distributed computing implementations

[Robin Becker]

In article <d116fbae.0304100704.68ef4833 at posting.google.com>, Uche
Ogbuji <uche at ogbuji.net> writes
>martin at v.loewis.de (Martin v. Löwis) wrote in message news:<m3r88kk24p.fsf at mira.
>informatik.hu-berlin.de>...
>> escalation746 at yahoo.com (robin) writes:
>> 
>> > If you want the biggest, boldest approach and don't care about
>> > overhead, use CORBA.
>> 
>> I can't agree with that analysis: CORBA has, of all your alternatives,
>> the least network bandwidth requirements.
>
>True, and Mike Olson did some pretty thorough analysis on this.  See:
>
>http://www-106.ibm.com/developerworks/library/ws-pyth9/
>
>
>--Uche
>http://uche.ogbuji.net

I'm interested to find out why the python server/client pair is so
abominably slow when going across our lightly loaded 100/10 Mbs
ethernet.


I modified the time-client.py script to allow setting of the server name
using an environ script.

When using the time-client on the same machine I see
C:\Python\tmp\test_servers>time-client.py
Connecting to ('localhost', 8080)
Time to connect: 0.040000

Sending a long string to the server
Time to send a string of 21000 chars, 0.000000

Recieving a long stirng from the server
Time to receive a string of 22000 chars, 0.000000

Sending lots of ints to the server
Time to send 5000 ints, 32.297000 (0.006459 per call)



On a machine on the same local net  I see
R:\Python\tmp\test_servers>time-client.py
Connecting to ('192.168.0.3', 8080)
Time to connect: 0.000000

Sending a long string to the server
Time to send a string of 21000 chars, 0.000000

Recieving a long stirng from the server
Time to receive a string of 1455 chars, 0.000000

Sending lots of ints to the server
Time to send 5000 ints, 1007.937000 (0.201587 per call)


What am I missing that causes such painfully slow connections?
-- 
Robin Becker