Also: DHCP, dynamic addressing, DHCP lease, address pool, IP reservation
DHCP is the protocol by which a device joining a network asks for and receives an address, a subnet mask, a default gateway and DNS servers from a server, for a limited lease period, so that nobody has to configure addresses by hand.
Our take. DHCP is the protocol that makes networks usable and the one nobody thinks about until it fails, at which point every symptom looks like something else. A room full of devices with self-assigned addresses is a DHCP failure wearing a hundred disguises, and the engineer who checks the lease first saves an afternoon.
A device that joins a network has no address and cannot ask for one by address, so it broadcasts. Every device on the segment hears the request; a DHCP server answers with an offer, the device requests the offered address, and the server acknowledges it. Four messages, and the device now has an address, a subnet mask, a default gateway and the addresses of DNS servers, all of which it would otherwise need typed in. The address is a lease with an expiry, renewed automatically while the device stays, and returned to the pool when it leaves, which is how a network with a few hundred addresses serves thousands of devices over a month.
The pool: the range of addresses the server may hand out, usually most of the subnet with the infrastructure addresses excluded.
The lease time: how long an address is held before renewal. Short for guest networks where devices come and go, long for offices where they stay.
Reservations: a fixed address tied to a device's hardware address, so a printer always gets the same one without being configured by hand.
Options: everything else the server can hand out, from DNS servers to the address of a boot server for devices that load their software over the network.
Relay: since the request is a broadcast and broadcasts stop at the subnet edge, a router forwards them to a central server so one server can serve many subnets.
In practice
Half the devices on an office floor lose their connection on a Monday morning. Each shows a link-local address, the one a device gives itself when no server answered, so the network is up and the address server is not reachable from that floor. The other floors are fine, which points at the relay on that floor's router rather than the server. The relay had been removed during a weekend change. Every user's symptom was a website that would not load, a printer that could not be found, a call that would not connect; every one of them was the same four-message exchange failing before it began, and the address on any one laptop said so.
Check the address first. A link-local address means the exchange failed; a valid address means the problem is elsewhere.
Check the relay when one subnet fails and others work.
Check the pool when only new devices fail: the pool is exhausted, and shortening the lease or widening the pool fixes it.
DNS turns names into addresses; DHCP hands out the addresses themselves, along with the DNS servers to use. A device with a working address and no DNS can reach sites by number and not by name.
Addressing is the scheme; DHCP is the automation that assigns addresses from it. Every address DHCP hands out is an ordinary address that could have been typed in by hand.
Translation rewrites addresses at the network edge for traffic leaving; DHCP assigns them inside the network to devices arriving. Both are usually done by the same home router, and they are unrelated functions.
Key takeaways
→Discover, offer, request, acknowledge: four messages give a device its address, mask, gateway and DNS.
→A link-local address on a client means the exchange failed; check that before anything else.
→Servers and infrastructure get static addresses or reservations; clients get leases.
Certifications that test this
Vendor exams whose syllabus covers this concept — facts, cost and a preparation path on each page.
A rotating selection from the course directory, drawn from the subcategories where this concept is taught rather than picked for it. Details, price and the provider link are on the course page.
>>> Start Sending Data Over Long Distance using Arduino via Wired and Wireless Connection and extend Arduino Capabiliti…
Udemy
FAQ
Should servers use DHCP?
Servers and network infrastructure should hold fixed addresses, either configured statically or reserved on the server by hardware address, because other systems refer to them by address and a change breaks those references. Client devices should lease, because nobody refers to a laptop by its address.
What happens when the lease expires?
The device renews it, normally halfway through the lease and without anyone noticing. If the server is unreachable at renewal, the device keeps the address until the lease ends, then loses it and falls back to a self-assigned one, which is when users notice.
Why is DHCP a security concern?
Because any device on the segment can answer a discovery broadcast, a rogue server can hand out its own gateway and DNS addresses and route every new device's traffic through an attacker. Managed switches can restrict which ports may act as a DHCP server, and the security exams ask about exactly that control.
Sources
The primary text this definition rests on. Read it before you trust ours.
