Malay wrote:
> We are planning to buy GCG package for our institute. We are not sure which
> computers to buy and what should be the minimum configuration. Our budget is
> limited. We would appreciate any comments in this regard. We also expect
> comments on Seqlab and Seqweb, the former being the GUI of GCG and whether
> there is any free or cheap X-windows server, capable of running Seqlab.
Hi Malay,
We are running GCG 9 currently but will be upgrading to 10 in the next
couple of days.
Our current hardware is a Sun E450 with 512Mb RAM, 2 processors and 50Gb
disk, and the only problems we have been having is with a lack of disk
space. The GCG Data is around 15Gb at the moment, but is expected to
grow out to 22Gb in the next 12 Months (I think). Also, if you plan to
process your own database updates daily/weekly, then you will need to
keep local mirrors of the databases in their raw format, and have enough
room to process the raw data to GCG format, so aim for plenty of disk
space. The prcessing power of the machine is fine for the number of
people currently working with the system. If the E250 was available at
the time of our purchase we may have gone with that instead (would have
saved $AUS12000 on the case), but I guess that the E450 leaves us with
more room for expansion in the future.
For an X client, we use VNC from the AT&T research labs (was the
Olivetti and Oracle Research Labs). It is free, and works really well.
It basically creates another display on the host machine and then just
sends a "picture" of the display to the client, so you end up with a
dislpay in a window. Nice thing is that you can get clients and servers
for Dec, Solaris, Win32, Mac and Linux, and the clients are really small
(I can carry win32 client around on a floppy for emergency access to the
Sun Server). The advantage of this software is that it is not as
network intensive as running a remote X display. Anyway, have a look at
the web pages, they explain it better than I can:
http://www.uk.research.att.com/vnc/
There is also an alternative to SeqWeb called w2h, we have installed it,
but haven't really tried it out propperly, so I can't comment on it.
The web page is: http://industry.ebi.ac.uk/w2h/
Anyhow, Hope this information was helpfull. Email me directly if you
want any more information.
Regards,
James.
James Mercer
Systems Administrator
Murdoch Bioinformatics Research Group
Murdoch University, Perth, Western Australia.
http://arginine.it.murdoch.edu.au
