name resolution to pull images from attacker infra. |
3.3.1.2 MITM attacks on image downloads |
High |
Medium |
High |
Intercepts registry traffic to inject malicious layers. |
3.3.1.3 Cache poisoning for malicious images |
Medium |
Medium |
High |
Seeds caching proxies with tampered image manifests. |
3.3.2.1 Malicious library injection via DNS |
High |
Medium |
High |
DNS redirection leading to malicious package sources. |
3.3.2.2 Dependency confusion attacks |
Medium |
Low |
High |
Public package names overshadow private ones using DNS hints. |
3.3.2.3 Package manager DNS hijacking |
High |
Medium |
High |
Alters resolver path for package indexes to attacker-controlled hosts. |
4.1.1.1 Cloudflare token theft |
Medium |
Medium |
High |
Stolen API tokens change DNS at registrar/hosted zones. |
4.1.1.2 AWS Route 53 key leakage |
High |
Medium |
High |
Exposed access keys enable authoritative DNS manipulation. |
4.1.1.3 Google Domains API abuse |
Medium |
Medium |
High |
Misused API credentials alter domain/records. |
4.1.2.1 Registrar support impersonation |
Low |
Low |
High |
Social engineering resets ownership or enables transfers. |
4.1.2.2 Post-GDPR WHOIS information gaps |
Low |
Low |
Medium |
Exploits limited contact visibility to facilitate impersonation. |
4.1.2.3 Phone number porting attacks |
Medium |
Low |
High |
SIM swap/port-out to intercept registrar 2FA challenges. |
4.1.3.1 EPP protocol exploitation |
High |
Medium |
High |
Targets registrar-registry channel; requires ecosystem access. |
4.1.3.2 Registry lock bypass |
High |
Medium |
High |
Circumvents lock controls through process gaps or insider help. |
4.1.3.3 Transfer process manipulation |
Medium |
Medium |
High |
Abuses transfer windows/auth codes to seize domains. |
4.2.1.1 CNAME mapping to unused cloud resources |
Low |
Low |
High |
Classic dangling CNAME takeover; easily automated discovery. |
4.2.1.2 NS record pointing to decommissioned servers |
Medium |
Low |
High |
Orphaned NS enables zone control for subdomains. |
4.2.1.3 MX record targeting disabled services |
Medium |
Low |
Medium |
Mail path hijack for phishing/data collection. |
4.2.2.1 GitHub Pages site cloning |
Low |
Low |
Medium |
Rebind subdomain to attacker’s pages for brand abuse. |
4.2.2.2 S3 bucket takeover |
Low |
Low |
High |
Recreates deleted buckets to host attacker content. |
4.2.2.3 Azure Blob Storage hijacking |
Low |
Low |
High |
Claims unbound storage names referenced by CNAMEs. |
4.2.3.1 SSL certificate procurement |
Medium |
Low |
High |
Validates control over taken subdomain to obtain certs. |
4.2.3.2 DNS record obfuscation |
Medium |
Low |
Medium |
Hides persistence with nested CNAMEs/TTL tricks. |
4.2.3.3 Monitoring evasion techniques |
Medium |
Low |
Medium |
Low-and-slow changes and selective responses to evade detection. |
4.3.1.1 Edge server cache poisoning |
High |
Medium |
High |
Poison CDN edge with malicious DNS/HTTP artefacts via control-path issues. |
4.3.1.2 Origin DNS spoofing |
High |
Medium |
High |
Redirect CDN to attacker “origin” by DNS manipulation. |
4.3.1.3 GeoDNS manipulation |
Medium |
Medium |
Medium |
Geographic split-horizon abuse to isolate victims. |
4.3.2.1 SAN certificate abuse |
High |
Medium |
High |
Misuses shared SAN certs for unintended hostnames. |
4.3.2.2 CDN SSL termination bypass |
High |
Medium |
High |
Forces traffic around expected TLS termination points. |
4.3.2.3 Multi-CDN configuration conflicts |
Medium |
Medium |
Medium |
Exploits inconsistent DNS/TLS between providers. |
5.1.1.1 DNS query pattern manipulation |
Medium |
Medium |
Medium |
Pollutes reputation systems with crafted benign-looking query mixes. |
5.1.1.2 Behavioural model contamination |
