On Wed, Jun 10, 2009 at 11:54 PM, Andrej<[email protected]> wrote:
> I'm a bit puzzled by the whole thing, as the behaviour seems to be > diametrically opposed to things I've seen on Linux/Intel. We ran some > tests, comparing NFS to scp between the same two instances, and > scp beats NFS hands down, which I can't understand, really, by about > a factor of 15. Looking into changing NFS mount options, not that > they differ much from what we'd do on Intel When you do data transfer through scp you would expect CPU for encryption and decryption to be your major factor. If NFS is so much slower, it would be interesting to understand what is holding it back. >From my network measurements I see that effective packet size makes a big difference because of the per-packet cost involved (which seems more significant than moving the data around). If you're able to collect performance data from it, I'm always willing to have a look. Drop me a note off-list to see how we can exchange data. Rob -- Rob van der Heij Velocity Software http://www.velocitysoftware.com/ ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
