Greetings,

Does anybody have any literature that compares the performance of
IDMS and DB/2? Specifically, I am looking for CPU usage or MIPS
required to perform like tasks. That is to say, if you had two
functionally identical databases running side by side, which one would
use how many MIPS or CPU hours versus the other to do the same basic
work?


Many thanks,
Joel Zecher

"jtzecher" <joel.zecher (AT) ge (DOT) com> wrote


Isn't I/O the big bottleneck now? Even 15KRPM isn't enough anymore.
Heh, anyone try using solid-state drives on a DMBS server yet? :-)

jtzecher wrote:

For the same functionality, DB2 will require about an order of
magnitude more cycles than IDMS. Always been that way. Don't expect it
to change.

But it takes two orders of magnitude more labor effort to add an
equivalent piece of new functionality to an IDMS based application than
to a DB2 application. Always been that way. Don't expect it to change.

DBMS_Plumber wrote:
Thank you, DBMS_Plumber. I had approximately the same conclusion but
was just wondering if anybody formalized any part of that into a study
of some sort.

DBMS_Plumber wrote:
I'm not surprised with your first conclusion regarding performance, but
I have not seen any shop that has moved from IDMS to DB2 reduce their
number of developers (although there might be some, it certainly is not
the norm). It could be that they are doing more with these developers,
but I would suspect that it is not an order of magnitude more than what
they could have done on IDMS.

Can you point me to any emperical data to even suggest that less labor
is required to add an equivalent piece of new functionality using DB2
instead of IDMS?

Thanks. --dawn