Below is a real world example of a multicast flow. 

​

Multicast Packet Flow (Simplified Diagram)

Step-by-step flow:

1. Source sends traffic to multicast group (e.g. 239.1.1.1)
2. Routers build a distribution tree (via PIM)

3. Hosts join using IGMP (IPv4) or MLD (IPv6)

   • Host1, Host2, Host3 → JOIN
   • Others → no traffic

4. Traffic is forwarded only where needed

   • Router A sends one stream, not multiple copies
   • Replication happens closer to receivers

Key insight: duplication happens in the network, not at the source

Scenario

A telecom operator delivers live TV channels using multicast.

Components

• Video Headend → Source of streams
• Core Network → Runs PIM-SM
• Access Network (DOCSIS/DSL/FTTH)
• Set-Top Box (STB) → Receiver

Flow Breakdown

Step 1: Channel Streaming

• Headend sends:

  Channel 1 → 239.10.10.1
  Channel 2 → 239.10.10.2

Step 2: User Changes Channel

• User selects Channel 1

• STB sends:

  IGMP JOIN 239.10.10.1

Step 3: Network Reaction

• Access router receives IGMP JOIN
• Signals upstream via PIM Join
• Multicast tree is extended toward the user

Step 4: Traffic Delivery

• Stream starts flowing:

  Headend → Core → Access → STB

• If multiple users watch the same channel:
  👉 Only one stream per link, replicated near users

Step 5: Channel Change

• User switches channel:

  IGMP LEAVE 239.10.10.1
  IGMP JOIN 239.10.10.2

• Network:

  • Stops old stream (prune)
  • Starts new stream

• IGMP = last-mile control (user joins/leaves)
• PIM = backbone distribution
• Efficiency comes from shared streams
• Fast channel change depends on IGMP + network tuning