How does the receiving switch keep the groups apart?
In our example a PC and printer belong to VLAN 10, separated by SW1 and SW2. Their ordinary access links carry untagged data. The trunk between the switches also carries VLAN 20, so SW2 needs to know which group each non-native data frame belongs to.
A tagged frame includes an IEEE 802.1Q field identifying its VLAN. An untagged frame has no such field on that cable. A frame can belong to VLAN 10 inside the switch without carrying a VLAN 10 tag on every link. Classification inside the switch and the bytes transmitted on the cable are different things.
Tagged and untagged describe the frame on this link
An ordinary switched data frame is classified into a VLAN, but it need not carry its VLAN number. That's the whole difference:
The explorer uses a VLAN 10 PC-to-printer path. Choose each cable and keep native VLAN 99 on both switches initially. The native VLAN is the receiving trunk's classification for ordinary untagged data when native tagging is disabled; it is not the VLAN that every device must use.
Inspect a frame on each cable
The single tag is four bytes, inserted after the source MAC address. Its TPID (Tag Protocol Identifier), normally 0x8100 for this customer VLAN tag, identifies the tag format. TCI (Tag Control Information) contains the priority, drop-eligibility and VLAN fields below. Hexadecimal is a compact way of writing bits; 0x8100 is not an IP address or application port.
- PCP (Priority Code Point) is three bits, with values 0–7. It can help a configured quality-of-service policy classify traffic; the bits do not guarantee bandwidth by themselves.
- DEI (Drop Eligible Indicator) is one bit used by supported congestion policies.
- VID (VLAN Identifier) is 12 bits. Values 1–4094 identify ordinary VLANs; 0 means priority-tagged traffic without a VLAN ID assignment, and 4095 is reserved. A switch may support a smaller number of simultaneously active VLANs.
For the journey, assume the PC already knows the printer's MAC address and the switches have learned its location. Tagging does not replace ARP (Address Resolution Protocol) or MAC learning. Those processes find the local receiver and output port; the tag preserves the VLAN context on the shared trunk.
VLAN 10TypePayloadFCS
VLAN 10 is not native, so its data frame carries an 802.1Q tag across the trunk.
On small screens, swipe sideways or focus the diagram and use the arrow keys.
Why both exist: a typical endpoint uses untagged data, while a VLAN-aware device can handle tags. Switch port configuration determines what is expected on each link; the device type alone does not decide the port mode.
Where each one is used
On small screens, swipe sideways or focus the diagram and use the arrow keys.
Reading other vendors' VLAN settings
Some platforms describe tagged and untagged VLAN membership instead of access and trunk modes. Ingress means traffic entering a port; egress means traffic leaving it. The PVID (Port VLAN Identifier) classifies ordinary untagged incoming data. Egress membership separately determines whether that VLAN leaves tagged or untagged; setting one does not necessarily configure the other. For a conventional trunk, tagged VLANs 10 and 20 with untagged VLAN 99 and PVID 99 correspond to a Cisco trunk allowing 10,20,99 with native VLAN 99.
Checking which VLANs a port tags
The following is a read-only Cisco Catalyst IOS XE example. Administrative Mode is the configured intent; Operational Mode is the actual port role. Read both. The Access Mode VLAN field can still display VLAN 1 on a trunk; it does not mean all trunk traffic belongs to VLAN 1.
SW1#show interfaces GigabitEthernet1/0/23 switchport Name: Gi1/0/23 Switchport: Enabled Administrative Mode: trunk Operational Mode: trunk Administrative Trunking Encapsulation: dot1q Operational Trunking Encapsulation: dot1q Access Mode VLAN: 1 (default) Trunking Native Mode VLAN: 99 (NATIVE) Trunking VLANs Enabled: 10,20,99 ...
show interfaces trunk. Native VLAN 99 is normally untagged; check native-tagging configuration separately, using show vlan dot1q tag native where supported or the global vlan dot1q tag native setting. A capture of the link can confirm the actual data-frame tags.Reference: Cisco IOS XE VLAN trunking guide. Examples assume ordinary data frames with at most one tag; control traffic and priority-tagged frames need separate treatment.
Short lab: predict the bytes before changing the explorer
- Keep access VLAN 10 at the PC and printer, native VLAN 99 on both trunks, and native tagging off. Select PC → SW1, SW1 → SW2 and SW2 → Printer. Predict untagged, tagged 10 and untagged, then compare the result.
- Change the explorer's native VLAN to 10. This changes both ends together. With native tagging off, the trunk data is now untagged; the receiving switch still classifies it into VLAN 10. The access frames are unchanged.
- Enable native tagging in the explorer. Observe tag 10 on the trunk. This teaching control changes only the model, not a real switch. Actual native-tagging commands can be global and platform-dependent, so do not toggle them on a live network merely to reproduce the animation.
- For a real isolated lab, read the switchport settings at both ends and compare an authorised packet capture on the intended link. NIC offload or capture placement may hide or remove tags, so an untagged-looking host capture alone does not prove the trunk sent no tag.
Tags let many groups share a cable, but require agreement between the connected devices. They identify a VLAN, not an encrypted channel. Keeping ordinary endpoints untagged reduces endpoint configuration; the switches carry the extra responsibility.
Test yourself on tagged and untagged frames
A tagged frame for VLAN 10 is sent out an access port in VLAN 10 to a PC. What does the PC receive?
A switch port tags VLANs 10 and 20, sends VLAN 30 untagged, and assigns untagged ingress to PVID 30. With native tagging disabled, what is the conventional Cisco equivalent?
A hypervisor hosts VMs in VLANs 10 and 20 over one cable. How should the switch port be set?