Let's say that's roughly 315kbps of upload, that's still enough to make 10kbps far too slow !
Remember 10Kb is actually eight times 10kbps (80kbps).
There's probably a 'multiple connections' or 'concurrent transfers' setting or something
similar you could look for.
That might open up a bottleneck in the software. A wild guess I'm afraid but think along
those lines initially and it can't do too much harm.
I think user "s" asked if it was one big file in case your MTU settings were wrong. Check
your local router is set as per your ISP's recommended MTU properly too.
I've moved away from the sever and have checked the FireFTP settigns too where it is possible
to further increase the number of multiple connections. On retesting the speed, things have
improved - up to 280kbps on the upload which isn't so bad - so things are looking much
better. The two things may not be related, but what matters is there seems to have been an
Thanks all for the suggestions :-)
Just to clarify - have tried small files (<50kb), larger files (2-3mb) and combinations of
sending just one file and more than one file as well as sending mixed file sizes at the same
The MTU thing hadn't occurred to me. Both the routers are as supplied by the ISP's though so
won't the MTU on each already be configured to the optimum for that ISP?
There is a performance setting called "number of threads" which after reading something about
increasing this to help when the server is busy I changed to 10 from the default 2. I've just
tried changing this back to 2 but it's had no effect on the speed of upload.
I wish FileZilla Server, speed test sites and the diagnostic programs would report their
speeds in the same units - it would help us all when we're doing comparisons!
My ISP maximum upload speed is about 0.36megs.
I've only tested it with one file transfer at a time rather than having multiples on the go.
Doesn't seem to matter what the file type or size is, the same low transfer speed results.
Are you transferring one big file or plenty of small files?
Yes, you're right.
All you want to check is that your ISP is giving you what you expect (ie more than 10kbps !).
You're checking the network is working correctly before looking at the application.
So if it's an ADSL line and you connect at 8Mbps, the chances are you'll get 4Mbps in the UK
as a yardstick.
The results from these tools are just in Mbit or Kbit. Remember kbps and KB/s are different !
Try and stick to kbps and megs for ease.
I think these tools might be a bit sophisticated for me as this is my first outing with any
kind of server. A case of if I ran the tool, would I understand the results!? :-)
If I understand correctly, I need to check the hardware settings for the network controllers
on the server and the client. The server one says '100Mbps / Full Duplex'. I can't check the
client one I used at the moment but can on Monday. I seem to remember last time I looked it
was similar to this though.
What other places should I look for possible causes, and how will I recognise the cause when
I see it?
Should have mentioned - I'm running the server in Windows so any tools that can work there
would be great.
A couple of thoughts spring to mind:
Don't be too certain about the connection starting correctly and then throttling back. The
difference between what you can fit down the pipe and the maximum speed of the pipe is always
pretty different, in simple terms.
Check your speed and duplex on the NICs. You might get fast UDP and erratic TCP if you have
one NIC accidentally at HD not FD. You might not notice this in a LAN because latency is far
The best tools are iperf and netcat (use nc -l ...)
for throughput testing. I know it sounds rudimentary but spend 10 minutes checking your
*actual* connection speeds with one or both of these tools. You'd be surprised how poor some
DSL performance is .... really :)
You can then get deeper into config parameters if you're 100% certain that the network is as