The Fat Pipe - Most Popular Packet Pushers Pods · Packet Pushers

N4N052: Multicast Part 2

April 2, 2026·1 hr 25 min·5 clips
IPv4 multicast uses 28 bits, but Ethernet multicast gets only 23 bits.
1. The Fat Pipe - Most Popular Packet Pushers Pods episode “N4N052: Multicast Part 2” continues the multicast discussion from episode 50. 2. Ethan Banks and Holly Metlitsky host Lenny Giuliano, described as a senior distinguished systems engineer at HPE Juniper Networks, because he fills in multicast details from the earlier episode. 3. The episode asks what PIM, IGMP snooping, IPv4 and IPv6 multicast, and SSM are for, and why older ASM still matters. 4. Lenny explains that PIM means protocol independent multicast because it uses the unicast routing table as the RPF table. 5. He contrasts PIM with DVMRP, which had its own unicast routing protocol, and says protocols like BGP, OSPF, and ISIS can now fill that role. 6. Lenny says multicast on a LAN is treated like broadcast, so it floods to all ports unless something smarter intervenes. 7. IGMP snooping makes switches inspect IGMP messages so they can forward multicast only to ports with interested hosts. 8. He gives the concrete example of a LAN with 100 hosts where only two receivers get the traffic when IGMP snooping is enabled. 9. Lenny and Ethan then shift to multicast addressing and the Layer 2 and Layer 3 mismatch. 10. Lenny says IPv4 multicast uses 224.0.0.0/4, while Ethernet multicast uses only 23 bits, creating a 32-to-1 oversubscription. 11. He says that means one multicast MAC address can map to 32 different IPv4 multicast addresses. 12. The practical guidance he gives is to pick multicast addresses at random rather than using simple patterns like .1 or consecutive numbers. 13. The conversation then compares IPv4 and IPv6 multicast behavior. 14. Lenny says IGMP has versions 1, 2, and 3, while IPv6 uses MLD, with MLDv1 aligning to IGMPv2 and MLDv2 aligning to IGMPv3. 15. He says PIM works for both IPv4 and IPv6 without a separate PIMv6 protocol. 16. Lenny argues that IPv6 multicast reduced the need for MSDP-style inter-RP communication by using embedded RP addresses so the RP can be derived from the group address. 17. The tone stays conversational and explanatory, with Ethan and Holly pressing for simpler explanations while Lenny answers with historical detail and operational context. 18. The format is a host-led technical interview with frequent back-and-forth, side comments, and short clarifications for newer listeners. 19. Listeners who work with multicast, routing, or enterprise switching will get the most from it. 20. Listeners looking for light, non-technical audio will probably skip it.

As heard by us

A careful multicast follow-up that connects source discovery, application behavior, and IGMPv3 reports.

Ethan Banks and Holly Metlitsky return to multicast with Lenny Giuliano, a senior, distinguished systems engineer at HPE Juniper Networks, for a second pass at ideas left open in their earlier conversation.

Read the full review in PlayNext →

Why you'd press play

You want multicast source discovery explained at the point where the application tells the host both the group address and the source address.

Read the full recommendation in PlayNext →
Listen to the show on