Does the length of cat6 cable make a difference in a home network?

1

I have a Dlink DIR-655 router for my wired and wireless router for my home network. I have a 100ft cat6 ethernet cable. I pulled it from the router to my home server in my basement.

Does the 100ft cable make a difference or does the fact that it's a cat6 cable make a difference?

When moving similar files at work, it's faster. So I'm trying to find out what's the bottleneck at home. What speeds should I expect for a home network?

vlash

Posted 2011-12-11T00:21:51.657

Reputation: 11

2Moving files to and from what? It's not clear what you're actually doing. If you're talking about speed talking to wireless nodes or nodes over the Internet, the Ethernet is irrelevant. – David Schwartz – 2011-12-11T04:13:24.237

Answers

2

100ft is not a problem for running 1Gbps over Cat-6, which should be able to be run over 100 meters. However, the likely culprit here is the NIC in your home server, and/or the DIR-655 itself, either of which may not be very fast. The SmallNetBuilder guys were only able to squeeze 250Mbps through the router; so the internal switch is at least that fast, but probably not much more.

Your work router is almost certainly far more powerful than your $50 DIR-655.

joev

Posted 2011-12-11T00:21:51.657

Reputation: 543

Note that the test you refer to describes routing speeds, not local switching speeds. That might be different. – Eroen – 2011-12-11T02:46:10.340

To be honest, I don't think I need 100ft of the cable. As some of you have mentioned, it could be an issue either in the router or in the NIC card. Let me run some test tonight on other networked machine and see what the speeds are. (have both wired & wireless setups). – vlash – 2011-12-13T14:40:46.340

2

The speed will be limited to the lowest of:

  • Hard drive speed, compensated for seeking and caching
  • Any part of your network between your locations, including switches, hubs, and network cards.
  • In extreme cases, bus and cpu speeds of the connected devices.

Only in very rare cases would the cable quality or length be any real issue, in that case you would not observe reduced speed but increased lost/damaged data. (You could check that by continuously pinging your server while copying a large file from it.)

I suggest you first check that all your network components (computer, router and server) support Gigabit speeds (or whatever your think you should have). Secondly, I would suspect the router not actually offering the speeds it claims (seen that more than I should have). You could then connect a computer directly to the server to see if speeds increase.

Another likely possibility is low speeds on the wireless connection to you computer, likely due to poor reception.

Eroen

Posted 2011-12-11T00:21:51.657

Reputation: 5 615

2

I pulled it from the router to my home server in my basement.

As already mentioned, 100' of Cat6 UTP cable is well under the 100m Ethernet limit by more than two thirds. But the methods you used while "pull"ing this UTP cable could be an issue. There are simple and clearcut guidelines for handling UTP to avoid damaging the cable. Cable damage can result in poor network performance. The basic rules while installing UTP (and coax) cable are:

  • do not kink the cable.
  • avoid tight or sharp bends; for UTP the bend radius should be at least 1".
  • do not pull too hard on the cable; it's not a rope! UTP typically has a limit of 25 pounds of force.
  • do not flatten or deform the cable with staples or cable ties.
  • keep signal cables away from power cables and sources of EMI, e.g. power supplies, transformers, CFL lights.

If you terminated the cable yourself, then there's another set of best practices to follow.

More info is at Cat5/5E/6 Cabling Tutorial.

Use the ifconfig command in *nix to see if the network interface has any detected transmit/receive errors.

Addendum: a full-featured US$100+ wireless router w/Gigabit ports from a name-brand network equipment manufacturer came with a Cat5e patch cable that was tightly bundled (not coiled). The bends had radii as small as 1/4 inch. Such patch cables should be considered DOA, and trashed.

sawdust

Posted 2011-12-11T00:21:51.657

Reputation: 14 697

1

The length and type of cable will limit its maximum bit rate. Normally the devices on each end negotiate the data rate, assuming the cable will handle it, and if the cable can't then the connection simply fails.

But it's possible that the D-Link unit may have automatic fallback if it detects too many errors, etc. (I confess to not being up on the latest networking stuff.) But if that's happening then you should be able to sign onto the router and query the data rates of its ports somehow, or else there should be LEDs for the ports to indicate what data rate they're running at. If the port in question is running at 1G (and there's no indication of excessive errors) then the cable is not causing a performance problem.

But I'd be surprised if 100 feet of Cat-6 is a problem.

Daniel R Hicks

Posted 2011-12-11T00:21:51.657

Reputation: 5 783

0

The degradation in speed might be because of how you pulled the cable up to your server.

AC motors or tube chocks can create distortion over RJ-45 and you must check if RJ-45 is going parallel to any electrical supply cable for more than three meters. If that's not the problem, after that, check your NIC.

Amit Bhanushali

Posted 2011-12-11T00:21:51.657

Reputation: 9