Re: Web apps on the AS/400

WOW,

I'm glad it was a long holiday weekend in Ireland, I'm sure there are some
great debates in there but it's difficult to attempt to digest them in one
go! Anyone volunteering to summarise?

Anyhow Nathan & Brad asked early on if I'd share more info. on the stress
tests I quoted at the start of this debate so here is brief summary-

Purpose of test
----------------------
To establish bottlenecks in an Extranet solution based on following
equiptment:

Web Server- 270 (1024M/25gb) running WebSphere HTTP server & LANSA
Application Server- 270 (1024M/25gb) running Websphere & LANSA
Database Server- 720, only accessed by Java/WebSphere inquiry programs (via
MQ), LANSA programs execute on App. server & also access databases on
application. server.
Firewalls in front & behind Web server.

Test procedure
----------------------
Simulate up to 45 concurrent users over 10 min. periods by generating
transactions automatically using scripts at a remote site.

Findings (at max no. of users)
------------
No stress on Webserver, Max CPU%= 6.9
On Application server-
LANSA: Max CPU%=25, Av CPU%=21.2
JAVA: Max CPU%=71, Av=68 (MQ accounts for over 20% CPU in all cases)
(Load on database server not measured)

The Java programs are simple inquiry programs passing read requests to the
database server, the LANSA programs are complex validation & calculation
programs updating databases on the Application server. We expected the
findings above to be reversed!

I agree with Jim Franz who said 'Java will overtake CGI someday' and at that
point we will simply let LANSA generate Java code, no fuss, no hassle, no
learning.

Mike Dunnion.   www.mdc.ie



+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---