Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Hardware

Prioritizing Computer Replacements? 46

Posted by Cliff from the swapping-ordinality dept.
cuzality asks: "The public library I work for is (hopefully) going to get some money in the coming months to replace some of our computers. Since we can't replace all of them at the same time, we need to decide which machines to replace first. We have begun working on a survey instrument to help us accomplish this, but since this is the first time we are replacing computers (most were bought in 1999), this is also the first time we have tried to put together a survey for this purpose. What kind of criteria have you used to prioritize replacement of computers? Examples of surveys you have used would also be very helpful..."
This discussion has been archived. No new comments can be posted.

Prioritizing Computer Replacements? More

Prioritizing Computer Replacements?

Comments Filter:
  • at least that's what we always do....
  • To know how to replace the computers, you're doing a survey. But to do the survey, you're surveying slashdot. But you needed a computer to do that.

  • This survey... (Score:3, Insightful)

    by hookedup ( 630460 ) on Wednesday July 14, 2004 @08:40AM (#9695576)
    Is it for the general public using the computers? Or workers for the library?

    If it's for the public, perhaps leave some photocopies with a few simple questions.

    "What are you using this system for"

    "What software would you like to see offered"

    Etc..

  • Usually (Score:1, Funny)

    by Anonymous Coward
    I push to get mine replaced first...
    • Reminds me of what I used to do, when I was a one-man IT department. Everyone wanted a shiny new machine. So, they got the shiny new machine... with my old motherboard.

  • sounds good enough (Score:3, Insightful)

    by dotgod ( 567913 ) on Wednesday July 14, 2004 @08:43AM (#9695605)
    a public library with computers built in 1999? i thought all libraries had to have a minimum of 15 year-old technology. count your blessings.
    • nope, any library that was smart applied to the Bill and Melinda Gates Foundation and got a grant for new networking equip, several pre-setup gateway 's w/ encyclopedia software, and a new windows gateway server. All for free. You know he has to give away millions a DAY just to avoid tax issues. He actually saves himself money by giving it away. :)
      • "any library that was smart applied to the Bill and Melinda Gates Foundation"

        Except that when the support period runs out, they won't even talk to you. Plus, they didn't ship the library I support any original disks or images to reload the dozens of games and other programs that came with them in case you had to reload it. So when the pre-loaded NT 4.0 finally crashes we have no way to rebuild the machines as they were shipped. I know we could Ghost them, but we shouldn't have to. The disks and/or images

        • jesus, its FREE HARDWARE. I know we could Ghost them, but we shouldn't have to...... Good lord man, thats like "I know I should run windows update, but damnit, I shouldnt have too" Take the free hardware for what its worth. How can free servers, 15 free PC's w/ monitors, and 3000 dollars to upgrade our network be a bad thing?!?!?

  • hmm! (Score:3, Insightful)

    by denominateur ( 194939 ) on Wednesday July 14, 2004 @08:48AM (#9695644) Homepage
    Now quite frankly.. what kind of stupid question is that?

    1. Are there broken computers or computers which are beginning to show critical signs of wear?

    2. Are the computers sufficiently fast for the tasks assigned to them?

  • basic criteria to determine (Score:4, Informative)

    by truffle ( 37924 ) on Wednesday July 14, 2004 @08:58AM (#9695711) Homepage
    How many hours a week is each computer used?

    What functions does each computer perform (surfing web, accessing custom library application, etc.) and what's the breakdown of those functions in percentages?

    Assemble a list of all functions the computers are used for, determine how each function would benefit from running on newer hardware.

    What new functions would the library like to incorporate utilizing newer computrs? Assemble a wish list, and then determine through a panel which of those functions are likely to be implemented in the next year if new computing resources are available.

    With this information you should be able to determine which computers will benefit most from an upgrade. Other factors to consider are that it's more estheticly pleasing to have similar computers together (mild preference for grouping computers of the same type). Publicly usable computers that are upgraded can make a positive statement to library patrons (preference for visibility).

