What you'll learn today

Six interactive checkpoints along the way — quizzes, a DORA stepper, a DNS visualizer, and more.

Why every host needs an IP

• An IP address is the logical identity of a device on a network — without one, the device cannot send or receive any IP traffic.
• Every laptop, phone, server, printer, IP camera, smart bulb, and IoT sensor needs one.
• Two big questions for every device:
  • Who decides which IP it gets?
  • Does that IP stay the same over time?
• Those two questions lead us to Static vs Dynamic assignment.

Static vs Dynamic — at a glance

When to use a Static IP

Use a static IP when other devices need to find this device by a fixed address.

• Servers — web, file, database, DNS, DHCP itself
• Network printers — so the print queue always reaches them
• Routers and switches — for management access
• Wireless access points — for the controller to reach them
• Security cameras / NVR — for stable recording

When to use a Dynamic IP

Use dynamic for any device that mostly initiates connections (a "client") and doesn't need a fixed address.

• Laptops, desktops, phones, tablets
• IoT and smart-home devices
• Guest Wi-Fi clients
• Anything that comes and goes from the network frequently

The DHCP server hands out an address, default gateway, subnet mask, and DNS server in one go — so the device "just works" the moment it joins the network.

Imagine configuring 200 phones in an office by hand… 😩

What is DHCP?

• Dynamic Host Configuration Protocol.
• A service running on a server (or your home router!) that hands out IP addresses to devices that ask for them.
• Also delivers everything else a device needs to work: subnet mask, default gateway, and DNS server(s).
• Each handout is called a lease and lasts a configurable amount of time.
• Uses UDP ports 67 (server) and 68 (client).

The DORA process

When a client joins a network, it negotiates a lease with the DHCP server in four messages:

Just remember: Discover → Offer → Request → Ack. We'll walk through each on the next slides.

D — DHCP Discover

• The client just woke up on the network. It has no IP yet.
• It can't address a specific server, so it sends a broadcast packet.
• Source IP: 0.0.0.0   Destination: 255.255.255.255
• Includes the client's MAC address so a server can reply directly to it.
• Every device on the LAN sees this packet — but only DHCP servers act on it.

O — DHCP Offer

• One or more DHCP servers receive the Discover and reply with an Offer.
• The Offer contains a candidate IP address for the client, plus subnet mask, gateway, DNS, and lease length.
• It's still a broadcast (the client has no IP yet), but it's addressed to the client's MAC.
• If multiple DHCP servers exist, the client may receive multiple offers — it will accept just one.

R — DHCP Request

• The client picks one offer (usually the first to arrive) and announces its choice.
• It sends a Request — still as a broadcast, so all DHCP servers learn which offer was accepted.
• Servers whose offers were not chosen can release the IP they reserved back into the pool.
• The client says, in effect: "I want 192.168.1.42 from server 192.168.1.5."

A — DHCP Acknowledge

• The chosen server confirms with a final Acknowledge packet.
• This officially commits the lease: the IP is now reserved for that client's MAC for the lease duration.
• The client now has a full configuration: IP, subnet mask, gateway, DNS, lease time.
• The client can begin using the network normally.

That's the full D-O-R-A!

APIPA — when DHCP fails

• If a client sends Discovers and gets no Offer (server down, cable unplugged, wrong VLAN…), it doesn't sit there with no IP forever.
• It picks a random address from 169.254.0.0/16 — the Automatic Private IP Addressing range.
• That lets it talk to other APIPA-using hosts on the same wire — but it cannot reach the gateway, so no internet.
• Seeing 169.254.x.x in ipconfig is your big red flag: "DHCP isn't working."

What's actually in a DHCP scope?

A "scope" is everything the server is configured to hand out:

Home router vs Enterprise DHCP server

Same protocol — very different scale.

Case study — small office of 50 users

You're designing the IP plan for an office on the 10.20.30.0/24 network.

Static for "infrastructure," dynamic for "people." Reservations bridge the gap.

Why we need DNS

• Computers route on IP addresses like 142.250.190.46.
• People remember names like google.com.
• We need a system that translates one to the other — that's DNS: the Domain Name System.
• Think of it as the phone book of the internet — you look up a name, you get back a number.
• Without DNS, every URL you typed would have to be a raw IP address — and IPs change!

What DNS does — in one sentence

DNS turns a human-readable name into an IP address (and back).

DNS uses UDP port 53 (and TCP 53 for larger replies / zone transfers).

The DNS hierarchy

• DNS is a distributed, tree-shaped database. No single server knows everything.
• Root servers — 13 logical clusters, "." — they know who runs each TLD.
• TLD servers — Top-Level Domains: .com, .org, .net, .uk, … — they know who runs each domain inside their TLD.
• Authoritative servers — own the actual records for a specific domain like example.com.

Walk-through: resolving www.example.com

1. Your laptop asks its local resolver (often your router or your ISP's DNS) for www.example.com.
2. Resolver checks its cache. If it's there and still fresh — done.
3. Otherwise, the resolver asks a root server: "Who handles .com?" → root replies with the IP of a .com TLD server.
4. Resolver asks the .com TLD server: "Who handles example.com?" → TLD replies with the authoritative server for example.com.
5. Resolver asks the authoritative server: "What's the A record for www.example.com?" → it replies with the IP.
6. Resolver caches the answer (with a TTL) and hands the IP back to your laptop.
7. Your laptop opens a TCP connection to that IP. Done!

All of that usually happens in well under 100 ms.

Common DNS record types

A and AAAA records

• A record — maps a hostname to one IPv4 address.
• AAAA record — same idea, but for IPv6.
• A name can have multiple A records → simple load balancing (DNS round-robin).

Sample zone file entries:

www.example.com.   IN  A     93.184.216.34
www.example.com.   IN  AAAA  2606:2800:220:1:248:1893:25c8:1946
api.example.com.   IN  A     203.0.113.10
api.example.com.   IN  A     203.0.113.11

The bottom two together = round-robin between two API servers.

CNAME records

• Canonical NAME: an alias from one name to another name.
• Resolver follows the chain until it lands on an A or AAAA record.
• Useful when you want www.example.com and example.com to resolve to the same place — change the IP once, both follow.
• Also common with hosting providers: myshop.com → myshop.shopify.com.
• Rule: a name with a CNAME can't have any other records (except DNSSEC).

www.example.com.   IN  CNAME  example.com.
example.com.       IN  A      93.184.216.34

MX records (mail)

• MX = Mail eXchanger — names the mail server(s) for a domain.
• Multiple MX records can coexist, each with a priority (lower number = tried first).
• The MX value is always a hostname, never a bare IP.

Sending server tries priority 10 first; falls back to 20 if it's unavailable.

Key takeaways