|
Richard, with all due respect, what you really have said in your note is that JDE implementation using SQL/ODBC is horrible. This cannot be any proof or disproof that SQL is inherently good or bad on AS/400. Lo's message was that properly designed SQL code can perform reasonably well on AS/400, and I tend to support his point of view. Alexei Pytel > -----Original Message----- > From: Richard Jackson [SMTP:richardjackson@richardjackson.net] > Sent: Thursday, August 31, 2000 11:43 PM > To: MIDRANGE-L@midrange.com > Subject: RE: RPG native to SQL > > "By the way, there is nothing in the AS/400 SQL implementation that would > make it as inefficient as you're implying." > > What experience do you base that statement on? I spent five years > benchmarking the JDE World software applications (RPG) and then three > years > benchmarking the JDE OneWorld applications (SQL/ODBC). Between 1992 and > 1995, I owned CPU sizing for JDE. During 1996 and 1997, I performed > additional World software benchmarks for JDE and customers. In 1996, I > performed the first large scale OneWorld benchmark (24 users) on AS/400. > I > then performed the same benchmarks on HP/Oracle, Digital NT and Alpha, and > IBM RS6000, AS/400 and S/390 MVS and DB2. We reached about inquiry-only > 1,500 users on MVS. I have performed other large-scale benchmarks with > World and OneWorld code as recently as late 1999 and continue to do > performance work on both products when they run on the AS/400. > > From the database point of view, JDE OneWorld (SQL/ODBC) performs about > the > same IO operations as World does using native IO - READ is replaced with > SELECT and so forth. The average CPW per JDE World (RPG) user is about > 0.65 > (the last measurement that I personally reported to JDE management was .57 > but that was for a product release about 5 years old so I rounded up). > The > average CPW per OneWorld (SQL/ODBC) user ranges from 3 to 45 with a mean > around 10. For OneWorld (SQL/ODBC), the numbers are for database only, no > application code was running on the server. The 0.65 number for World RPG > includes the RPG code and the native database code. > > My statements about memory are based on the same three years of OneWorld > testing. Earlier this year, I spent three weeks working on a problem > where > 275 users did not comfortably fit into a 11 gigabyte memory pool dedicated > to ODBC - only ODBC code ran in that pool. Under certain abnormal > circumstances, three to six jobs in this pool would make extremely high > memory demands and non-database faulting reached 1,200 per second in that > that pool. These events lasted from 5 to 20 minutes. Interactive (that > is > to say, ODBC) jobs would slow way down and batch throughput was seriously > effected. > > We might disagree but I think that I know what I am talking about. > > Richard Jackson > mailto:richardjackson@richardjackson.net > http://www.richardjacksonltd.com > Voice: 1 (303) 808-8058 > Fax: 1 (303) 663-4325 > +--- | 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 +---
As an Amazon Associate we earn from qualifying purchases.
This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].
Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.