Why is my network so SLOW?
Discussion
Quickie, and I may be barking up entirely the wrong tree here but ....
... why is my network so slow?
I have a network mounted drive on my computer. I run Linux Mint. It's hooked through Cat5 cable (OK, not up to the usual PH standard but come with me here) to a TP Link 10/100 Mbps switch (one of these ).
My wife's computer is also running Mint and hooked directly to the same switch.
I've just tried to copy the contents of my HDD from my machine to hers and it's showing 5.5MB/sec, meaning 15 minutes or so for 11.2GB of data. You can see the file transfer count ticking up - OK, it's ticking quite quickly but it's going in single digits, not tens of files at a time. These are word files, not huge videos.
Am I wrong to expect it to go quicker than this?
(Note: I have no need for it to go quicker, I'm just curious.)
... why is my network so slow?
I have a network mounted drive on my computer. I run Linux Mint. It's hooked through Cat5 cable (OK, not up to the usual PH standard but come with me here) to a TP Link 10/100 Mbps switch (one of these ).
My wife's computer is also running Mint and hooked directly to the same switch.
I've just tried to copy the contents of my HDD from my machine to hers and it's showing 5.5MB/sec, meaning 15 minutes or so for 11.2GB of data. You can see the file transfer count ticking up - OK, it's ticking quite quickly but it's going in single digits, not tens of files at a time. These are word files, not huge videos.
Am I wrong to expect it to go quicker than this?
(Note: I have no need for it to go quicker, I'm just curious.)
100Mb Ethernet has a theoretical max (ignoring protocol overhead) or 12.5MB/s. So you are only achieving about 50% of that (real world maximum allowing for protocol overhead is about 10MB/s)
Is your cabling good? Are both end points configured correctly? Perhaps the transfer protocol is less than ideal. You could try iperf to get a raw speed reading and rule out hardware/cabling issues which would just leave the protocol or implementation
Is your cabling good? Are both end points configured correctly? Perhaps the transfer protocol is less than ideal. You could try iperf to get a raw speed reading and rule out hardware/cabling issues which would just leave the protocol or implementation
Small files are slower than big files (more overhead). Getting 5.5MB/sec of word documents on a 100 Mbit network is about right.
Swap the switch for a Gb one, they’re cheap:
https://www.amazon.co.uk/NETGEAR-Ethernet-Unmanage...
Swap the switch for a Gb one, they’re cheap:
https://www.amazon.co.uk/NETGEAR-Ethernet-Unmanage...
As above, I'd swap it for a gig switch for the sake of £20.
Check the driver in Mint is the correct one for your network interface.
Some years ago when 100meg switches were common, some devices would drop to half duplex when running at 10 or sometimes 100 meg, so check the negotiated settings the NIC has chosen, fix it manually to 100meg full duplex if it's chosen half duplex. This won't halve the speed but it will reduce it significantly.
But first, check the speed using a single large file, like a few videos zipped together to give you a single large file.
Check the driver in Mint is the correct one for your network interface.
Some years ago when 100meg switches were common, some devices would drop to half duplex when running at 10 or sometimes 100 meg, so check the negotiated settings the NIC has chosen, fix it manually to 100meg full duplex if it's chosen half duplex. This won't halve the speed but it will reduce it significantly.
But first, check the speed using a single large file, like a few videos zipped together to give you a single large file.
Guys,
Thanks for the answers. I'm not fussed by the situation; I rarely do large file moves between the two machines and this is the first time I've even noticed it, a year after the network was set up. However I am curious.
For reference, this is between two machines, not a NAS. Both of them have SATA2 drives, mounted by fstab and shared across the network by SAMBA. I don't know what the transfer protocol or implementation are - as in, I don't know what these terms refer to; I have simply stuck network cables in the back of the machines and hooked them up to the switch.
The cables are commercially bought ones but not expensive - the sort of thing that comes with a router. I'm not sure whether it's relevant but one computer is connected to the switch via two RJ45 sockets mounted on two sides of a wall, with a small bit of network cable in the wall between them. The terminations in the RJ45 sockets were done by me and I think they are good but have no way of knowing.
iperf3 sounded interesting - thanks for the suggestion robbieduncan and CaptainMorgan. I've installed it and this is the result.
