Network file transfer rate comparisons

[Chris Lindblad <cjl@WHEATIES.AI.MIT.EDU>]

    Date: Thu, 26 Jan 89 10:09 EST
    From: Len Moskowitz <Len@HEART-OF-GOLD>
 
    It seems to me that your performance test will reflect disk access times just as much as network
    performance.  I tried it with a 3675 accessing its own two large drives (a 515 and a 470) using
    NFILE and found the following performance:
 
    (io-performance-test (fs:parse-pathname "h:>len>temp>temp.temp"))
    Wrote a file of 1048576 bytes in 15.166667 seconds for 69136.875 bytes-per-second.
    Read a file of 1048576 bytes in 12.466666 seconds for 84110.375 bytes-per-second. 
 
    These times are roughly twice as fast as your 3650.  Assuming the network end of things is the
    same on both our machines, the difference is made up by the disk, no?  When you're comparing the
    different network protocols, you might keep in mind that disk performance is a significant
    factor.
 
    I tried the test on DNA's DAP between the 3675 and our VAXen:
 
    (IO-PERFORMANCE-TEST (FS:PARSE-PATHNAME "tsd1:disk$irandd:[lm26446]temp.temp"))
    Wrote a file of 1048576 bytes in 37.816666 seconds for 27727.88 bytes-per-second.
    Read a file of 1048576 bytes in 43.683334 seconds for 24004.03 bytes-per-second.
 
Here's the additional information:  The 3650 file server was on the same
subnet as my machine, with the files on 470 MB Fujitsu Eagles.  The UNIX was a
sun 3/280 on another subnet with 800 MB CDC Drives.