High |
Medium |
High |
Inserts poisoned samples into training/feedback loops. |
5.1.1.3 Feedback loop exploitation |
Medium |
Medium |
Medium |
Exploits automated block/allow updates to drift policies. |
5.1.2.1 GAN-generated benign-looking queries |
High |
Medium |
High |
ML-generated traffic mimics normal distributions to evade filters. |
5.1.2.2 CDN traffic mimicry |
Medium |
Low |
Medium |
Routes via popular CDNs to blend with noise and whitelists. |
5.1.2.3 Legitimate domain spoofing |
Medium |
Low |
Medium |
Uses typo/homograph names that resemble legitimate endpoints. |
5.1.3.1 Reinforcement learning for evasion |
Very High |
High |
High |
Trains agents to adapt queries against defences; research-heavy. |
5.1.3.2 Genetic algorithm optimisation |
High |
Medium |
Medium |
Evolves traffic features to reduce detection scores. |
5.1.3.3 Transfer learning across networks |
High |
Medium |
Medium |
Reuses models between environments to shorten tuning time. |
5.2.1.1 LLM-generated homograph domains |
Low |
Low |
High |
Automates convincing domain suggestions at scale. |
5.2.1.2 Context-aware typosquatting |
Low |
Low |
High |
Uses user/brand context to pick likely typos; increases click-through. |
5.2.1.3 Cultural adaptation algorithms |
Medium |
Low |
Medium |
Localises domain choices to regions/languages. |
5.2.2.1 Dynamic DNS fast-flux networks |
Medium |
Medium |
High |
Rapidly changing DNS answers to resist takedown. |
5.2.2.2 Automated certificate procurement |
Low |
Low |
High |
Scripted DV issuance increases trust signals for phishing. |
5.2.2.3 Multi-CDN abuse for resilience |
Medium |
Medium |
Medium |
Spreads infrastructure across CDNs to survive blocking. |
5.2.3.1 NLP-based brand monitoring |
Medium |
Low |
Medium |
Identifies hot targets to register convincing domains. |
5.2.3.2 Social media sentiment analysis |
Medium |
Low |
Medium |
Times campaigns around trending events. |
5.2.3.3 Employee behaviour prediction |
High |
Medium |
Medium |
Tailors lures using internal cadence/meeting patterns. |
6.1.1.1 Record ECDSA-P256 signatures to forge later |
Low |
Low |
Medium |
Public DNSSEC signatures, kept to forge once P-256 is broken. |
6.1.1.2 Record RSA-2048 DNSKEYs to forge later |
Low |
Low |
Medium |
Public DNSKEYs, kept to forge once RSA breaks; nothing is decrypted. |
6.1.1.3 NSEC3 chain enumeration |
Medium |
Low |
Medium |
Gathers structure of signed zones for later targeting. |
6.1.2.1 Long-term encrypted data storage |
Low |
Medium |
Medium |
Warehouses captured encrypted DNS/DoH traffic for future decryption. |
6.1.2.2 Future decryption capability planning |
Medium |
Medium |
Medium |
Organises key materials and compute pipelines anticipating PQ-era. |
6.1.2.3 Harvest-then-decrypt campaigns |
Medium |
High |
High |
Strategic programs at scale to capture now and decrypt later. |
6.1.3.1 Algorithm confusion attacks |
High |
Medium |
Medium |
Exploits transition periods where multiple DNSSEC algs coexist. |
6.1.3.2 Hybrid scheme weaknesses |
High |
Medium |
Medium |
Attacks mis-implemented hybrid (classical+PQ) deployments. |
6.1.3.3 Backward compatibility exploitation |
Medium |
Medium |
Medium |
Forces fallback to pre-PQ algorithms/policies. |
7.1.1.1 Traditional DNS (TXT, NULL records) |
Low |
Low |
Medium |
Simple tunnelling via classic records; widely detected by mature defences. |
7.1.1.2 DoH/DoT/DoQ encrypted tunnelling |
Medium |
Low |
High |
Encrypts payloads to bypass middleboxes; harder to monitor. |
7.1.1.3 ICMP-based DNS manipulation |
Medium |
Low |
Medium |
Covert data over ICMP with DNS semantics; niche and noisy. |
7.1.2.1 Query rate limiting bypass |
Medium |
Low |
Medium |
Distributes queries to evade per-source throttling. |