    • Excellent starting point.

      I might add that computers which, by nature of their hardware configuration, are more difficult to maintain (trouble incidents per hour of use) should be ripe candidates for replacement.

    • Here's how it goes here: 1st: Management who read's word documents and plays solitaire get new systems 2nd: Relatives of management get new systems, their old systems are handed down to people who do all the work, their systems handed down to the shop, and the shop's systems thrown away and lit on fire. 3rd: Anyone that actually needs a new system is told that we already bought systems, and dont have the cash to get new ones.

  • Scoring (Score:2, Insightful)

    by mwvdlee ( 775178 )
    If you ask the public, they'll probably want Dual-64bit processors and the latest GFX card or something else they don't need.

    Please note that you need not just "replace" but you could also move old computers to less demanding tasks.

    Ideally your survey would end with a number of scores; the suitability for the task it is used for and the importance of that particular system to the library.

    I assume you have some sort of server in the library, if so then you should consider how the CPU power, files and othe

  • Why are you replacing them? (Score:3, Insightful)

    by furry_marmot ( 515771 ) on Wednesday July 14, 2004 @09:02AM (#9695754) Homepage
    1. Figure out what you are doing or want to do with the computers. Is that what the survey is for?
    2. Figure out what you would replace them with (including networking, licenses for proprietary software, etc.) and figure out what it will cost and thus how many you can buy.
    3. Figure out which of your current machines are serving you the least in this capacity. If you have a bunch of unstable Win95 or Win98 machines that crash a lot, I'd say those are good candidates. If you have older machines that are stable, consider leaving them alone if it will allow you to replace more of the unstable ones.
    4. Go over the list you make, poke holes in it, and start again. When you're done, show it to others and don't look at it for a week, then come back to it and see what you think of your logic.
    There. Any other part of your job you'd like us to do for you? I mean, how hard was that?
    • If you have older machines that are stable, consider leaving them alone if it will allow you to replace more of the unstable ones.

      Yes, this sounds like common sense that everyone should follow. It's not. Those older and more stable machines are accidents waiting to happen. The first critera I use when considering machine replacement is age. There's a mentality among people who haven't been in the position of systems administration that keeps them thinking "well, if it worked for five years, it'll kee
      • You're right, of course. But the poster's question was how to prioritize which computers to replace because he couldn't replace them all. The thrust of my comment (in intent, anyway), was simply to replace the most broken/unstable first.

  • Replacing machines not always necessary (Score:4, Insightful)

    by sbryant ( 93075 ) on Wednesday July 14, 2004 @09:04AM (#9695762)

    While some machines are broken, or so slow that they need replacing, the time taken to replace other machines is often non-trivial for a benefit that is minimal. Sometimes, simple upgrades have much more effect for the amount of time and money spent. A common one is more RAM, but a lot of users would benefit a lot from a new monitor or a flat screen - or a second screen even. Little things, like replacing ball mice with optical ones go down well.

    You might also look at some the infrastructure. If everything is slow because the network is overloaded, replacing machines won't make any difference.

    In a lot of firms, the top people get new kit regularly (which they don't need), and the bottom people suffer with old cast-offs. Look at the entire setup, and see where people are needing upgrades the most, then replace those parts.

    -- Steve

  • It would probably be cheaper to buy enough new ones at one time to replace all of your old ones, even if you don't do the actual replacing for many of them for a while. You can find nice price cuts in bulk, and the bigger the bulk, the bigger the cut. If you replace half now and half later, you'll spend a whole lot more in the long run. Better to get them all now and slowly phase the new ones in over time, even if this process takes a year or more. If all your users will be doing is some casual web browsing
    • That's understandable, but you'd still want to test each one, that way you can get the parts replaced before you have to deploy. Obviously you can do this before the install, but depending on how you put the machines together (piece by piece, or complete machines ala dell?) it may be a pain....

