Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

What is the maximum and/or ideal length of an ethernet cable? Is there a distance that data cannot be transferred over an ethernet cable, say over X number of feet?

share|improve this question
As far as ideal length. The ideal length would be as long as necessary and no longer. Be it for cost, interference, cable management or latency (for all the hard core gamers out there). Excess unneeded cable can cause problems. – Scott McClenning Sep 25 '10 at 23:03
There is also a MINIMUM length, to avoid problems like reflection. I encountered this years ago, when I was doing up my house. I cn't remember the value, but I THINK a cable one foot long was not recommended! – user86822 Jun 21 '11 at 8:58
which is odd, as we've got some 10cm patch cables in our server room/switch bays. – tombull89 Jun 21 '11 at 9:36
up vote 12 down vote accepted

There is no ideal length of a cat 5 cable but the specification states that it should not be more than 100m (328 feet).

More info on the wiki page

The specification of 328 feet has to do entirely with collision detection in a CSMA/CD (Carrier Sense Multi Access / Collision Detection network. Basically, the length is limited by the fact that the shortest possible frame size (64 bytes) can be sent out on the wire and if a collision occurs, the sending node will still be sending that frame when it hears the collision (due to a jam signal or higher than normal amplitude). If a collision occurs during the first 64 bytes of a frame, this is a normal collision. If something is out of spec such as the cabling is too long and the collision occurs after the first 64 bytes, this is a late collision and will not be retransmitted until an upper layer of the OSI model detects that the packet did not make it to its destination. You can run longer cabling and the network will still function, but there will be issues.

share|improve this answer
True, if the cable is used for ethernet (which is very likely, of course). The maximum length of 100 meter ensures that collision detection mechanisms work as expected (and maybe also ensures the signal strength is okay?). – Arjan Sep 25 '10 at 16:23
The statement "the length is limited by the fact that the shortest possible frame size (64 bytes) can be sent out on the wire and if a collision occurs" applies in gigabit networks but is not true for 10/100 networks. The speed of a signal in copper cables = 200,000,000 m/s. 64 bytes = 512 bits. (200,000,000[m/s]/100,000,000[bits/s]) * 512[bits] = 1024 meters (or 512 meter round trip) maxim distance to detect collisions during a 64 bit packet. – Keith Reynolds Aug 19 '15 at 19:48
Before there were network switches, hubs would repeat signals which along with round trip propagation delay is the reason for the 543 rule. that you can have a maximum 5 network segments daisy chained with 4 repeaters, and 3 mixed networks, though the mixed network rule does not apply to twisted pair. Furthermore, switched networks buffer and queue packets which removes even the 543 restriction. – Keith Reynolds Aug 19 '15 at 20:04
Instead the main limiting factor is a cable's impedance, which attenuates the signal, and a Network cards (NIC) ability to differentiate between high and low signals at the other end. the specs specify 100 meters, but with quality cable that have lower impedence and good termination with minimal cros talk or alien talk, two NICS can sometimes communicate well over cables longer than 100M – Keith Reynolds Aug 19 '15 at 20:06

The ideal length is exactly as long as you need (up to 100 meters) and no longer. Every extra 11.9 inches adds another nanosecond of extra latency.

share|improve this answer
I just swapped out my 25' cable for a 10' cable here at home. Those extra 15 nanoseconds are going to help me to be the first answerer on Stack Overflow. Thanks for the suggestion, Joel! – James McNellis Sep 25 '10 at 19:38
Signals travel down copper wire at about 200,000,000 meters a second that is approximately 656168000 feet or 1.524 nano seconds a foot. – Keith Reynolds Aug 19 '15 at 20:31
While latency is important to network performance, it is not a limiting factor in determine a 100 m specification. – Keith Reynolds Aug 19 '15 at 20:32

Ideal: As short as possible, as isolated as possible.

Maximum: 100 meters without repeaters.

share|improve this answer

To be in spec, a CAT 5 UTP cable should not exceed 100 meters. Cable can be connected with repeaters and you can get another 100 meters and so on. However, if you network is too large, then the TCP/IP packet will take so long to go from end to end computers will reach the timeout before they get a reply back. At that point other devices will have to be used to retransmit the packets, like switches/routers... I'm not sure about that distance/time before packets are considered lost.

share|improve this answer
I truly doubt the maximum length is about timeouts. I've always learned it's to ensure proper collision detection. (Timeouts limits are much, much longer than the microseconds used for collision detection.) – Arjan Sep 25 '10 at 16:25

You must log in to answer this question.

protected by studiohack Jun 22 '11 at 3:05

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged .