[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Network ssh tunneling speed test



Well,
the file i was sending is a paw-data analysis file, it's not an ASCII file. I don't have an ASCII file as big as that one (150MB), neither the network is that busy.... :-\

Regards,
MC

Radu Brumariu wrote:

   The speed of your file transfer is influenced also by the medium on
which it's getting transfered. So if you're on a busy network, your
timing may not be accurate; unless that's your intention( test on a busy
network ) . It may be the case of a faulting network device, that's of
course not intended...
   Is there a difference in sending an ASCII file over sending a JPEG ?
I don't think so, but I am not an expert in encryption.

Radu

Marco Calviani wrote:

Thanks however,
  now i've increased from 1.5 MB/s to 2.1 MB/s switching from 3des to
blowfish. As i understood there is not so much difference in security
profile between these two algorith.

Regards,
MC

Radu Brumariu wrote:

So why don't you do something like this :

time scp -c <alg> file1 user@machine:

It wil tell you the time spent in user space / kernel space , etc.

It's still not clear what exactly you are trying to measure, and this
thing won't do for you. If you want intimate information from scp, such
as how much time it spends in a certain routine, then recompile ssh with
profiling support and run the above commands inside gprof .


Radu

Marco Calviani wrote:



Robert Vangel wrote:

Marco Calviani wrote:

Thanks for your help, but i need some more precise benchmark other
than this. I wonder if there exists some "time" command. I'll try
googling another time.

Regards,
MC

What do you mean more precise? Once the file(s) have finished
copying, it tells how the speed of each file and how long each one
took...
There is a large spread of speed if you repeat the measure over time.
I did not know on what it depends. I would like a more appropriate
tool, to have a look at the various connections that occur. For
example different speed with different crypto algorithm....

Regards,
MC










Reply to: