mldMulticast Listener Discovery v1 & v2 (IPv6)IPv6 Multicast Listener Discovery (querier <-> listener mld)
What It Does on the Wire
What igmp does for IPv4, mld does for IPv6 (ff02::/16 -> 33:33:xx:xx:xx:xx Ethernet MACs at Hop Limit 1), making it easy to check whether your leaf switches actually snoop IPv6 multicast.
Multicast Listener Discovery (mld) is the IPv6 counterpart to IGMP, carried inside ICMPv6 (Next Header 58) with a link-local Hop Limit of 1. IPv6 nodes use MLD not only for streaming multicast applications, but also to join solicited-node multicast groups (ff02::1:ffxx:xxxx) required by IPv6 Neighbor Discovery.
FrameAssembly constructs 24-byte MLD messages (query Type 130, report Type 131, and done Type 132) with automatic RFC 2464 33:33:xx:xx:xx:xx multicast Ethernet MAC derivation and IPv6 pseudo-header checksums.
Switch MLD Snooping & Solicited-Node Group Audits
Verify data-center leaf switches properly maintain MLD snooping entries for ff02::1:ff00:0/104 solicited-node groups so IPv6 Neighbor Discovery never drops.
Dual-Stack Multicast Migration Testing
Test IPv6 multicast receivers transitioning from IPv4 igmp to IPv6 mld alongside dual-stack pim Sparse-Mode routers.
Firewall ICMPv6 Link-Local Policy Verification
Validate host firewall rules permit RFC 4890 link-local ICMPv6 types 130, 131, and 132 with Hop Limit 1.
Minimal Quickstart (mld_quickstart.frasm)
Hit Preview Packets to compile in-browser and eyeball timestamps + wire byte lengths, or ship it straight to the Playground.
host rtr_v6 { ip fe80::1 }host node_v6 { ip fe80::25 }compile { rtr_v6 -> node_v6 mld query maxrespdelay 10000 wait 3ms node_v6 -> rtr_v6 mld report multicastaddr ff02::1:ff00:25}What the Compiler Calculates So You Don't Have To
- Sets IPv6 Next Header to
58(ICMPv6) and defaults Hop Limit (ttl) to1per RFC 2710. - Encodes ICMPv6 Type
130(query),131(report), or132(done), 16-bitMaximum Response Delay(milliseconds), and 128-bitMulticast Address. - Derives RFC 2464
33:33:xx:xx:xx:xxdestination MACs automatically when sending toff00::/8multicast hosts and computes the IPv6 pseudo-header checksum.
Link-Local MLD General Query, Solicited-Node Join, and Listener Done
Simulates an IPv6 multicast router querying all link-local nodes (ff02::1), a subscriber reporting membership in ff05::100, and cleanly departing with mld done.
host querier_v6 { ip fe80::1c73:1 }host all_nodes { ip ff02::1 }host all_rtrs { ip ff02::2 }host listener { ip fe80::1b21:50 }compile { // 1. General MLD Query to ff02::1 (multicastaddr :: = all groups, 10000ms max delay) querier_v6 -> all_nodes mld query multicastaddr :: maxrespdelay 10000 vlan 200 wait 4ms // 2. Host joins site-local multicast stream ff05::100 listener -> querier_v6 mld report multicastaddr ff05::100 maxrespdelay 0 vlan 200 wait 25ms // 3. Host sends MLD Done to all-routers multicast ff02::2 upon leaving listener -> all_rtrs mld done multicastaddr ff05::100 vlan 200}# 1. Eyeball relative timestamps, ports & wire lengths in your terminal
frameassembly show mld_snooping_ipv6.frasm
# 2. Compile .frasm -> classic .pcap (tcpdump/tcpreplay/Suricata) or nanosecond .pcapng
frameassembly mld_snooping_ipv6.frasm mld_snooping_ipv6.pcap
frameassembly mld_snooping_ipv6.frasm mld_snooping_ipv6.pcapng
# 3. Round-trip .pcap <-> .frasm whenever you want to scrub or tweak an existing capture
frameassembly reverse mld_snooping_ipv6.pcap mld_snooping_ipv6_edited.frasmAll mld Keywords & Header Options
