NDP exploitation (IPv6)¶
Attack pattern¶
NDP is IPv6’s version of ARP, and it inherits ARP’s original sin: it asks the local link who owns an address and believes whatever answers. The same messages that let a host find its neighbours and its router, neighbour solicitations and advertisements, router advertisements, redirects, let an attacker on the segment claim any address, poison a neighbour cache, or fail a host’s duplicate-address check so it never gets an address at all. No message is signed, so a forged one is as good as a real one.
1. Ndp exploitation [OR]
1.1 Neighbour cache poisoning [OR]
1.1.1 Spoofed neighbour advertisement (NA)
1.1.1.1 Overriding legitimate entries in neighbour caches
1.1.1.2 Redirecting traffic to attacker-controlled MAC addresses
1.1.2 Gratuitous NA attacks
1.1.2.1 Unsolicited NA messages to inject false bindings
1.1.2.2 Persistent cache poisoning with periodic NA floods
1.1.3 Neighbour solicitation (NS) spoofing
1.1.3.1 Triggering unnecessary NA responses
1.1.3.2 Forcing address resolution for non-existent targets
1.2 Denial-of-service (DoS) [OR]
1.2.1 Neighbour cache exhaustion
1.2.1.1 Flooding with spoofed NS/NA messages
1.2.1.2 Filling cache with invalid entries to disrupt communications
1.2.2 Duplicate address detection (DAD) attacks
1.2.2.1 Responding to DAD probes to prevent address assignment
1.2.2.2 Causing address conflicts and host initialisation failures
1.2.3 Router advertisement (RA) flooding
1.2.3.1 Overwhelming hosts on the link with fake RAs
1.2.3.2 Disrupting legitimate router advertisements and prefix assignments
1.3 Man-in-the-middle (MitM) & interception [OR]
1.3.1 Default gateway impersonation
1.3.1.1 Spoofing router MAC addresses via NA messages
1.3.1.2 Redirecting all traffic through attacker nodes
1.3.2 Host impersonation
1.3.2.1 Taking over IPv6 addresses of critical servers
1.3.2.2 Intercepting sessions for SSH, HTTPS, or VoIP
1.3.3 Redirect attack exploitation
1.3.3.1 Sending ICMPv6 redirect messages to optimise malicious paths
1.3.3.2 Bypassing security controls via route manipulation
1.4 Reconnaissance & topology mapping [OR]
1.4.1 Neighbour unreachability detection (NUD) exploitation
1.4.1.1 Probing active hosts via NUD stimuli
1.4.1.2 Mapping live IPv6 addresses and their MAC bindings
1.4.2 Passive NA monitoring
1.4.2.1 Sniffing NA messages to build network maps
1.4.2.2 Identifying critical infrastructure targets
1.5 Evasion & stealth techniques [OR]
1.5.1 Extension header abuse
1.5.1.1 Embedding malicious ND messages in hop-by-hop options
1.5.1.2 Bypassing NDP inspection tools
1.5.2 Low-rate attacks
1.5.2.1 Slow poisoning to avoid detection thresholds
1.5.2.2 Timing-based evasion of monitoring systems
1.6 IoT & resource-constrained targeting [OR]
1.6.1 Battery exhaustion attacks
1.6.1.1 Forcing continuous ND activity on IoT devices
1.6.1.2 Draining power via repeated DAD or NUD triggers
1.6.2 Limited stack exploitation
1.6.2.1 Crashing devices with malformed ND packets
1.6.2.2 Abusing minimal ND implementations in embedded systems
1.7 Cloud & virtualisation attacks [OR]
1.7.1 Virtual switch exploitation
1.7.1.1 Poisoning neighbour caches in hypervisors
1.7.1.2 Breaking isolation between virtual machines
1.7.2 SDN controller targeting
1.7.2.1 Spoofing IPv6 addresses of controllers
1.7.2.2 Disrupting management plane communications
1.8 Transition mechanism attacks [OR]
1.8.1 NAT64/DNS64 exploitation
1.8.1.1 Poisoning NDP bindings for translated addresses
1.8.1.2 Redirecting IPv6-to-IPv4 traffic
1.8.2 Tunnelling protocol abuse
1.8.2.1 Injecting NDP messages into 6to4 or Teredo tunnels
1.8.2.2 Spoofing endpoints for tunnelled communications
1.9 Advanced persistent threat (APT) techniques [OR]
1.9.1 Long-term espionage
1.9.1.1 Sustained traffic interception via persistent cache poisoning
1.9.1.2 Targeting specific organisations (e.g., government, telecom)
1.9.2 Tool customisation
1.9.2.1 Developing bespoke tools for NDP exploitation
1.9.2.2 Adapting attacks to evade signature-based detection
1.10 Rogue DHCPv6 server attacks [OR]
1.10.1 Combining NDP and DHCPv6 exploitation
1.10.1.1 Spoofing both NA and DHCPv6 responses
1.10.1.2 Forcing hosts to use malicious configuration parameters
1.10.2 DNS server injection
1.10.2.1 Providing rogue DNS servers via DHCPv6 and NDP
1.10.2.2 Hijacking DNS resolution for entire subnets
Believes whoever answers¶
NDP messages carry no authentication, so any node on the link can claim any IP or MAC.
It is a plain request-response with no session state, which makes a spoofed reply trivial.
The message set (NS, NA, RS, RA, Redirect) is large, and each type is its own way in.
IPv6 assumes the local link is trusted, but joining that link, by cable or by wireless, is often easy.
Few networks monitor IPv6 or NDP specifically, so the attacks run without an audience.