slow gigabit, help

Skilldibop said:
That's true of structured cabling as that's solid core, but patch cables are multicore and are designed to be flexible.
Click to expand...
Theyre designed to be flexible yes, but they only have a certain degree of flex, you bend them too much and you damage them. The same with getting kinks and knots in the cable.

You tread on them, on a hard floor, and they're ******.
 
This thread is getting extremley pedantic lol... too put it in perspective I work with 1Gb and 10Gb Ethernet fibre and copper supporting over 9000 workstations, I have NEVER seen a damaged network cable "slightly reduce the throughput" especially in a scenario such as this, if you've got a broken cable the chances are it'll just not work or be very very slow... (bits per second) Even if you have a semi broken cable udld usually detects it and stops it from forming a link.
I don't quite understand how standing on a cable makes it not conform to Cat5E standard? the 5E simply means that it uses all 4 pairs (1000base T4) so it can support short run gigabit, if you stand on the cable does this mean the other 2 pairs cease to work?
Also QOS isn't going to make any real difference as its not a multilayer switch, and can only do 802.1p TOS classification, it has 4 queues and will only start dropping frames when the queues start getting to 75% full. Again, not enough traffic for this to make a difference.
It could quite possibly be a duplex mismatch on the switch or one of the PCs but again I suspect that you wouldn't even get close to 100Mbit with a duplex mismatch.
I'd seriously say its something to do with one of the pcs or hard disk limitation or a duplex mismatch..

I give up lol
 
Last edited:
V-Spec said:
I have NEVER seen a damaged network cable "slightly reduce the throughput" especially in a scenario such as this, if you've got a broken cable the chances are it'll just not work or be very very slow... (bits per second) Even if you have a semi broken cable udld usually detects it and stops it from forming a link.
I don't quite understand how standing on a cable makes it not conform to Cat5E standard? the 5E simply means that it uses all 4 pairs (1000base T4) so it can support short run gigabit, if you stand on the cable does this mean the other 2 pairs cease to work?
Also QOS isn't going to make any real difference as its not a multilayer switch, and can only do 802.1p TOS classification, it has 4 queues and will only start dropping frames when the queues start getting to 75% full. Again, not enough traffic for this to make a difference.
It could quite possibly be a duplex mismatch on the switch or one of the PCs but again I suspect that you wouldn't even get close to 100Mbit with a duplex mismatch.
I'd seriously say its something to do with one of the pcs or hard disk limitation or a duplex mismatch..

I give up lol
Click to expand...
I'm not sure what your comprehending as "damaged" the cable doesn't have to be noticabley damaged in any way. The cable can be "damaged" by excessive bending or pulling as well as pinching, crushing etc, as doing so will disturb the twists and molecular composition of the cable which will result in abnormal resistance, capacitance and impedance, which will damage the cable and stop it conforming to the Cat standard. This is Cat6 and will probably be tested upto 250mhz so will be able to take more damage than Cat5e and function to a reasonable extent.

True it's not going to be due to QoS, It was just an afterthought.

It could be one of the client adapters or a pc setting, but it isn't going to be the hard drives.
 
Last edited:
to settle it once and for all i have just taken a 0.75m cat5 patch cable, jumped on it on a vynl floor (much to the amusement of coleagues) works fine.
Test 2 bend the cable completely in 2 so that both halfs are in contact their entire length (so much so the outer insulation ahd actually split) and low and be told it still works.
Test 3 bent it back and forth a few times on the same place. it now doesn't work at all.

It was easier to do that than explain in terms of youngs modulus and tensile stress/strain why it would either work or not.
 
It's to do with "necking" of materials and the fact that resistance is related directly to the cross sectional area of the conductor.

Of all the mind messing stuff i learnt in physics that made me think "when will i ever use this", i use a suprising amount in real life.
 
Skilldibop said:
to settle it once and for all i have just taken a 0.75m cat5 patch cable, jumped on it on a vynl floor (much to the amusement of coleagues) works fine.
Test 2 bend the cable completely in 2 so that both halfs are in contact their entire length (so much so the outer insulation ahd actually split) and low and be told it still works.
Test 3 bent it back and forth a few times on the same place. it now doesn't work at all.

It was easier to do that than explain in terms of youngs modulus and tensile stress/strain why it would either work or not.
Click to expand...
I never said it wouldn't work, it just stops the cable complying with Cat6 standards, i.e. it won't work at 250mhz, so won't transfer data at 1gb/s.
 
untitlededit1td.png


That’s the sort of speed I get across my Gigabit (40,000 - 50,000KB/s) I could get it a bit faster if I tweaked it but cannot be bothered really as I am more than happy with transferring a DVD (4.36Gb) in under 120 seconds. :D

gun8cc.png
 
Last edited:
Phnom_Penh said:
I never said it wouldn't work, it just stops the cable complying with Cat6 standards, i.e. it won't work at 250mhz, so won't transfer data at 1gb/s.
Click to expand...

You don't need Cat6 to send at 1Gbps, you only need Cat6 for 10gb Ethernet
Cat5e is used for 10/100/1000Mbps
 
I never said it wouldn't work, it just stops the cable complying with Cat6 standards, i.e. it won't work at 250mhz, so won't transfer data at 1gb/s.
Click to expand...
In theory that is true of the physical cable properties, but if it cant take a 250MHz signal 1Gbit will not be negotiated in the first place. Speed negotiation is done by sending signals of relevant requencies and seeing whcih ones are recieved.
When 250MHz signal is sent, if the medium isn't up to it the other end won't see a 250Mhz signal (or at least not an intact one) and it will not reply over 250MHz thus a lower speed will be negotiated and the next available speed will be a 100Mbit duplex.
 
V-Spec said:
You don't need Cat6 to send at 1Gbps, you only need Cat6 for 10gb Ethernet
Cat5e is used for 10/100/1000Mbps
Click to expand...
That wasn't my point anyway :\ And you can get 10G over Cat5e, although the op is using cat6, so I don't see your point.
Skilldibop said:
In theory that is true of the physical cable properties, but if it cant take a 250MHz signal 1Gbit will not be negotiated in the first place. Speed negotiation is done by sending signals of relevant requencies and seeing whcih ones are recieved.
When 250MHz signal is sent, if the medium isn't up to it the other end won't see a 250Mhz signal (or at least not an intact one) and it will not reply over 250MHz thus a lower speed will be negotiated and the next available speed will be a 100Mbit duplex.
Click to expand...
Ive tried this before, and found that Windows would still say that it was connected at 1Gbps even though it was working at a lower rate. Also, how come Cat5e can connect at 1gbps yet is only certified to 100mhz?
Also, if a wire is bent or pulled etc etc the twists will be moved. This is important as pairs will interefere with each other, causing lower data speeds, yet will not mean a lower speed at negociation.
 
Just thought I'd mention that in the past I have had a 'dodgy' cat5 cable. It wouldn't work properly as a cable for our LAN(10 or 100mbit, can't remember which) but was fine for an ISDN connection (64/128kbit). Or it might have been vice-versa, was over 6 years ago now.

However, I'm inclined to agree that it's possible that the bottleneck could be HD speed. I very rarely defrag (maybe once or twice a year) and some of my fuller HDs (say 99% full) perform rather poorly especially when cpu utilization is high. Might be worth testing drive speed to rule out the possibility.
 
You must log in or register to reply here.
Back
Top Bottom