4.2DHCP, DNS, and NTP

Manually configuring an IP address on every device is impractical, so in practice we distribute them automatically via DHCP. We can reach hosts by name because DNS resolves hostnames to IPs behind the scenes, and we can correlate log timestamps because NTP synchronizes device clocks. CCNA tests whether you can, for these three "under the hood" services, judge configuration and fault isolation grounded in how they work. DHCP relay in particular is a frequent real-world trouble spot.

4.2.1DHCP DORA operation and the relay agent

• DHCP leases an address between client and server via the four DORA messages (Discover -> Offer -> Request -> Ack). The initial Discover is sent as a broadcast because the destination is not yet known (the server uses UDP 67, the client UDP 68).
• Because a router does not forward broadcasts to other segments, if the DHCP server is on a different subnet from the client, the Discover never reaches it. Bridging this is the job of the DHCP relay agent: configuring ip helper-address <server-IP> on the client-facing interface makes the router convert the received DHCP broadcast into a unicast destined for the server and forward it.
• To make a Cisco router itself a DHCP server, define ip dhcp pool <name> -> network / default-router / dns-server, and exclude addresses you do not want to hand out (such as the router's IP) with ip dhcp excluded-address.

4.2.2The role of DNS

• DNS is the mechanism that resolves a hostname such as www.example.com into an IP address. A device receives the address of the DNS server to query (often via DHCP) and converts names to IPs for each communication.
• A key isolation cue: if ping 8.8.8.8 (a raw IP) succeeds but ping www.example.com (a name) fails, you can conclude that IP reachability is fine and only DNS name resolution is failing. On Cisco devices, specify the DNS server with ip name-server and enable resolution with ip domain-lookup.

4.2.3NTP and stratum

• NTP synchronizes device clocks to an accurate reference time. Distance from the time source is expressed as a hierarchical value called stratum: the reference clock (GPS/atomic) is stratum 0, a server directly attached to it is stratum 1, and the number increases by one per hop thereafter. A lower stratum value means closer to the time source and higher trustworthiness (16 means unsynchronized).
• To make a Cisco device an NTP client, simply configure ntp server <IP>. Verify synchronization with show ntp status (shows "Clock is synchronized" and the device's own stratum) or show ntp associations. Accurate time is essential for Syslog, digital-certificate validation, and log correlation.

Suppose you operate an L3 switch (router) aggregating several subnets, distributing IPs company-wide from a single DHCP server in the head-office server room. One day, only the PCs on a newly added Sales segment (10.20.0.0/24, a different subnet from the server) show "cannot obtain an IP address." A PC on that segment configured manually can ping the DHCP server, so the path (L3 reachability) is fine. Automatic acquisition still fails because DHCP's initial Discover is a broadcast, and a router/L3 switch does not forward broadcasts to a DHCP server on another segment. That other departments' existing segments get addresses fine while only the new segment fails strongly suggests a forgotten ip helper-address <DHCP server IP> on the new segment's gateway interface. Applying ip helper-address 10.10.1.5 (the server's IP) to the Sales SVI makes the router convert the received DHCP broadcast into a unicast to the server and relay it, so DORA completes and addresses are handed out. Thinking "ping to the server works, so DHCP must reach it too" is wrong here—the crux of isolation is the behavioral difference that ping is unicast L3 reachability, while a DHCP Discover is a broadcast that will not cross an L3 boundary without a relay. Likewise, if a site's clock drifts and Syslog cannot be correlated, check whether that device has ntp server configured and whether show ntp status reports synchronized.

4.2.4Section summary

• DHCP leases addresses via DORA. Since Discover is a broadcast, a server on a different subnet requires an ip helper-address relay (converting the broadcast into a unicast)
• DNS resolves hostnames to IPs. If raw-IP ping works but name ping fails, reachability is fine and DNS resolution is the problem
• NTP synchronizes time, and a lower stratum is closer to the source. The client uses ntp server <IP>; check state with show ntp status