strange problem in internet connection

the ingredients:

Pc with windows 7
A raspberry py ver.2
two hdsl 8Mb
A label printer datamax

Install virtualhere on Rpi ok
install windows drivers on w7 ok
connect Rpi on hdsl 1 and open port on 7575 and connect the label printer on usb of Rpi
connect W7 on hdsl 2

Set the driver for see the Rpi and printer ok

... the print queue ...too slow 3 or 4 minute for a print.
connect all on a local lan 2 sec for print

I do not have enough bandwidth? (ironic)

test , from and to w7 vs Rip
result, no, the bandwidth is full work correctly

I went to see how much bandwidth you are using the connection w7 vs rpi.... 150kb max!!!!! and no the full 8mb band.!!!

Simulate in local lan limiting the bandwidth at 10Mb work correctly
Simulate in local lan limiting the bandwidth at 1Mb work correctly
Simulate in local lan limiting the bandwidth at 500kb work correctly

The question is:

Because if equipment work in hdsl1 to hdsl 2 no work in full bandwidth and only max 150 kb?????

#2

Sounds to me like a latency issue. Latency and bandwidth are two different things. Latency is the time (e.g ping time) between client and server and bandwidth is how much data it can carry per packet (simplified). Your connection over the internet to a printer sends lots of small packets and this makes it latency sensitive. Ideally i think around 10-20ms latency between server and client would be ideal...If you have more than that then you will see some printing slowness.

#4

OK try downloading the trial of usblyzer.com or similar usb software analyzer and you can trace the messages going over the network one by one if you are interested....

> OK try downloading the trial

Too late, I have already purchased. :-D

> OK try downloading the trial of usblyzer.com

I think not needed. Your reasoning is 100% correct, it's just that I had missed.

In effects in, not being a print spool, but a data natively bypassing any spooling mechanism, the single-byte portion when available and not put in a buffer and then "fired."

The ideal solution would be a buffer implemantare program Easylabel of Tharo. But I have already spent months and months (years) to fix a problem on a really basic. Let alone an implementation of a buffer.

Ok, now imagine me as a puppy with big eyes gleaming and supplicants.

by case but just by case, it is that "you" implement a buffer configurable in size and device/port?
....by case but just by case.

:-D

#6

The usb protocol is a essentially a "state machine" i dont know what the next message will be and the message previous very likely influences the next message sent. These messages are generated by the print driver which is proprietary to the printer manufacturer hence the messages cannot be buffered as buffering would stop the flow entirely.

The "download" of fonts or images, which is then composed with a stream of ASCII data. Nothing exceptional or special.

Anyway I'm in trouble. :-)

Patience.

Thanks anyway.