Ethernet is based on CSMA/CD. When using VLANs, the number of stations can become very high, making CSMA/CD inefficient. Are there any mechanisms that VLANs use to overcome this?

  • VLANs create separate smaller collision domains. Also, almost all modern networking is full-duplex, so CSMA/CD does not apply; a switch can receive a frame on every interface simultaneously, with sufficient buffer space. (true hubs haven't been manufactured for decades.) – Ricky Beam Jan 14 at 2:28
  • 1
    @RickyBeam VLANs do not affect collision domains. You may be thinking of broadcast domains. – user253751 Jan 14 at 12:35
  • Hubs don't do VLANs, and if they did, it would create partitions allowing more than one port (one per partition) to transmit at the same time. (actually had a 10b2 hub like that -- port groups, not vlans) If P1=V10, P2=V20, then as long as no port is in both vlans (and the vlans are quiet), P1 and P2 can transmit at the same time. – Ricky Beam Jan 14 at 15:40

VLANs segregate a shared link-layer (L2) topology into smaller broadcast domains.

CSMA/CD is the physical-layer (L1) access method for half-duplex Ethernet, normally only used with repeaters, hubs, or ancient Ethernet variants. Practically, it isn't even supported with current gigabit-and-faster variants any more.

Since VLANs can only (reasonably) be used with switches and switches predominantly use full-duplex links without CSMA/CD, both technologies don't really come into touch with each other.

Of course, you could configure a VLAN trunk to use half duplex on purpose, but why would you do so? Additionally, CSMA/CD would work just as well with or without VLAN tags.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.