Classful Addressing

Classes A, B, and C all share a fundamental assumption: each IP address identifies a single network interface on a single host. This unicast model works well for direct communication between two parties—web browsing, email, file transfers—but becomes grotesquely inefficient when one source needs to send identical data to many recipients.

Consider streaming a live video to 10,000 viewers. With unicast, the server transmits 10,000 identical streams—overwhelming bandwidth, server resources, and network infrastructure. This is where Class D multicast addresses provide an elegant solution.

Class D addresses (224.0.0.0 – 239.255.255.255) don't identify hosts. Instead, they identify groups. Any number of hosts can join a multicast group, and a single transmission to that group's address reaches all members efficiently—routers replicate packets only where necessary, minimizing bandwidth consumption.

Class D addresses are identified by their leading four bits being '1110'. Unlike Classes A-C, Class D has no network/host division—the entire address identifies a multicast group.

Binary anatomy of a Class D address:

Mathematical derivation of the range:

• Minimum first octet: 11100000 = 224
• Maximum first octet: 11101111 = 239

Therefore, any address with first octet between 224 and 239 is definitively Class D (multicast).

Critical distinction from unicast:

Understanding multicast requires contrasting it with the two other transmission models—unicast and broadcast. Each serves different communication patterns:

The efficiency advantage:

Consider sending 1 Gbps video to 1,000 recipients:

• Unicast: Server sends 1,000 × 1 Gbps = 1 Tbps (impossible for most servers)
• Broadcast: Works but reaches everyone, including uninterested hosts (wasteful, policy issues)
• Multicast: Server sends 1 × 1 Gbps; routers replicate only toward interested recipients

Multicast achieves broadcast's bandwidth efficiency while limiting delivery to interested parties—the best of both worlds.

The multicast address space (224.0.0.0 – 239.255.255.255) is subdivided into ranges with different purposes and scopes. Understanding these ranges is essential for proper multicast deployment:

Scope zones in detail:

1. Link-local (224.0.0.0/24): Packets with TTL=1 stay on the local network segment. Used by routing protocols like OSPF (224.0.0.5, 224.0.0.6) and RIP (224.0.0.9). Routers never forward these.

2. Site-local / Administratively Scoped (239.0.0.0/8): Similar to RFC 1918 private addresses for unicast. Organizations can use these internally without IANA coordination. Never forwarded outside the administrative domain.

3. Global (224.0.1.0 – 238.255.255.255): Internet-wide multicast for applications like IPTV, financial data feeds, or global collaboration tools. Requires multicast routing infrastructure.

IANA maintains a registry of well-known multicast addresses for standard protocols. Network engineers should recognize these addresses when analyzing traffic or configuring multicast routing:

Why these matter in practice:

• 224.0.0.1 (All Hosts): Sending to this address reaches every multicast-capable host on the link. Used for general announcements.

• 224.0.0.5/224.0.0.6 (OSPF): If you see traffic to these addresses, OSPF routing is active. Firewall rules on internal interfaces must permit this.

• 224.0.0.251 (mDNS): Apple Bonjour and Linux Avahi use this for zero-configuration networking. AirPrint, AirPlay, and service discovery depend on it.

• 224.0.0.22 (IGMPv3): Modern multicast group management. If you see this, hosts are actively joining/leaving multicast groups.

How do hosts join multicast groups? How do routers know which segments have interested receivers? IGMP (Internet Group Management Protocol) provides the signaling mechanism that connects hosts to multicast distribution trees.

IGMP versions and evolution:

IGMP operation overview:

When a multicast packet reaches a local network, it must be encapsulated in an Ethernet frame. Unlike unicast (which uses ARP to resolve IP to MAC), multicast uses a direct mathematical mapping from IP multicast address to a special MAC address range.

The mapping algorithm:

The 23-bit overlap problem:

IP multicast addresses have 28 bits of group identification, but only 23 bits map to the MAC address. This means:

2^28 / 2^23 = 2^5 = 32 IP multicast addresses map to each MAC address

For example, these 32 addresses all map to the same MAC:

• 224.0.1.1, 225.0.1.1, 226.0.1.1, 227.0.1.1, ..., 239.0.1.1

Implication: If a host joins 224.0.1.1, its NIC will receive packets for all 32 overlapping addresses. The IP layer must filter unwanted groups. This rarely causes problems in practice, but explains occasional 'extra' multicast traffic.

Multicast enables applications that would be impractical or impossible with unicast. Here are the most significant real-world uses:

The Internet multicast gap:

Despite its efficiency, Internet-wide multicast never achieved widespread deployment. Most ISPs don't enable multicast routing, and NAT devices complicate group membership. As a result, Internet video streaming (YouTube, Netflix) uses unicast with CDN caching rather than true multicast. Multicast thrives in controlled enterprise networks where administrators can enable and manage it end-to-end.

Traditional multicast (called Any-Source Multicast or ASM) allows any host to send to a group. This creates security and engineering challenges: who can send? How do you prevent unauthorized sources?

Source-Specific Multicast (SSM) addresses these issues by letting receivers specify both the group AND the source:

SSM channel format: (Source, Group) = (S, G)

SSM operation:

1. Receiver knows the source (e.g., from a TV guide or application)
2. Receiver sends IGMPv3 report: "Join (10.1.1.100, 232.1.1.1)"
3. Router builds shortest-path tree directly from source to receiver
4. Only packets from 10.1.1.100 to 232.1.1.1 are delivered

Benefits:

• No rendezvous point infrastructure required
• Inherent source authentication (only specified source accepted)
• Simpler, more scalable routing
• Shorter join latency (direct path, no RP detour)

Class D represents a fundamentally different addressing paradigm—not host identification, but group communication. This one-to-many model enables applications impossible with unicast efficiency. Let's consolidate the key concepts:

What's next:

With Class D (Multicast) complete, we'll explore the final category: Class E (Reserved/Experimental) addresses in the 240-255 range. You'll understand why this large block remains unusable, the historical reasons for its reservation, and ongoing debates about reclaiming it for the exhausted IPv4 pool.