Lesson 2 — NAT and PAT Fundamentals
Week 6 outline
- Week 6 overview
- Lesson 1DHCP, DNS, and NTP Basics
- Lesson 2NAT and PAT Fundamentals (current step)
- QuizAddressing Services & NAT
- Lesson 3IPv6 Addressing and Dual-Stack
- Lesson 4IPv6 Static Routing
- QuizIPv6 Addressing & Routing
- Lesson 5Network Management Services: SNMP, Syslog, TFTP/FTP, QoS
- CheckpointWeek 6 Checkpoint
NAT and PAT Fundamentals
Configure static NAT, dynamic NAT, and PAT overload; mark inside and outside interfaces by role; read show ip nat translations field by field; and isolate a NAT failure to interface role, ACL, or routing.
Lesson orientation
What you'll learn (5 objectives)~59 min: video → lesson → check → apply → lab prep
Learning objectives
- Explain why NAT exists and define inside local, inside global, outside global, and outside local with one worked example
- Mark the inside and outside interfaces and choose static NAT, dynamic NAT, or PAT overload for a given requirement
- Configure PAT overload so many private hosts share one public address, and read the port numbers in the translation table
- Interpret every field of show ip nat translations and show ip nat statistics
- Diagnose a NAT that does not translate by checking interface role, the ACL, and Layer 3 routing in order
Terms you will see
Time breakdown
- Read the notes26 min
- Walk the PAT configuration and translation table14 min
- Work the three-case decision practice9 min
- Work the troubleshooting scenario10 min
Assigned video

NAT Part 1 (Day 44)
By Jeremy's IT Lab · Opens externally on YouTube
- Assigned watch
- ~8 min
Where to stop
PAT overload and reading the translation table are developed in the written sections below, not in this video.
The source marks this segment timestampStatus="unresolved" with no start or end, so no bounds are shown here.
The link opens in a new browser tab. Return here when you finish watching.
Watch for these concepts
- That inside and outside are read from interface role, not diagram direction
- A static one-to-one mapping being built
Go beyond the video
The assigned video introduces the idea. CCNA Practice Labs completes the learning path: clarify core concepts, explore how each device behaves, visualize the communication path, check your understanding, and prepare for hands-on lab work.
Watch the concept → understand it → visualize it → practice it → apply it.
Why NAT exists, and the four address terms
RFC 1918 reserved 10.0.0.0/8, 172.16.0.0/12, and 192.168.0.0/16 for internal reuse. Every organisation can use 10.10.10.0/24 precisely because those addresses never appear on the public internet — an ISP drops a packet sourced from them. NAT is the translation at the edge that rewrites the private source address, and for PAT the source port too, as packets cross from inside to outside, then rewrites the reply on the way back in.
The four address terms are pure exam vocabulary, so anchor them to one concrete flow: inside host 192.168.1.10 reaching web server 8.8.8.8 through a NAT router whose public address is 203.0.113.5.
| Term | Meaning | In this example |
|---|---|---|
| Inside local | The inside host's real, private address | 192.168.1.10 |
| Inside global | The public address the inside host appears as after translation | 203.0.113.5 (plus a port such as :50001 for PAT) |
| Outside global | The real public address of the outside destination | 8.8.8.8 |
| Outside local | The outside host as seen from inside (usually identical) | 8.8.8.8 |
Three types, and the interface roles that make any of them work
| Type | Mapping | Use it when |
|---|---|---|
| Static NAT | Permanent one-to-one | An inside server must be reachable *from* the internet at a stable public address — only a permanent mapping is reliably reachable from outside |
| Dynamic NAT | Temporary one-to-one from a pool | You need real distinct public source addresses per host and have a pool to spend; pool size caps concurrent hosts |
| PAT (overload) | Many-to-one, kept apart by source port | Many inside hosts just need to reach the internet behind one address — the normal case |
Work that table top to bottom and the first match wins. Must an outside host initiate to a specific inside host? Static. Do you genuinely need distinct public sources and own a pool? Dynamic. Otherwise — and this is the overwhelming default — PAT overload on the outside interface address.
Configure PAT overload and read the translation table
Topology: R1 is the edge router. Gi0/0 faces the private LAN 10.10.10.0/24 (R1 = 10.10.10.1). Gi0/1 faces the ISP with public 203.0.113.2/30 and a default route to the provider. Hosts 10.10.10.11 and 10.10.10.12 need internet access, sharing R1's single public address.
