| This highly depends on how yu meassure it,
I have seen several tests where RMS outperformed RMS with factors.
But If the customer is looking at RMS WITHOUT journaling compared with
Rdb, then MAYBE RMS can be faster soing the same thing.
THisngs to wathc out for is:
RMS OPEN CLOSE compared against rdb ATTACH/FINISH (RMS much faster as
it is doing letarally nothing compared to Rdb)
Searct one record in RMS against search one in Rdb.
If this is the first search in the table, Rdb will have to load
metadata, check table ACL's, column ACLS aso. so RMS always faster.
BUT IF you let Rdb load the metadata, attach to the database and then
compare, I would be surprised if RMS would be able to hang on to Rdb's
speed in any respect (insert, update and delete). This is bacause of
- Better indices
- Better usage of indices
_ Better Join strategies
_- Better buffer management
So I would bet on Rdb as the winner in comparable tests.
An Rdb fan (though RMS is good for TEMP files aso but real data, noooo)
Lars
|
| I think it's very dangerous to get sucked into comparisons between RMS and
Rdb/VMS. That's like comparing an apple and a banana. (Or maybe an armadillo
and a wombat?)
In some applications, RMS will outperform Rdb/VMS, in others Rdb/VMS will
outperform RMS. In general, RMS journaling seems to be slower than Rdb/VMS',
from what I hear.
Rdb/VMS is doing a lot more for you, however, in terms of consistency,
manageability, flexibility, and a lot of other dimensions. And you can write
high level queries using Rdb/VMS much more easily than you can in RMS.
On the other hand, if you know what you're doing, and don't really have true
database needs, you can probably write yourself a very fast system using RMS.
If you want to expand it, it may not be very easy, but that's part of the
requirements you determine before you start.
|