In article <9306050426.AA18145 at net.bio.net>, MBSMITH at ECUVM.CIS.ECU.EDU (jeff Smith) writes:
...
|> answer, but .... we are planning to get a new computer/work station
|> to be dedicated to running the GCG software and a few other molecular
|> biology programs such as gel readers. At the present time we have about
...
Make a list what you *REALLY* need in three phases :
1) current status
o what is the currently available software
o why do you think that you should abandon the current setup
o what is the maximum you are able (not 'willing') to pay for
o what is your user community wrt to ambition and flexibility
2) current desires
o what software do you not currently run but have requests to do so
o is there anything you cannot do in your current setup (devices?)
o what is the growth rate of resources you need to cover this year
o are you politically backed to do what you want or are there
constraints
3) future
o if you have a bigger machine, what could you afford to run?
o would your users be able to switch to another OS
o would your licenses be transferrable to another OS
o what is the anticipated cut in running cost for your service
(don't tell me that your budget is always growing!)
o do you need to run all and everything on a monster or can
you affor to set up a cluster,farm, mainframe+PC, etc.
Once you have made that list, revise it as brutally as you can.
Next, set up a profile of needs (without thinking of a model
and a vendor) and lower the amount of money you want to spend by 30%.
Give this list to vendors and ask for ___SOLUTIONS___ not
for a paricular hardware and software. If GCG is a prerequisite,
mention this. Be aware ofthe fact that many companies approach
the end of the fiscal year at some day, and might not be
always of the same responsiveness.
Once you got the replies, ask for references of the top three
choices and intervew the site managers there about their
confidence and history why they did choose this and not another
way of going. Leave the top three in the list and ask your
environment (computing center, building department etc)
whether any of the three might cause problems.
If you are confident that you think that you know at least
two top candidates, ask your users. Not in a poll on something
they do not know, but pick those high-end guys who know what
they do ans ask them whether they see problems or benefits
of any of the choices. Make clear to them that pricing is
also an issue, as well as maintenance.
This approach make take you some time but I am sure it will
pay off. I personally run GCG software on both VMS and SGI
clusters and support all other platforms where GCG runs
on. I would not claim that I know all and everything but the
site-specific environment has such a large effect on the
choice that I think it would be unfair to tell you that I
am satisfied with what I have and you must buy similar.
Reagards
Reinhard
--
+----------------------------------+-------------------------------------+
| Dr. Reinhard Doelz | RFC doelz at urz.unibas.ch |
| Biocomputing | DECNET 20579::48130::doelz |
|Biozentrum der Universitaet | X25 022846211142036::doelz |
| Klingelbergstrasse 70 | FAX x41 61 261- 6760 or 267- 2078
| CH 4056 Basel | TEL x41 61 267- 2076 or 2247 |
+------------- bioftp.unibas.ch is the SWISS EMBnet node ----------------+
-----------------------------------------
