Network Layer Addressing

Every device on a network needs an IP address to communicate—but where do these addresses come from? Unlike MAC addresses, which are burned into hardware at the factory, IP addresses must be configured somehow. This configuration can happen manually, automatically, or through a combination of both approaches.

The mechanisms for address assignment have evolved dramatically from the early Internet's manual configuration to today's sophisticated dynamic systems that can configure millions of devices without human intervention. Understanding these mechanisms is essential for network administrators, system engineers, and anyone working with networked infrastructure.

This page explores address assignment comprehensively: from manual static configuration to the DHCP protocol that powers most networks today, from IPv6's stateless autoconfiguration to enterprise address allocation strategies.

Static addressing assigns a fixed IP address to a device through manual configuration. The address is stored in the device's configuration and persists across reboots until manually changed.

Configuration Components:

Static configuration requires setting multiple parameters:

1. IP Address: The device's unique identifier (e.g., 192.168.1.100)
2. Subnet Mask: Defines network/host division (e.g., 255.255.255.0 or /24)
3. Default Gateway: Router for reaching other networks (e.g., 192.168.1.1)
4. DNS Servers: For name resolution (e.g., 8.8.8.8, 8.8.4.4)

When Static Assignment is Appropriate:

Static Assignment Challenges:

While static addressing provides predictability, it introduces significant operational challenges:

1. Manual Tracking Required Every address assignment must be documented and tracked. Without centralized management, conflicts become inevitable as administrators forget which addresses are in use.

2. Configuration Errors Manual entry invites typos. A transposed digit (192.168.1.100 vs. 192.168.1.10) can cause duplicate addresses or unreachable devices.

3. Network Changes are Painful If the network is re-addressed (subnets change, new gateway), every statically configured device must be manually updated.

4. Mobility is Difficult Moving a device to a different subnet requires reconfiguration. Dynamic environments (laptops, phones) are impractical to manage statically.

5. No Automatic Failover If a device's configured DNS server or gateway becomes unavailable, the device has no way to discover alternatives.

The Dynamic Host Configuration Protocol (DHCP) automates address assignment, eliminating the need for manual configuration on most devices. DHCP is defined in RFC 2131 and is ubiquitous in modern networks.

DHCP Components:

DHCP Server: Maintains a pool of available addresses and configuration parameters. Responds to client requests with address offers and manages the lease database.

DHCP Client: Software on the end device that requests configuration from servers. Built into virtually all operating systems.

DHCP Relay Agent: Forwards DHCP traffic between subnets when the server isn't on the local network. Essential for centralized DHCP in multi-subnet environments.

Address Pool (Scope): The range of addresses a server can offer. Administrators configure pools, exclusions, and reservation rules.

Lease: The temporary assignment of an address to a client. Includes an expiration time after which the address may be reclaimed.

DHCP Message Types:

DHCP uses eight message types, but four are essential for address acquisition:

Additional message types handle special cases:

DHCP address acquisition follows the DORA sequence: Discover, Offer, Request, Acknowledge. This four-way handshake ensures reliable, conflict-free address assignment even with multiple DHCP servers on the network.

Step-by-Step Analysis:

Step 1: DISCOVER The client broadcasts a DHCPDISCOVER message. Key fields:

• Source IP: 0.0.0.0 (client has no address yet)
• Destination IP: 255.255.255.255 (broadcast)
• Source MAC: Client's hardware address
• Transaction ID: Random number to match responses

Step 2: OFFER
Servers receiving the DISCOVER select an available address from their pool and respond with DHCPOFFER:

• Offered IP address
• Subnet mask, gateway, DNS, and other parameters
• Lease duration
• Server identifier (server's own IP)

Step 3: REQUEST The client selects one offer (typically the first received) and broadcasts DHCPREQUEST:

• Specifies the requested IP and server identifier
• Broadcast so all servers learn which offer was accepted
• Servers whose offers weren't chosen return those addresses to their pools

Step 4: ACKNOWLEDGE The selected server sends DHCPACK confirming the assignment:

• Final configuration parameters
• Lease time begins now
• Client can begin using the address

DHCP leases are temporary by design. This temporariness serves important purposes: it allows address recovery when devices leave the network, enables configuration updates, and adapts to changing network conditions.

Lease Timeline:

Lease Database:

DHCP servers maintain a lease database recording:

• Client hardware (MAC) address
• Assigned IP address
• Lease start time and duration
• Client hostname (if provided)
• Client identifier (if provided)

This database is persistent—it survives server reboots. Modern DHCP implementations support database replication between redundant servers.