As I said, I'm curious rather than bothered. And certainly not curious to the tune of the £20 it would cost to buy a Gigabit switch (not being a powerfully built company director, using Linux Mint because it's free and noting the Cost of Living Crisis ....
) .
Thanks for the answers. I'm not fussed by the situation; I rarely do large file moves between the two machines and this is the first time I've even noticed it, a year after the network was set up. However I am curious.
For reference, this is between two machines, not a NAS. Both of them have SATA2 drives, mounted by fstab and shared across the network by SAMBA. I don't know what the transfer protocol or implementation are - as in, I don't know what these terms refer to; I have simply stuck network cables in the back of the machines and hooked them up to the switch.
The cables are commercially bought ones but not expensive - the sort of thing that comes with a router. I'm not sure whether it's relevant but one computer is connected to the switch via two RJ45 sockets mounted on two sides of a wall, with a small bit of network cable in the wall between them. The terminations in the RJ45 sockets were done by me and I think they are good but have no way of knowing.
iperf3 sounded interesting - thanks for the suggestion robbieduncan and CaptainMorgan. I've installed it and this is the result.
iperf3 on the server said:
me@mycomputer:~$ iperf3 -c 192.168.10.11 -f K
Connecting to host 192.168.10.11, port 5201
[ 5] local 192.168.10.10 port 44468 connected to 192.168.10.11 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 11.8 MBytes 12055 KBytes/sec 0 90.5 KBytes
[ 5] 1.00-2.00 sec 11.3 MBytes 11526 KBytes/sec 0 126 KBytes
[ 5] 2.00-3.00 sec 11.4 MBytes 11645 KBytes/sec 0 130 KBytes
[ 5] 3.00-4.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 4.00-5.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 5.00-6.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 6.00-7.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 7.00-8.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 8.00-9.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 9.00-10.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 113 MBytes 11540 KBytes/sec 0 sender
[ 5] 0.00-10.00 sec 112 MBytes 11480 KBytes/sec receiver
iperf Done.
To my eyes that looks like 11.5Meg transfer, which is about par for the course. Unless I'm reading it wrong, that is. Connecting to host 192.168.10.11, port 5201
[ 5] local 192.168.10.10 port 44468 connected to 192.168.10.11 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 11.8 MBytes 12055 KBytes/sec 0 90.5 KBytes
[ 5] 1.00-2.00 sec 11.3 MBytes 11526 KBytes/sec 0 126 KBytes
[ 5] 2.00-3.00 sec 11.4 MBytes 11645 KBytes/sec 0 130 KBytes
[ 5] 3.00-4.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 4.00-5.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 5.00-6.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 6.00-7.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 7.00-8.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 8.00-9.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
[ 5] 9.00-10.00 sec 11.2 MBytes 11454 KBytes/sec 0 151 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 113 MBytes 11540 KBytes/sec 0 sender
[ 5] 0.00-10.00 sec 112 MBytes 11480 KBytes/sec receiver
iperf Done.
As I said, I'm curious rather than bothered. And certainly not curious to the tune of the £20 it would cost to buy a Gigabit switch (not being a powerfully built company director, using Linux Mint because it's free and noting the Cost of Living Crisis ....
) .Pretty much any modern ISP router will have gigabit ports if you can rearrange things to use that. Alternatively, if you have a spare one, turn DHCP and wireless off on it and it will act as a switch.
I had a quick hunt on Amazon. FWIW TP-Link do a 5 port plastic cased gigabit model for £8 if that is any more appealing. Does make it easier and hence more likely you will copy files to the other PC or NAS so you have a backup.
I had a quick hunt on Amazon. FWIW TP-Link do a 5 port plastic cased gigabit model for £8 if that is any more appealing. Does make it easier and hence more likely you will copy files to the other PC or NAS so you have a backup.
Tempted by £8 ?
