| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 81.1 | Something | AUNTB::BOOTH | A career of MISunderstanding | Wed Mar 02 1988 21:51 | 13 |
| Oracle has told it's user group that its new transaction processing
software will be (originally) 10 times as fast as its general purpose
database. At that meeting, they quoted the new software rate at
28 TPS (Debit/Credit). That means that the general purpose Oracle
must perform at 2.8 TPS. Oracle rarely quotes a machine size. But
if they are using the numbers, you can bet they were generated on
a high-end VAX. I would add that they are now quoting the new software
at 6 times the rate of the old. That would be 16.8 TPS.
If the 2.8 TPS figure is correct, then Oracle is approximately half
as fast on Debit/Credit as Rdb/VMS.
---- Michael Booth
|
| 81.2 | TPS is becoming as meaningless as MIPS | 32371::BERENSON | Rdb/VMS - Number ONE on VAX | Thu Mar 03 1988 16:08 | 12 |
| DebitCredit performance comparisons are meaningless without a full
description of how the benchmark was run. Are the numbers for 1 second
95% response times (the benchmark spec), 1 second 90% response times
(the Tandem way of reporting), maximum throughput independent of
response times? Is it just Database performance or is forms processing,
etc. included? Where is the response time measured, at the backend or
at the terminal. On and on.
Depending on how the measurements are done, how tricky one is willing to
be, etc., the Rdb/VMS V2.3 DebitCredit numbers run from 2-9 TPS on an 8700.
Who knows what the range is for ORACLE or which numbers they are quoting
for *TPS.
|
| 81.3 | Another source of information... | BISTRO::KIRK | Diamonds on the sole of my keyboard | Tue Mar 08 1988 15:36 | 9 |
| Jordi,
You may also want to take a look at the other products that Oracle
may be selling with the database, such as SQL*Forms, and compare it to
Rally. Take a look at
BISTRO::RALLY_COMPETITION
Richard
|