Lease Operations:

Release: When a device shuts down gracefully (e.g., through OS shutdown), it may send DHCPRELEASE to return its address immediately. This is optional—many devices simply let leases expire.

Decline: If a client detects its assigned address is already in use (via ARP probe), it sends DHCPDECLINE. The server marks that address as conflicted and doesn't offer it again without administrator intervention.

DHCP's broadcast-based discovery creates a challenge: broadcasts don't cross routers. In a multi-subnet network, this would require a DHCP server on every subnet—impractical for large networks.

The Solution: DHCP Relay Agents

A DHCP Relay Agent (RFC 3046) is a router or dedicated device that:

1. Listens for DHCP broadcasts on local interfaces
2. Encapsulates them in unicast packets
3. Forwards them to a configured DHCP server on another subnet
4. Receives responses and broadcasts them to the client

This allows a single centralized DHCP server to serve many subnets.

The Gateway Address (giaddr) Field:

When the relay forwards a DHCP message, it sets the giaddr (gateway address) field to the IP address of its interface on the client's subnet. The server uses this to:

1. Determine which address pool to use (e.g., giaddr = 192.168.1.1 → offer from 192.168.1.0/24 pool)
2. Know where to send the response

Multi-Server Architecture:

Enterprise networks typically deploy multiple DHCP servers for redundancy. Two common approaches:

Split Scope (80/20 Rule):

• Each server owns a portion of the address range
• Primary server has 80%, backup has 20%
• If primary fails, backup provides limited addresses
• Simple but wastes 20% capacity during normal operation

DHCP Failover (RFC 3074/7031):

• Servers synchronize lease databases in real-time
• Both servers can offer any address
• Seamless failover without address conflicts
• More complex but efficient

IPv6 introduces fundamentally new address assignment mechanisms, with Stateless Address Autoconfiguration (SLAAC) as the primary innovation. Understanding these mechanisms is essential as networks transition to IPv6.

SLAAC (Stateless Address Autoconfiguration):

SLAAC allows devices to configure themselves without any server:

1. Link-Local Generation: Device creates fe80:: address using its Interface ID
2. Router Discovery: Device sends Router Solicitation; routers respond with Router Advertisements
3. Prefix Learning: RA contains network prefix(es) and flags
4. Global Address Generation: Device combines prefix + Interface ID
5. DAD: Duplicate Address Detection confirms uniqueness
6. Address Activation: Device begins using the address

The 'stateless' aspect means no server maintains a database of assignments—devices self-configure.

Stateful DHCPv6:

For environments requiring centralized address management, logging, or complex policies, stateful DHCPv6 operates similarly to DHCPv4:

1. Client sends SOLICIT (like DISCOVER)
2. Server(s) respond with ADVERTISE (like OFFER)
3. Client sends REQUEST
4. Server responds with REPLY (like ACK)

Key Differences from DHCPv4:

• Uses UDP ports 546 (client) and 547 (server), not 67/68
• Uses ICMPv6-based relay rather than IP helper
• Can operate in stateless mode (parameters only, no addresses)
• Device may have SLAAC and DHCPv6 addresses simultaneously
• No DHCPDECKINE equivalent—DAD handles conflicts

Large organizations require systematic approaches to address allocation, balancing ease of management, security, scalability, and operational efficiency.

Address Space Planning Principles:

1. Leave Room for Growth: Allocate larger blocks than currently needed; address reconfiguration is painful
2. Maintain Summarizability: Assign contiguous blocks to related entities for route aggregation
3. Separate by Function: Different address ranges for servers, users, IoT, guests, management
4. Document Everything: Address assignments, VLAN mappings, DHCP scope configurations
5. Plan for IPv6: Even if not deploying today, structure IPv6 assignments in parallel

Allocation Tools and Practices:

IP Address Management (IPAM) Systems:

• Centralized database of all allocations
• Integration with DHCP and DNS
• Visual representation of usage
• Conflict detection and history tracking
• API integration with provisioning systems

VLAN-Based Segmentation:

• Align address ranges with VLANs for clarity
• Example: VLAN 100 → 10.100.0.0/16
• Simplifies firewall rules and troubleshooting

Naming Conventions:

• Embed location/function in hostnames
• Example: web-prod-dc1-001.example.com
• Enables understanding from DNS records alone

Address assignment is the practical mechanism by which the theoretical addressing scheme becomes operational reality. Let's consolidate the essential concepts:

What's Next:

With addresses assigned, the final challenge remains: address resolution—translating between IP addresses and MAC addresses for actual frame delivery. The next page explores ARP, NDP, and the mechanisms that bridge the logical and physical addressing worlds.