TP-Link LS1005G 5-Port Desktop/Wallmount Gigabit Ethernet Switch/Hub, Network Splitter, Plug and play, Plastic Case https://www.amazon.co.uk/dp/B07VWB347G/ref=cm_sw_r...
TP-Link LS1005G 5-Port Desktop/Wallmount Gigabit Ethernet Switch/Hub, Network Splitter, Plug and play, Plastic Case https://www.amazon.co.uk/dp/B07VWB347G/ref=cm_sw_r...
thebraketester said:
Tempted by £8 ?
TP-Link LS1005G 5-Port Desktop/Wallmount Gigabit Ethernet Switch/Hub, Network Splitter, Plug and play, Plastic Case https://www.amazon.co.uk/dp/B07VWB347G/ref=cm_sw_r...
Nope, not even at £8. It's money I don't need to spend, but thanks for the link! TP-Link LS1005G 5-Port Desktop/Wallmount Gigabit Ethernet Switch/Hub, Network Splitter, Plug and play, Plastic Case https://www.amazon.co.uk/dp/B07VWB347G/ref=cm_sw_r...
Out of curiosity (and this may be a daft Q) but what is the relationship between 10/100 vs. Gigabit switches and Cat5/Cat6 cabling? I am aware that I only have Cat5 cables.
2Btoo said:
Nope, not even at £8. It's money I don't need to spend, but thanks for the link!
Out of curiosity (and this may be a daft Q) but what is the relationship between 10/100 vs. Gigabit switches and Cat5/Cat6 cabling? I am aware that I only have Cat5 cables.
Are you 100% sure it's Cat5? That's moderately hard to get hold of these days. Cat 5e is much more common (and happily gigabit-capable).Out of curiosity (and this may be a daft Q) but what is the relationship between 10/100 vs. Gigabit switches and Cat5/Cat6 cabling? I am aware that I only have Cat5 cables.
Haven't a clue, actually. The cables are whatever came with the routers I have had over the last few years and the installed bits (i.e in the walls) are from a bundle of old scruffy stuff that came off a drum about 4 years ago. It could be Cat5e, or Cat5, or Cat4 for all I know. (Was there a Cat4? Shows how much I know about this sort of thing!)
Look at the plugs. As long as there are 8 conductors, you can expect it to run at gigabit.
They should be wired to one of these specs: https://en.wikipedia.org/wiki/ANSI/TIA-568#T568A_a... , but if there are 8 conductors, the odds of that not being the case are minute.
They should be wired to one of these specs: https://en.wikipedia.org/wiki/ANSI/TIA-568#T568A_a... , but if there are 8 conductors, the odds of that not being the case are minute.
xeny said:
Look at the plugs. As long as there are 8 conductors, you can expect it to run at gigabit.
They should be wired to one of these specs: https://en.wikipedia.org/wiki/ANSI/TIA-568#T568A_a... , but if there are 8 conductors, the odds of that not being the case are minute.
Helpful, thanks. They should be wired to one of these specs: https://en.wikipedia.org/wiki/ANSI/TIA-568#T568A_a... , but if there are 8 conductors, the odds of that not being the case are minute.
Yes, 8 conductors in all cases. (Although I did see a cable the other day that seemed to have only 4 conductors going into each RJ45 plug. Wasn't mine, but was noteworthy as I don't think I've seen such a thing before.)
2Btoo said:
Helpful, thanks.
Yes, 8 conductors in all cases. (Although I did see a cable the other day that seemed to have only 4 conductors going into each RJ45 plug. Wasn't mine, but was noteworthy as I don't think I've seen such a thing before.)
They will typically run at 100 mbit, but definitely not gigabit. At one point in my career I was responsible for a building that had been flood wired with the stuff :-(.Yes, 8 conductors in all cases. (Although I did see a cable the other day that seemed to have only 4 conductors going into each RJ45 plug. Wasn't mine, but was noteworthy as I don't think I've seen such a thing before.)
Gassing Station | Computers, Gadgets & Stuff | Top of Page | What's New | My Stuff