      Also, you might find that even if you can get the prices in bulk to come down, it may be cheaper still to wait one year and buy then. At that point, you'll get more for your money. I know the library probably won'

  • How we've done this before... (Score:5, Insightful)

    by brigc ( 30780 ) <brigc@bcmccoy.com> on Wednesday July 14, 2004 @09:17AM (#9695846) Homepage
    FWIW, I used to be automation consultant for a regional library system, and we generally recommended folks look at things in this order:

    PUBLIC machines first
    As much for PR reasons as any.

    SHARED STAFF MACHINES second
    Usually most cost-effective distribution.

    DEDICATED STAFF MACHINES third
    Meaning the director's desktop usually gets upgraded last! :)

    Within each category, we then looked at how old the BIOS was and how much RAM each machine had. Then we looked at budget issues.

    These are, of course, general guidelines... sometimes things are done differently because you either have a funding source wanting to push a particular program or service... sometimes you have a machine which fails or is unreliable. Common sense should always override a plan, if it hasn't already been incorporated within the plan. ...brig
    • When I was working in the university library as a student employee, trying to code up web apps on a 120MHz Pentium while 2GHz P4s were in public areas used for little more than searching the catalog, I was told that it was because of some sort of funding issue. Easier to get grants for things students touch or somesuch.

  • Why ask here? (Score:1, Flamebait)

    by presearch ( 214913 )
    How about you do your own fscking job? What are they paying you for?
    This is such a difficult job that you need to ask what old computers to throw out?
    Boy that's such a hard problem...

    • A bit nasty, but basically what I was going to say. As IT Manager I can tell you at any given time which machines are the oldest, which are the least stable, which get the heaviest use and could do with being faster, which still have room to upgrade and which are maxxed out. Sounds like you need an IT Manager. Or possibly a new one.
  • 1. Director/Commissioner/Board Members
    2. You or other IT staff
    3. Other important people
    4. Supplement computers in busy areas
    • Interesting...This list is the exact opposite of this one [slashdot.org]. Well, I guess we know who respects their users (ie, understands WHY they have a job in the first place) and who feels that the library denizens are "one of those things" that the admins must "deal with" from time to time...
  • A new computer every 3 months works for me. ;-)

  • My criteria.... (Score:2, Interesting)

    by dark404 ( 714846 )
    Well, first of all it depends on what machines you have now and what they are being used for. I'm going to assume you have PII 233-400ish systems as I seem to remember those being the mid range back then.

    Assuming the systems are not optimal for what they're needed for, I would:
    1. Replace any broken machines.
    2. Before replacing machines, determine if you need MORE machines than you have. In that case I would buy new machines but not use them to replace anything. I would just add them to the network.
    3. If yo

  • LTSP - Reuse Old Equipment Effectively! (Score:5, Insightful)

    by runswithd6s ( 65165 ) on Wednesday July 14, 2004 @10:34AM (#9696643) Homepage
    If I were given free reign to decide how to design the network, I would quickly install the Linux Terminal Server Project [ltsp.org]. Without going into a lot of details, here are some of your benefits:
    • Reuse older equipment
    • Operate diskless terminals (less moving parts, less breakage)
    • Centralized Management
    • Put money where it counts, in the servers and network infrastructure
    • Leverage the security of Linux (No Windows boxen getting infected)

    IMHO, 90% of the operations office staff and library patrons perform do not require specialized, personal PC's. Diskless "thin" clients are by far the least problematic with respect to hardware failure or OS maintenance.

    Unless you absolutely require Windows clients, which I highly doubt given that most Library databases are searchable via Internet or Intranet sites, stay away from them. If the recent IE exploits hadn't scared you away yet, nothing will.

    • Good points. One of the older (by not really crap) boxes should be put in place as a firewall, and if they already have a firewall that is old, it should be upgraded. You don't want to have to rebuild your firewall when you least expect it. The firewall should *not* be some box you buy and cannot control, but preferably iptables on Linux.