! 1. Mark interfaces by role
R1(config)# interface GigabitEthernet0/0
R1(config-if)# ip nat inside
R1(config-if)# interface GigabitEthernet0/1
R1(config-if)# ip nat outside
! 2. Define which inside sources to translate (standard ACL)
R1(config)# access-list 1 permit 10.10.10.0 0.0.0.255
! 3. Translate them to the outside interface address, overloaded
R1(config)# ip nat inside source list 1 interface GigabitEthernet0/1 overloadR1# show ip nat translations
Pro Inside global Inside local Outside local Outside global
tcp 203.0.113.2:50001 10.10.10.11:50001 8.8.8.8:443 8.8.8.8:443
tcp 203.0.113.2:50002 10.10.10.12:51544 8.8.8.8:443 8.8.8.8:443
icmp 203.0.113.2:12 10.10.10.11:12 8.8.8.8:12 8.8.8.8:12
R1# show ip nat statistics
Total active translations: 3 (0 static, 3 dynamic; 3 extended)
Outside interfaces: GigabitEthernet0/1
Inside interfaces: GigabitEthernet0/0
Hits: 148 Misses: 3- Pro — the transport protocol of the flow. PAT keys on protocol plus port, so each flow is a distinct row.
- Inside global — both TCP rows share the same address 203.0.113.2 but carry different ports, :50001 and :50002. That port is the whole trick: it is how the router demultiplexes returning traffic back to the right inside host.
- Inside local — the real private socket. Host .12 used source port 51544 internally and was given :50002 on the outside; the router keeps the original port when it can and rewrites it when it must stay unique.
- Outside local and Outside global — both 8.8.8.8:443, identical because the server is genuinely public and no outside translation is in play.
- show ip nat statistics confirms the interface roles and the translation counts.
For an internal server that must accept connections *from* the internet, PAT alone is not enough — an outside host needs a permanent target, so add a static mapping: ip nat inside source static tcp 10.10.10.50 80 203.0.113.2 80.
A PAT that installs nothing
Symptom: inside hosts cannot reach the internet, show ip nat translations shows nothing, and the statistics show misses climbing with zero active translations and zero hits. An empty table means traffic is not matching a NAT rule, so walk the pipeline in order.
R1# show ip nat statistics
Total active translations: 0 (0 static, 0 dynamic; 0 extended)
Outside interfaces: GigabitEthernet0/1
Inside interfaces: GigabitEthernet0/0
Hits: 0 Misses: 212
R1# show access-lists 1
Standard IP access list 1
10 deny 10.10.10.0, wildcard bits 0.0.0.255
20 permit anyR1(config)# no access-list 1
R1(config)# access-list 1 permit 10.10.10.0 0.0.0.255
R1# clear ip nat translation *What you should retain
- NAT rewrites non-routable private addresses to public ones at the edge; PAT/overload shares one public IP using unique source ports.
- Inside local = real private, inside global = public face; local and global name the side you are looking from.
- Static = permanent 1:1 and reachable from outside; dynamic = temporary 1:1 from a pool; PAT = many:1 with ports, the default for user internet.
- Roles first — ip nat inside and ip nat outside by interface function — then the source ACL, then the source rule ending in overload.
- Verify with show ip nat translations (rows are flows; watch the port column) and show ip nat statistics.
- Misses count lookups that had to create a translation, so they rise on a healthy router. The fault signature is misses rising with 0 active translations and 0 hits.
- Troubleshoot in order: interface role, then the source ACL permits the inside subnet, then Layer 3 routing to the ISP.
Before you read on
Three cases on the R1 edge above. (1) Forty inside users need internet through R1's one public IP — which NAT type? (2) You configure ip nat inside source list 1 interface Gi0/1 but omit overload — what happens to the second and later hosts? (3) An outside customer must reach an inside web server at a fixed public IP — which NAT type?
See it happen
NAT/PAT Lab
Step through the translation illustrator and watch inside local, inside global, and the port field rewrite as you switch between static, dynamic, and PAT modes — the port column is the one to keep your eye on.
Check this section before moving on
Addressing Services & NAT
The quiz opens on its own page so you can focus on it. It is a Week 1 milestone, tracked separately from this lesson's own completion.
Topics covered
- • The DORA exchange and which messages are broadcast
- • DHCP relay and when it is required
- • The four NAT address terms
- • PAT overload and the port field
- • Isolating a NAT that installs no translations
Study deeper
Topic guides extend this lesson — they do not replace the first-party walkthrough above.
NAT
For the full four-term table with the local/global mnemonic and additional worked translation examples