Every sub-keyword and header option is lowercase [a-z0-9]+ with zero underscores — mix and match them in any order after mld:
src_host -> dst_host mld [query | report | done] [multicastaddr <ipv6>] [maxrespdelay <0..65535>]| Keyword / Option | Accepted Value | Default | Effect on Packet |
|---|---|---|---|
| query | report | done | MLD message selector | query (Type 130) | Selects MLD Multicast Listener Query (130), Multicast Listener Report (131), or Multicast Listener Done (132). |
| multicastaddr <ipv6> | 128-bit IPv6 multicast address (or ::) | :: (General Query) | Sets the 16-byte Multicast Address field (:: in General Queries, specific ffxx:: group in Group-Specific Queries, Reports, and Done messages). |
| maxrespdelay <0..65535> | Whole number (0..65535) (milliseconds) | 10000 (10 seconds) | Sets the Maximum Response Delay in milliseconds on MLD Queries (0 on Reports and Done messages). |
Cross-Layer Wrappers Valid with mld
802.1Q VLAN / QinQ / MPLS tags, IPv4/IPv6 QoS & TTL fields, wait gaps, and transport overrides allowed on this protocol
8 options
Cross-Layer Wrappers Valid with mld
802.1Q VLAN / QinQ / MPLS tags, IPv4/IPv6 QoS & TTL fields, wait gaps, and transport overrides allowed on this protocol
| Option | Range / Format | Default | Effect on Packet |
|---|---|---|---|
| wait <number><ns|us|ms|s|m> | Time gap (ns, us, ms, s, m) | 10ms | Advances the packet timestamp clock after sending this packet and before the next packet (wait 0ns sends back-to-back). |
| payload "<text>" | hexpayload "<hex>" | Quoted text or hex bytes (choose one) | Omitted | Appends or overrides raw payload bytes ("deadbeef" or "de:ad:be:ef") and updates packet lengths and checksums. |
| srcmac <mac> | dstmac <mac> | Colon, hyphen, or Cisco dot MAC | Host MAC / RFC multicast MAC | Overrides the 14-byte Ethernet II source or destination MAC address for this single packet. |
| vlan <0..4095> [pcp <0..7>] [dei <0|1>] | VLAN ID 0..4095, PCP 0..7, DEI 0..1 | Untagged | Wraps the packet in an IEEE 802.1Q VLAN tag (0x8100) with optional 802.1p QoS priority. |
| qinq <0..4095> [innervlan <0..4095>] | Outer S-VLAN 0..4095, Inner C-VLAN 0..4095 | Untagged (innervlan 1 when qinq is set) | Double-tags the packet with outer IEEE 802.1ad S-Tag (0x88a8) plus inner IEEE 802.1Q C-Tag (0x8100). |
| mpls <0..1048575> [exp|tc <0..7>] [bos <0|1>] [mplsttl <0..255>] | Label 0..1048575, exp/tc 0..7, bos 0..1, ttl 0..255 | Unlabeled (at most 1 mpls tag per packet) | Inserts a 32-bit MPLS label header (0x8847) ahead of the Layer 3 header. |
| ttl <0..255> [dscp <0..63>] [ecn <0..3>] | TTL/Hop Limit 0..255, DSCP 0..63, ECN 0..3 | ttl 1, dscp 0, ecn 0 | Sets IPv4 TTL or IPv6 Hop Limit plus DiffServ (DSCP) and ECN QoS markings. |
| flowlabel <0..1048575> | 20-bit IPv6 Flow Label (0..1048575) | 0 | Sets the 20-bit IPv6 Flow Label for ECMP routing tests (ignored when running over IPv4 hosts). |
Gotchas & Quick Fixes
IPv6 Endpoints Are Mandatory on mld
host4_a -> host4_b mld queryhost6_a -> host6_b mld query multicastaddr ::MLD runs inside ICMPv6, so both hosts must have IPv6 addresses. Use igmp for IPv4 multicast group management.
Use multicastaddr (Not group) and maxrespdelay in Milliseconds
node_v6 -> rtr_v6 mld report group ff05::100 maxresp 100node_v6 -> rtr_v6 mld report multicastaddr ff05::100 maxrespdelay 1000MLD uses multicastaddr <ipv6> and millisecond-granularity maxrespdelay <0..65535>, whereas IPv4 igmp uses group <ipv4> and 1/10s maxresp <0..255>.