The protocols that glue a local network together were designed for a room full of colleagues. ARP and DHCP believe whoever answers first, and nothing in either protocol proves who sent a message. This chapter is about what that trust costs, and the switch, Wi-Fi and encryption features that take it back.
Common mix-up: an attacker on your local network does not need to break encryption to cause harm, and encryption does not stop them from redirecting traffic. What encryption does is make the redirected traffic useless to them: they see ciphertext and cannot forge the certificate that would let them read it.
Read first: this chapter assumes how switches learn MAC addresses from Ethernet, MAC & switches, how a device gets its address, gateway and DNS server from DHCP & DNS, and the layout of a typical home network.
On a local network, a device that wants to reach the gateway asks "who has this IP address?" and remembers whichever answer comes back. A device that needs an address asks for a DHCP server and takes the first offer. Neither protocol carries a signature. That is why MITRE ATT&CK lists both ARP cache poisoning and DHCP spoofing as ways an intruder already inside a network can put itself on the path between other devices and the gateway.
The defenses work at three levels:
Detection helps too: IPv4 Address Conflict Detection (RFC 5227) has hosts probe before they use an address and defend it afterwards, and a sudden change in which MAC address answers for the gateway is a classic thing to alert on (Chapter 17).
This chapter is being finished. Still to come are a narrated walkthrough, an ARP cache simulator where you can switch on each defense and see which ones stop redirection and which only make it useless, a DHCP race between a real and a rogue server, and full sections on 802.1X, WPA3 and evil-twin access points.
Firewall zones and VLANs are in Chapter 05; DNS spoofing at the resolver is in Chapter 07; VPNs and zero trust are in Chapter 11.