If you are comparing costs, there is a new TCO study out from IDC, at:
http://www.ibm.com/servers/eserver/iseries/conslt/pdf/idctco.pdf
It definitely shows strong bias towards the iSeries, but gives an idea of
the issues and costs.
As has been pointed out by others, it is difficult to get an apples to
apples comparison. If you say you want to compare an iSeries $ 70k box to
an Intel based $70k box...what do you include in the costs? At a minimum
O/S and DB licensing should be included, and any other extra software you
need to actually run things, but there are a lot of other costs that it
could make sense to add in. Likewise, if you say you want to test a 1-way
against a 1-way, that isn't really apples to apples either, since the
architectures are different, RISC vs CISC. And because of that, you can't
really compare processor speeds.
We do a lot of testing with various configurations and platforms and to be
perfectly honest, I don't see a lot of difference in real performance.
Response time is usually about the same. It can vary dramatically,
depending on what version, SP, fix pack, PTF level, etc, but in general it
is pretty close between iSeries, xSeries and pSeries. The thing that
differs is the number of users. sometimes we have a big pSeries in house
and it scales higher than any of the others, sometimes we are working with
a large iSeries box and it runs more users.
I've posted before that I work in a lab environment and not in a real
production setting, so my observations don't apply to all pieces of the
software in a real world environment. (plus I work for IBM, and have an
iSeries background, so I'm a tad biased). But I don't think performance is
really a critical factor in platform decision. You can probably get the
same performance out of any of the platforms if you tune it, put the right
service packs on, etc. (You might even be able to get pretty good
performance out of some of the non-IBM hardware if you work at it enough!
The platform choice should be based on your companies expertise,
strategic directions, other applications, % of downtime tolerable, need for
HA, effort of maintenance, relationship with vendor, stability of vendor,
total costs and all of the usual business cases.
Anyway, thought the IDC study might be worthwhile for those comparing.
Best regards,
Rob Jump
Sizing Specialist
IBM/J.D. Edwards International Competency Center
303-334-1054
[email protected]
www.ibm.com/erp/jdedwards