I don't think I have keystone jacks in my cat6 install...more like cheapstone.

[imagoon]

wrong post.

 

[JimKiler]

Getting data through is one thing, getting optimum performance (which is assumed you want wen you pay for it) is another. Some folks get a link light and think they're golden; it just ain't so.

Do whatever you want at home but for commercial installs, you gotta do it right or see your profit margin shrink or disappear.

I did my own runs in my house before they sheet rocked. I get the link light but have not tested anything speed wise. What is the best way to test it?

 

[kevnich2]

I did my own runs in my house before they sheet rocked. I get the link light but have not tested anything speed wise. What is the best way to test it?

The best way is iperf/jperf. It will truly show you the max speed your link is capable of. Run it between two computers and the slowest link will be what speed is shown.

 

[Papa Hogan]

We de-railed the OP enough here I think. He may have to start a new thread!

Sorry Papa. Carry on!

No, YOU carry on! I don't have too many rules for my threads...you guys are more than welcome to talk about whatever. Please continue! I've learned alot from you guys.

 

[ScottMac]

I did my own runs in my house before they sheet rocked. I get the link light but have not tested anything speed wise. What is the best way to test it?

The best way is a certification tester, but they're very expensive for a one-shot use. It might be possible to hire a cable person to bring their meter and scan it for you.

As mentioned above, a practical test (iperf/jperf) is as good for the home consumer, since you really aren't interested in the parametric info, you just want to see that the system is performing up to your expectation.

 

[brshoemak]

The best way is a certification tester, but they're very expensive for a one-shot use.

By "expensive" he means around $6K-$9K for an Fluke DTX Cable Analyzer or Ideal LanTEK (<- which are pretty nice btw).

I highly doubt a cable technician is going be carry one of those around for consumer testing, those are really only for cable plants or enterprise customers. Also, most consumers don't deal with NEXT issues.

Do a simple iPerf between two nodes and you'll have a decent understanding of performance. If you are using a managed switch you will be able to see CRC errors through the web interface which would point to actual line issues, but those switches are not terribly common in most homes.

 

[drebo]

I highly doubt a cable technician is going be carry one of those around for consumer testing, those are really only for cable plants or enterprise customers. Also, most consumers don't deal with NEXT issues.

Meh, we'll test and certify any cables any customer wants us to. We don't pay per cable.

 

[JimKiler]

The best way is iperf/jperf. It will truly show you the max speed your link is capable of. Run it between two computers and the slowest link will be what speed is shown.

I finally tested my cable runs and testing with my netbook only gives me 250Mbps max. So then i tested my two desktop which are connected to the same router and i get 450Mbps. Is that the max of the cable or is my router limiting me? I have an Asus RT-N16 which is only a couple years old.

I need to bring my work laptop home and test again.

 

[drebo]

It's probably a combination of processor in the netbook and your workstations and your switch.

Lots of things could be bottlenecking, including the cable.

 

[Carson Dyle]

It's probably a combination of processor in the netbook and your workstations and your switch.

Lots of things could be bottlenecking, including the cable.

Doesn't sound like a very useful test if that's the case.

 

[imagoon]

Doesn't sound like a very useful test if that's the case.

Well you start by testing the system and removing parts and looking at the results. It is a pretty typical test. If the netbook doesn't have enough CPU to generate 1gb/s of network flow, then right off the bat you can tell the netbook would be the bottleneck since even if it had a RAID0 of 12 SSD drives, you would never max out the NIC because the CPU simply can't handle the load so the file transfer would never max out the network.

Typically with iperf you would get a pair of laptops, do the test via a single cable, laptop to laptop and use that as a baseline (most centrino and better machines should be able to load up a 1gb/s network without issue.) Then plug them in over the segment you want to test and look at the results. It the flow is still 1gbps, then "the network is slow" isn't the network and you start checking disks and the like. If the speed drops to 100mb/s (and there isn't supposed to be a 100meg connection anywhere etc) you know that something is broke and you keep moving up the chain.