Re: Anyone using SQL*Net successfully?

From: <jaakola_at_cc.helsinki.fi>
Date: 30 Jun 92 11:23:56 GMT
Message-ID: <1992Jun30.132356.1_at_cc.helsinki.fi>


David, thanks for your information! I tried to reply to you via e-mail, but our system (mail and finger) somehow thinks your domain name is illegal.

I have gotten some replies to my questionnaire, and I will summarize them Real Soon Now. If you have experiences on SQL*Net (see my original posting), please e-mail them to me, I'll summarize!

I'll comment briefly David's comments:

In article <1992Jun29.171410.19128_at_srg.srg.af.mil>, dtrum_at_jetssrg.af.mil (David Trum x4772) writes:

>Unix clients might be faster.....just a suggestion.

Why?

We have tried OS/2 2.0 and MS-DOS 5.0 clients connected to the IBM mainframe DB server. In SQL*Net-intensive tasks the performance is about the same, because the bottleneck is not the PC. Changing 386SX (IBM PS/2 55SX) to 386SLC (IBM PS/2 57SLC) doesn't help. In CPU-intensive tasks, such as Forms 3.0 form generation, the generation time drops from 7 minutes to 6 minutes (386SX) or from about 4-5 minutes to 3-4 minutes (386SLC) by using OS/2. We need at least 80% better performance than that 386SX-mainframe system.

>Have you tried tuning your Oracle instance ? Unix (esp for multi-users)?
>Your SQL?

That's what we have been doing last 1.5 years... :-)

>How many simultaneous ACTIVE oracle users? What kind of Unix db server?

I speak of the PC-APPC-mainframe system. During the test I was the only Oracle user and the mainframe was virtually empty. And the response time was miserable. By the way, we tried a forms-benchmark, where time to execute query+post-query trigger dropped from 9 seconds (APPC to mainframe Oracle server) to 2.5 seconds (netbios to OS/2 1.3 Oracle server) by changing the server.

>|> I have seen a system like this which used an IBM 4381 mainframe as a
>|> server running RDBMS 6.0, SQL*Net for APPC 1.1 and Forms 3.0 on PCs. The
>|> response time for questions 1 and 3 was bad and connect time was over 30
>|> seconds. A single SQL statement took generally at least 2 seconds to
>|> execute, and a trigger with 6 SQL statements took at least 17 seconds to
>|> execute. This was measured on a weekend, when CPU load was very low (the
>|> tester was practically the only user!) and the SQL statements returned
>|> just a single row each and used carefully tuned indexes! So I want to
>|> know if there are better configurations or should I forget SQL*Net for
>|> heavy production use.
 

>|> Please reply via e-mail. I will summarize if anyone requests.

I will summarize, there has been some interest. If anyone has experiences on SQL*Net with PCs as clients, please tell me!

--
Juhani Jaakola
jaakola_at_cc.helsinki.fi
Received on Tue Jun 30 1992 - 13:23:56 CEST

Original text of this message