What are the two fundamental approaches to handling mobility in networks, and how do they differ?

Approach 1 (Network-based) lets routers handle mobility via routing tables; Approach 2 (End-system) uses agents at the network edge to redirect traffic.

Approach 1: Network-based mobility

• Routers announce the location of every mobile device via standard routing table exchanges
• Routers know the permanent address (e.g., 32-bit IP or Cell-ID) of each mobile node
• Routing tables show where every device is
• No changes needed on end systems — it's all handled in the network core
• Problem: doesn't scale — millions of mobile devices would explode routing table sizes

Approach 2: End-system mobility

• Uses additional functionality and agents (home agent, foreign agent) at the "edge"
• Two sub-variants:
  • Indirect routing — all traffic goes through the home network first, then gets forwarded to the visited network
  • Direct routing — the correspondent learns the care-of address and sends directly to the visited network

Modern mobile networks (4G/5G) use Approach 2 because the network approach simply can't handle billions of devices.

Go deeper:

• Mobile IP (Wikipedia) — contrasts the home-agent/foreign-agent edge approach with tunneling, and why pure routing-table mobility doesn't scale.