AP Networking 4.4: How Data Travel the Internet

Every card payment, price update, and web request that leaves a network crosses a chain of routers, each picking the next step from a routing table. This guide shows how a host decides a destination is remote, how NAT gets it out the door, how metrics choose between paths, and how to read traceroute when something breaks.

The short answer: a device compares its own IP address and subnet mask against the destination; if the masked networks differ, the packet goes to the default gateway. Each router then matches the destination against its routing table, prefers the most specific route, breaks ties with the lowest metric, and forwards the packet one hop closer. Traceroute shows the whole chain, one line per hop. All of Topic 4.4 runs on the grocery store network of scenario 4A.

• How a packet leaves the store network
• NAT: one public address for the whole store
• Routing tables: turn-by-turn directions
• Routing protocols and metrics
• Reading traceroute like an analyst
• Worked walkthrough: the Saturday outage
• Practice: Path Finder
• Sample practice questions
• Key terms
• Frequently asked questions

How a packet leaves the store network

A register terminal at 192.168.50.23 sends a charge to the payment processor at 192.0.2.80. First the terminal runs one test: it applies its subnet mask, 255.255.255.0, to its own address and to the destination. Its masked network is 192.168.50.0; the destination's is 192.0.2.0. They differ, so the destination is remote, and the terminal hands the packet to its default gateway, the store router at 192.168.50.1.

That handoff is the first hop. A hop happens every time a packet is forwarded from one router to the next; after the gateway, the hops usually belong to ISPs. At the destination network, the last router delivers the packet by its private address. Watch the trap: 10.20.30.7 and 10.20.31.9 share two octets and still sit on different /24 networks. The mask decides.

NAT: one public address for the whole store

The register's 192.168.50.23 is a private address, unreachable from the internet. As the packet leaves, the router performs network address translation: it replaces the private source address with its own public address, 203.0.113.10. The processor replies to 203.0.113.10, and the router translates that destination back to 192.168.50.23. Every register, scale, and office PC shares that one public address.

NAT conserves public IPv4 addresses, since a building full of devices needs only one, and it adds a layer of security, because outside hosts cannot address an internal device directly. Note the direction: NAT rewrites the source on the way out. Destinations are only rewritten on the return trip.

Routing tables: turn-by-turn directions

At each router, the packet meets a routing table: a list of known networks and the best path to each. Here is the store router's table, the artifact this lesson turns on:

STORE-R1# show ip route
Destination        Next hop            Interface   Metric
192.168.50.0/24    directly connected  lan0        0
172.16.0.0/16      203.0.113.1         wan0        5
0.0.0.0/0          203.0.113.1         wan0        10
0.0.0.0/0          198.51.100.1        lte0        50

Each entry carries four things: the destination network's address and mask, the next-hop address, the interface that reaches it, and the metric, that route's cost. Reading it is a fixed procedure:

1. Match the destination. 192.0.2.80 falls inside neither listed network, so only the two default routes (0.0.0.0/0, "anywhere else") match.
2. Prefer the most specific route. HQ traffic to 172.16.4.20 matches the /16 route and the defaults; the /16 wins.
3. Break ties with the lowest metric. Two defaults match 192.0.2.80. Metric 10 beats 50, so the packet leaves for 203.0.113.1 on fiber, not the LTE backup.
4. Forward and repeat. The next router runs the same procedure on its own table, hop after hop.

Routing protocols and metrics

Who writes the table? An administrator can enter routes statically, but at internet scale routers learn them through routing protocols, which define how routers share what they know and score paths. Each protocol computes its metric its own way, so two can honestly disagree.

Picture two paths to the processor: two hops over 10 Mbps microwave, or four hops over 10 Gbps fiber. RIP takes the two-hop path because it only counts routers; OSPF takes the fiber because it weighs bandwidth. Neither is broken. And because metrics track congestion and failures, the winning path can change during the day. This morning's route is not a promise about this afternoon.

Reading traceroute like an analyst

You can watch the path itself: tracert on Windows, traceroute on Mac and Linux list every hop to a destination with its response time. A healthy trace from the back office to the payment processor:

C:\Users\intern>tracert 192.0.2.80

Tracing route to paygate.example [192.0.2.80]
over a maximum of 30 hops:

  1     1 ms  192.168.50.1
  2     8 ms  203.0.113.1
  3    12 ms  isp-a-core1.example.net [203.0.113.65]
  4    14 ms  ix-peer2.example.net [198.51.100.201]
  5    18 ms  192.0.2.1
  6    19 ms  paygate.example [192.0.2.80]

Trace complete.

Six hops, all under 20 ms; the destination answers at hop 6. The "maximum of 30 hops" line is a probe limit, not a measurement. The same command on Saturday, while card readers time out:

  1     1 ms  192.168.50.1
  2     9 ms  203.0.113.1
  3    12 ms  isp-a-core1.example.net [203.0.113.65]
  4     *     Request timed out.
  5     *     Request timed out.
  6     *     Request timed out.

Read it top down. Hops 1 through 3 answer fast, so the store network, the router, and the ISP handoff are healthy. The silence starts after isp-a-core1: the failure sits beyond the ISP's core router, outside the store. The intern's move is a call to the ISP with this output, not an hour of rebooting registers. The same reading finds slowdowns: times jumping from 12 ms to 190 ms at hop 4 mean the delay begins there.

Worked walkthrough: the Saturday outage

Saturday, 11:40 am: prices display wrong, payments fail, and the manager fears an attack. You run the process instead.

1. Trace the payment path. tracert 192.0.2.80 breaks after hop 3, as above. Availability problem, beyond the ISP handoff. Evidence beats fear.
2. Check the router's view. When wan0 drops, the fiber routes via 203.0.113.1 are withdrawn and the metric-50 LTE default takes over. Traffic still flows, just slower.
3. Explain the wrong prices. HQ price syncs ride the 172.16.0.0/16 route over fiber. The morning trace ran through 203.0.113.1 at 8 to 19 ms; the afternoon trace reaches HQ through 198.51.100.1 at 41 to 75 ms. New path, slower sync: the prices are queued, not altered.
4. Verify after the fix. When ISP-A restores the fiber, routes return and a fresh trace shows the old path under 20 ms. Payments clear; the sync catches up.

Practice: Path Finder

Now read the artifacts yourself: decide where packets go, reroute around a dead link, and find where a path broke. Cases and choices shuffle every run, so the evidence is the only way through.

Sample practice questions

Three free questions in the graded quiz's style. Commit before you expand the explanation.

Key terms

Default gateway

The router a host sends to when the destination is on a different network.

Hop

One forward of a packet from one router to the next.

NAT (network address translation)

The gateway's swap of a private source address for its public address, letting many devices share one public IP.

Routing table

A router's list of known networks with the next hop, interface, and metric for each.

Next hop

The neighboring router a packet is forwarded to for a given destination.

Metric

The cost of a route, such as hop count or time. Lowest wins.

Routing protocol

Rules routers use to share reachability and pick paths: RIP, OSPF, BGP, IS-IS.

traceroute / tracert

Command line tools that list every hop to a destination with its response time.

Frequently asked questions

AP is a trademark of the College Board, which was not involved in the production of, and does not endorse, this resource.