As I said before once...

Copying via nfs is fast:

10.0.0.1:/usr/exports/bsd  17078856  2183217  14041696    13%    /mnt

-bash-2.05b$ time cp /mnt/netbsd ./test

real    0m8.190s
user    0m0.000s
sys     0m0.348s

-bash-2.05b$ ls -l test
-rw-r--r--  1 klingler  users  6017828 Oct 19 23:10 test

which gives a throughput of around 5.7MBit/sec


Same file with http:

-bash-2.05b$ time wget http://10.0.0.1/netbsd
--23:14:14--  http://10.0.0.1/netbsd
           => `netbsd'
Connecting to 10.0.0.1:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 6,017,828 [text/plain]

100%[======================================================>] 6,017,828     52.15K/s    ETA 00:00

23:16:07 (52.15 KB/s) - `netbsd' saved [6017828/6017828]


real    1m52.797s
user    0m0.088s
sys     0m0.703s


So why a big difference between NFS and other protocols on LAN?


rick