In general, the Ethernet Switch use RJ-45 link to connect the Hosts and the Switch uses Store and Forward mode.

So, why is there a saying:

Switch can segregate the collision domain

I thought the Switch Store and Forward mode and RF-45's full duplex mode eliminates collisions.


A collision domain is a network segment connected by a shared medium or through repeaters/switches where data packets may collide with one another while being sent. The collision domain applies particularly in wireless networks, but also affected early versions of Ethernet. A network collision occurs when more than one device attempts to send a packet on a network segment at the same time. Members of a collision domain may be involved in collisions with one another. Devices outside the collision domain do not have collisions with those inside.

Since every port in a switch is its own collision domain, a host will never collide unless it's medium is running half duplex.

That is why a switch segregates collision domains, because every port is an individual collision domain.


If you replace a repeater hub (long obsolete, of course) by a switch, you indeed break the larger collision domain into many smaller ones - one on each switch port when half duplex mode is used. Full-duplex links (to a FDX-capable NIC or another switch) are generally collision-free and don't take part in any collision domain.

The main point with switches is that they buffer network frames. That enables a switch to receive a frame at any time and then forward it later when the egress link is idle. This decoupling of receive and transmit operations enables a network that works with flows that are largely independent from each other and only compete for link bandwidth (assuming FDX links throughout).

This is in stark contrast to repeater hubs that repeat incoming bits as they are received. A hub can't buffer anything - when a single transmission is already in progress, a second transmission attempt cannot the repeated at the same time and causes a collision. The collision needs to disrupt reception on the ingress interface since data is already garbled, so the hub propagates it across all ports, including the one back to the source.

This way, all nodes connected to a hub (or potentially chained hubs) form a single, common collision domain. Only a single node at any time can transmit.

A switch connected to a hub (or another non full-duplex capable device) can still use half-duplex mode on any of its ports but due to the buffering between the ports, any collision will not propagate across the switch - the switch will simply retransmit the buffered frame later on. Therefore, a switch segregates collision domains or removes them completely on full-duplex links.

Note that half-duplex Ethernet and hubs are very much things of the past and only interesting for historical reasons or under very specific circumstances. Practically all Gigabit and (by standard) faster Ethernet links have dropped support for half-duplex communication and you need to use switched or point-to-point connections exclusively.


I thought the Switch Store and Forward mode and RF-45's full duplex mode eliminates collisions.

That's what "segregate collision domains" means. The switch eliminates collisions because the collision domain for each port is limited (segregated) to just that port. Combine this with full duplex mode and only one device on a port, and there's nothing left to collide with.

However, you can still have collisions if there is an old-school hub connected to the switch port and multiple devices connected to the hub. Thankfully, we try not to use hubs anymore, but this is why it's more correct to say "segregates the collision domain" over "eliminates collisions"; it's not that hard to contrive situations where collisions can still happen.


A bit of explanation here:

The main query I assume here is how switch avoids collision, not vague answers like : as each port is full-duplex. How is full-duplex achieved or how a port in switch doesn't share its media with other ports internally? There can be shared media inside switch.

And its not buffering that avoids collision : buffering basically is ability to store and forward data when memory of a port can't handle or input exceeds limit. The transmission is already serial. It controls the flow of data.

The magic that happens inside to make each port of a switch to run in full-duplex format without being interrupted by other ports is by use of cross-bar (non-blocking) silicon fabrics.

In gist, its the cross-bar internal silicon fabrics of a switch that enables each port of modern switches be in separate collision domain ( aka no internal shared bus/media at all between ports)

Have a look at this if you want deep-dive : https://etherealmind.com/what-is-the-definition-of-switch-fabric/

  • You are referring to modern switches that do everything in hardware, but the original bridges/switches that did switching/bridging in software also did not have collisions. – Ron Maupin Nov 3 '20 at 14:57
  • True that, software based sequencing have avoided collision when bus was somehow shared. But since question was asked two years ago and nobody talked about internal aspects of full-duplex, I decided to shed some light. :) – kushal khanal Nov 3 '20 at 15:04

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.