Understanding the Network Layer: From Packet Forwarding to IPv6 Migration

Network Layer Overview

The network layer is the third layer of the OSI model, situated between the transport and link layers, and its main purpose is to provide transparent host‑to‑host packet delivery.

It achieves this by performing two essential functions: forwarding (moving a packet from an input link to the appropriate output link) and routing (selecting the path a packet should take).

Two planes exist in a network: Data plane – responsible for forwarding traffic. Control plane – decides how traffic should be routed.

Router Operation

A router contains four main components:

Input port : handles line‑termination, data‑link processing, and lookup/forwarding functions.

Switching fabric : connects input ports to output ports, acting as the router’s internal network.

Output port : forwards packets onto the physical and data‑link layers.

Routing processor : runs routing protocols, maintains routing tables, and performs network‑management tasks.

Each input port holds a local copy of the forwarding table, enabling distributed forwarding without constantly consulting the routing processor.

Forwarding Table and Longest‑Prefix Matching

Routers use a forwarding table that maps destination prefixes to output interfaces. When a packet arrives, the router performs a longest‑prefix match; the longest matching entry determines the outgoing link.

Switching Fabric Types

Memory‑based switching – packets are copied to memory via the CPU before being sent to the output port.

Bus‑based switching – packets are placed on a shared bus with a label; the appropriate output port removes the label.

Network‑on‑chip (non‑blocking) switching – a cross‑bar matrix opens only the required intersection, allowing simultaneous independent transfers.

Queueing and Congestion

Both input and output ports maintain queues. When the switching fabric cannot keep up with incoming traffic, packets wait in the input queue (head‑of‑line blocking) or the output queue (buffer overflow). Common drop policies include tail‑drop and drop‑tail, which affect TCP congestion control.

Packet Scheduling

Routers schedule packets for transmission using various disciplines:

FIFO (first‑in‑first‑out) – simple queue order.

Priority queuing – packets are placed in separate priority classes; higher‑priority queues are served first.

Weighted round‑robin (WRR) – cycles through classes, providing service to each in turn while remaining work‑conserving.

IP Protocol Basics

The IP protocol provides host addressing and routing. It consists of three major functions: addressing, routing, and packet fragmentation/reassembly.

IP Addressing

IPv4 uses 32‑bit addresses written as four decimal octets (e.g., 156.197.1.1). The address space is divided into classes A, B, C, D (multicast), and E (reserved). Subnet masks indicate which bits belong to the network portion.

IPv4 Header Fields

Version

(4 bits) – value 4 for IPv4. Header Length – number of 32‑bit words. DSCP and ECN – service and congestion notification. Total Length – size of the entire datagram. Identification, Flags, Fragment Offset – support fragmentation. TTL – hop limit to prevent loops. Protocol – indicates the encapsulated transport protocol. Header Checksum – error detection for the header. Source and Destination addresses – 32‑bit IPv4 addresses. Options – optional extensions.

When a datagram exceeds the link‑layer MTU, routers may fragment it; reassembly occurs at the destination host.

IPv6 Enhancements

IPv6 expands the address space to 128 bits, removes the header checksum, and relocates fragmentation fields to an optional extension header. New fields include: Traffic Class (similar to DSCP). Flow Label – identifies packet flows for QoS. Payload Length, Next Header, Hop Limit.

Source and Destination addresses – eight 16‑bit groups separated by colons.

IPv6 also mandates built‑in IPSec support, stateless address autoconfiguration, and hierarchical address allocation (IANA → RIR → ISP).

Transition from IPv4 to IPv6

Because IPv4 address space is exhausted, networks must run both protocols simultaneously. The most practical migration technique is tunneling, where IPv6 packets are encapsulated inside IPv4 packets to traverse IPv4‑only segments.

Tunneling creates a virtual link between IPv6 endpoints; the outer IPv4 header carries the packet across the existing IPv4 infrastructure, while the inner IPv6 packet is extracted at the tunnel exit.

Summary

The network layer provides end‑to‑end packet delivery through forwarding, routing, and address management. Routers implement these functions with input ports, switching fabrics, output ports, and routing processors, while handling queueing and scheduling to avoid congestion. Understanding IPv4 header structure, address classes, and fragmentation is essential, and IPv6 introduces a larger address space, simplified headers, and built‑in security. Migration strategies such as tunneling enable a gradual transition to IPv6.