• ↑↓ pour naviguer
  • pour ouvrir
  • pour sélectionner
  • ⌘ ⌥ ↵ pour ouvrir dans un panneau
  • ←→ pour naviguer
  • esc pour rejeter
⌘ '
raccourcis clavier

The Problem

Network layers want to forward packets as fast as possible without the overhead of guaranteeing delivery. Some services (like IP) prioritize speed and simplicity over reliability.

Core Idea

A delivery model where the network makes its “best effort” to deliver packets but provides no guarantees — packets can be lost, duplicated, or arrive out of order.

How It Works

  1. Network layer receives packet from transport layer
  2. Attempts to route and forward packet toward destination
  3. If link is congested or error occurs: packet may be dropped
  4. No acknowledgments, retransmissions, or error recovery at this layer
  5. Higher layers (e.g., TCP) must handle reliability if needed

Visual Explanation

G Sender Sender Network Packet 2 LOST Sender->Network Packet 1 Sender->Network Packet 2 Dest Destination Network->Dest 1 arrives

Key Properties

  • No delivery guarantees
  • No order guarantees
  • No error recovery (just detection via checksums)
  • Minimal overhead — fast and simple

Connections

Edge Cases & Gotchas

  • Higher layers must implement reliability if needed (TCP does this on top of IP)
  • Packet loss is expected and must be handled by applications or transport layer
  • “Best effort” doesn’t mean “no effort” — routers still try their best to forward