7.1.2.2 Legitimate traffic blending |
Medium |
Low |
High |
Shapes tunnels to match benign client/protocol patterns. |
7.1.2.3 Multiple resolver rotation |
Low |
Low |
Medium |
Rotates upstreams to evade IP-based detection. |
7.1.3.1 Base32/64 encoding variations |
Low |
Low |
Medium |
Obfuscates payloads within label constraints. |
7.1.3.2 Compression with error correction |
Medium |
Low |
Medium |
Balances throughput vs. reliability over lossy paths. |
7.1.3.3 Fragmentation and reassembly |
Medium |
Low |
Medium |
Splits payloads across queries to bypass size checks. |
7.2.1.1 Query response timing modulation |
Medium |
Low |
Medium |
Encodes bits in inter-arrival/latency; low bandwidth but stealthy. |
7.2.1.2 DNS refresh interval exploitation |
Medium |
Low |
Medium |
Uses refresh/probe timings as covert clock. |
7.2.1.3 TTL value manipulation |
Medium |
Low |
Medium |
Encodes data in TTL fields; detectable via anomalies. |
7.2.2.1 DNS cache poisoning with data |
High |
Medium |
High |
Seeds caches with attacker-controlled encodings for later retrieval. |
7.2.2.2 NSEC3 gap exploitation |
High |
Medium |
Medium |
Stores bits via crafted non-existent name patterns. |
7.2.2.3 DNSSEC signature embedding |
High |
Medium |
Medium |
Hides data in optional fields/signature slack; raises validation risks. |
7.2.3.1 Query sequence encoding |
Low |
Low |
Medium |
Uses order of labels/queries to represent data. |
7.2.3.2 Resolver selection patterns |
Low |
Low |
Medium |
Chooses resolvers in specific sequences as a codebook. |
7.2.3.3 Domain name generation algorithms |
Medium |
Low |
Medium |
DGA-based channels for resilient command/data paths. |
7.3.1.1 Dynamic domain generation |
Medium |
Medium |
High |
Rotates domains rapidly to evade blocklists. |
7.3.1.2 DNS-based payload delivery |
Medium |
Low |
Medium |
Delivers stage payloads via TXT/NULL to reduce HTTP exposure. |
7.3.1.3 Dead drop resolvers |
Medium |
Medium |
Medium |
Uses specific recursive resolvers as covert mailboxes. |
7.3.2.1 Distributed exfiltration aggregation |
Medium |
Medium |
Medium |
Fan-out/fan-in architecture to assemble data outside perimeter. |
7.3.2.2 On-the-fly decoding services |
Low |
Low |
Medium |
Cloud functions decode/forward tunneled chunks in real time. |
7.3.2.3 Cloud function data processing |
Medium |
Low |
Medium |
Serverless transforms/filters exfil data before storage. |
7.3.3.1 Multiple exfiltration pathways |
Medium |
Medium |
Medium |
Redundant DNS + alt channels improve resilience. |
7.3.3.2 Fallback communication channels |
Medium |
Low |
Medium |
Automatic switchover to new domains/resolvers/CDNs. |
7.3.3.3 Anti-forensic techniques |
High |
Medium |
High |
Deletes artefacts, pads timings, rotates keys to hinder IR. |
DNS heatmap¶
| Attack Category | Example Attack Path | Risk Level | Likely Adversary |
|---|---|---|---|
| Exploit Protocol Weaknesses | TXID birthday poisoning, HTTP/2 HPACK, QUIC migration spoof | High | Nation-state / Skilled cybercriminal |
| Attack Encrypted DNS | RST/GOAWAY/QUIC-close downgrades, ML traffic fingerprinting | High | Nation-state / ISP-level actor |
| Cloud/SaaS Exploits | CoreDNS ConfigMap poisoning, Route 53 rule hijack | High | Cloud-savvy attacker / APT |
| Supply Chain Attacks | Registrar API key theft, dangling CNAME takeover | High | Cybercriminal / APT |
| AI/ML-Augmented Attacks | GAN-shaped queries, RL-based evasion | Medium | Well-resourced criminal / Research-grade actor |
| Post-Quantum Threats | Harvest-now-decrypt-later, transition-period confusion | Medium | Nation-state / Strategic actor |
| Data Exfiltration Techniques | DoH/DoT/DoQ tunnelling, cache-based storage channels | High | Cybercriminal / APT |