This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Nancy, You can get several nicely configured machines for $5K. We are running RedHat Fedora Core 3 here, which supports 64 bit processors in the 2.6 kernel. Redhat Fedora is free. Depending on your campus, you could go the supported RedHat Enterprise OS route. We generally recommend that you probably want to consider 64 bit processors for the future, such as the Xeon or Opteron64. Dual processor systems are quite affordable, and in the case of an LDM server that is also doing decoding/filing, and product generation, you will get the use out of a dual CPU machine. User desktop systems can go either way. For running models, the dual CPU systems are definitely going to be the way to go- and memory should be as much as possible in your budget. For an LDM server, it is a good idea to have as much RAM as the size of your product queue, and your queue should be able to hold an hours worth of data. If you were getting all IDD feeds, that would be up to 3GB an hour at present rates. Since you are receiving a smaller subset of data at this point, 2GB is OK for now. For a robust LDM server, just surfing the web, a Dell dual Xeon 2.8GHz Power Edge SC1425 with 2GB of RAM is running $1377 (my link was following the small business rack server links http://configure.us.dell.com/dellstore/config.aspx?c=us&cs=04&kc=6W300&l=en&oc=sc1425&s=bsd#bottom_anchor That link may not get you the same page if it was using cookies etc, but it gets you in the ballpark. >From starting around $1377, you can ad a couple of 250GB disks for about $400. For desktop systems where you'd need a graphics card and display, you'll probably add $750 for a 19" flat panel and graphics card, but be looking at tower cases, and smaller hard disks, maybe 36 or 72GB if the data is primarily on the server. You could probably get a couple of desktop systems with nice monitors etc for $1500. GEMPAK will run on 128MB, but You'd probably never want to get a new workstation with less than 512MB. You'd be happiest planning for the future with 1GB for a desktop system. If you plan on using the Java based IDV, I'd suggest 1GB. If you have a specific plan for number of seats in a lab, let us know and we may be able to get more specific, or if you are generally looking at remote access to the server that is doing LDM and modelling, then rack machines are a good use of space. Steve Chiswell Unidata User Support On Thu, 2005-04-07 at 15:05, Nancy Selover wrote: > Steve, > Do you guys make any hardware recommendations for LDM and > Gempak? We are going to dump the Compaq dinosaur after some RAID > failures, and get something faster with more disk capacity. We are > looking to spend $5000 or less. Since we can't get this all to work > on Mac OS-X, we will go with whatever flavor of Unix/Linux/Solaris is > recommended. We will be moving the operation to a location with a > Unix fluent system administrator, who will go with the recommendation > he's most comfortable with, but he doesn't know what LDM and Gempak > need.. Any ideas? I don't think we want to get into a component > system, but buy a package. > > If you don't make hardware recommendations, can you give me a > good idea of what specs I should be looking for in terms of processing > speed, memory and disk space for smoothly running LDM and Gempak. We > typically archive all the day's data including 4 radars, all zipped > and tarred. Then we move it to a tape backup on a weekly basis. We > also run MM5 and will be running WARF soon. > > Thank you for the advice. > Nancy > > Nancy J. Selover > Asst. State Climatologist > Office of Climatology tel: 480-965-0580 > Arizona State University fax: 480-965-1473 > Tempe, AZ 85287-1508 e-mail: address@hidden > >