  • Some questions: (Score:3, Interesting)

    by jbarr ( 2233 ) on Wednesday July 14, 2004 @10:36AM (#9696665) Homepage
    1. What are the computers being used for?
    If they are Public Access, you pretty much only need a web browser and possibly an office suite. As long as they are physically robust enough and can be effectivly secured, you don't necessarily need the latest and greatest hardware.

    If they are Internal use, you need to consider what you will be running. For most offices, it means an office suite, a web browser, and maybe some proprietary software. Again, the latest and greatest may not be needed.

    2. What are the current conditions?
    If you have broken computers, then these should be at the top of the list to be fixed or replaced. If you have computers that are working fine and doing the job, then you may not need to replace them.

    3. Do you have specific OS requriements?
    Are you running any proprietary software that is OS or OS-version specific? For example, if you have an accounting package that only runs on Windows 95, your hands may be tied for some of the computers. Can you reduce the cost by moving to Open Source solutions like Linux? Again, if they are publically accessible computers, you need to assess what access and features they will provide.

    4. Are all your OS's the same?
    Can you standardize on one OS and one OS version? It certainly makes management easier if you can keep your OS's to one single type or version. Try to get away from having many "specialized" OS's for different purposes. It'll drive up maintenance costs

  • 1) Look at replaceing the ones which run processes where the speed of the processor matters more than the ram. Or ones that will benifit from better video cards etc.

    2) take the ram out of the replaced ones and max out any of the remaining ones you can.

    3) See how much the ram boost made a difference in step 2.

    Decide if the ones with little effect are hampered by the processor, graphics card, or nature of the programs they run.

    If it's the processor or graphics card, replace the computer.
    If it's the progra
  • I assume you've already created a database that contains all the machines. Without trying to raise the ire of database purists, a spreadsheet would do nicely with one machine on each row. Enumerate the factors that are important as columns, like:
    • Does this machine work?
    • How frequently does it crash?
    • Does the machine behave erratically or consistently?
    • How important is the task it performs?
    • How often is it used?
    • How critical is it?
    • Are there other machines that perform the same task?
    • How much would an up
  • What a great example of how public institutions waste money! "We need to devise a survey" analysis paralysis. The time spend on devising a survey and then doing the surveying could equal the cost of a couple of machines!

    Try this list of priorities
    1. Replace ones that don't work
    2. Replace ones that are slower (replace a 300MHz PC over an 800MHz PC)
    3. Consider upgrading instead of replacing (maybe more RAM or a bigger hard drive would extend the life of the PC)


  • Priority 1: Computers that have stopped working

    Priority 2: FIFO (oldest first)

    Priority 3: Computers too slow to do their current work.


    Or you can be as illogical as ODOT and reverse the above priorities- the plus side is that this frees up some faster machines for redeployment onto slower machines.
  • Pop the cases open, look and see if the have the infamous IBM "deathstar" [techreport.com] Hard Drives in them, however, it seems like most of them that are going to die quick deaths already have. I've been running a 30GB DTLA since Jan 2001 with no problems, but I personally saw 2 crap out in a week at work in late '01.

  • This doesn't apply to all situations, but are you sure replace is the right option? If all the public machines are busy most of the time (or you have no public machines) perhaps you are better off getting some more tables and adding computers.

    Depending on who has what needs, you might replace a staff machine and put it in the public areas. If people start looking for fast machines you might need to start asking why they need speed and give those who really need speed the faster machines, but this is com

  • Replace the machines lottery-style.

    Randomly assign numbers to the machines (perhaps the IP address or last 3 digits of the serial number) and hold a raffle-like event. Getting all managerial and trying to "optimize" who gets what first is just a lot of wasted effort that makes you feel important but doesn't really improve anyone's productivity, least of all yours. Just start replacing machines.

    You're actually kind of lucky. Most library-computer-admin types are barraged by viruses, protect-the-kiddies
  • Haul off any broken/unwanted electronic stuff to Officemax -- they're having a promotional recycling offer.

Slashdot Top Deals

He has not acquired a fortune; the fortune has acquired him. -- Bion

Close