Slashdot Mirror
Generic PCs For Corporate Use?
porkThreeWays writes "I work for a government agency supporting about 1000 PCs. The economy has hit us just like everyone else and we are looking at ways to save money. We currently buy Dell computers and even with our government discounts end up spending about $1,000 for a pretty mediocre computer. I had the idea of building our own PCs for considerably less. We'd spec out a standard configuration that we'd use for 18 months. CPU speeds and RAM sizes may change during that time, but socket types, memory standards, hard drive interfaces standards, etc, etc would be required to stay the same. We have Dell warranties right now, but I could see just keeping spare parts on the shelf and building that into the cost of the PC. We'd also be able to transfer Windows licenses because the Dell installs are non-transferable. However, I couldn't find anyone on the large scale doing this. Is anyone on Slashdot using PCs they built themselves on the large scale?"
What is on them, a Core i7 with 12GB of RAM and an SSD?
Me thinks you're overpaying... Dell isn't that expensive, really it isn't...
be careful transferring windows licenses... they're all OEM licenses and the T&Cs don't allow you to transfer them to another machine (ever). Of course this is based upon my knowledge from a few years ago when i worked in the licensing field, so things might have changed (IANAL)
Wat?
With software licenses being what they are, I just don't see this as being feasible on such a large scale. If you were to use open source, Linux, etc., it may be do-able. And awesome.
I had the idea of building our own PCs for considerably less.
This is an awful idea. I had some experience with such an experiment; it didn't work. The computers were failing left and right, and the vendor distanced itself from the situaton. You will first be forced to maintain all that herd, and eventually you will become a scapegoat.
Business is all about using money to make other money. It is a legitimate expense to buy a computer; it's tax-deductible on corporate level, so you don't need to squint too hard at the prices. Buy good computers with a warranty and on-site support and be happy.
Use server based VMs or terminal servers. Then use winterms for the desktops. You can get those for a couple a hundred dollars and they last forever.
Historically, the claim of consensus has been the first refuge of scoundrels.
So long as you have an competent IT staff, you should be good. It's so much better to have internal people swapping out bad hardware and dealing directly with the RAM vendors, etc when possible. Not only do the vendors of specific hardware normally have longer warranties, they're normally faster than the end vendor at swapping out hardware. Think of it this way, if you keep paying Dell to do support and replacing bad hardware, then what is keeping another, less competent person from taking your job? So long as you create a good mechanism for keeping up with machines, parts, and vendor relations, you should be good. I highly recommend this approach. From my experience in government IT, it seems that the ones who go with vendors who provide a lot of support, the employee gets replaced with a drone who just interacts with said vendor. By doing what you're talking about doing, and doing it in a well executed and organized fashion, you are not only doing your employer a favor by saving money, you are also securing your job. You might want to do this in increments, maybe 100 computers at a time. That way you can find problems that you'd never think would ever be an issue. Doing things in small increments at first is probably the best move you can make. Also, think about keeping parts on hand, maybe enough parts to replace all the parts in every 25 computers or so. That way you can fix failed hardware on site, and then worry about swapping with the vendor later. If you keep up with your data, you can find out what fails the most and when, and then you can become more efficient in dealing with vendors, part swaps, and stocking of said parts.
Sig: I stole this sig.
I got our last computer at Best Buy for like 369$, dual core, 2gb ram, 320gb HD, more then adequate for running outlook and looking at craigslist.
software? dell wants like $150-$300 for office + over priced ram (Dell warranties may not like you having 3rd part ram)
We'd also be able to transfer Windows licenses because the Dell installs are non-transferable.
Do you even lift?
These aren't the 'roids you're looking for.
Mgmt killed the white boxes and bought IBM (now Lenovo) for half the machine at twice the cost. Failure rates are roughly comparable (lenovos probably have a slightly higher failure rate).
So, whatever was involved in that decision making process, cost of the hardware and reliablilty wasn't paramount. I think it is because my workplace is such a political environment, and they perceived risk to themselves by buying no-name.
They will set you up for considerably less than $1000/machine, as will Dell. When you add a service contract they aren't much cheaper though, but your equipment can get replaced with just a phone call.
If you want to build 1000 machines, go ahead, but it would probably take 1 person about 6-12 months to assemble them all. If you're spending a million bucks on PCs, hiring someone to assemble them full-time might be worthwhile. figure $700/pc for DIY, and $50k for the tech's salary. That's $750k for a lot less. And you can keep the tech around for support of the equipment instead of buying costly service contracts for every machine.
“Common sense is not so common.” — Voltaire
One thing you are not taking into account is labor. By buying a Dell (I don't like Dell, but this applies to any manufacturer), the time and expense in building is included in the cost. When you look at parts, it looks less, but add your time in building, time to diagnose an issue when the machine doesn't boot, and time to RMA parts and repair said machine when it breaks. This adds up quick. I work for a company that built its own machines for sale and we found a company that could build and warranty them for $10 more. That $10 extra was very well spent. So remember, you are getting more than just parts, you are getting the time to assemble, repair and replace.
I payed 400 for a toshiba a305 that is better than average
Software Assurance contract with Microsoft might actually be cheaper than paying for all those OEM Windows and Office licenses in the long run.
Dell prices aren't normally nearly that out of whack, probably a lot of that cost is Office. If possible the easiest cost savings would be a switch to Openoffice (Libreoffice). Many employees will not be up to the task of learning something new, so it will be a good way to thin the herd.
You might save some money but if you factor in the cost of a Windows 7 Professional license then the small (and I mean small) savings doesn't offset the amount of time you spent spec'ing, purchasing, stocking, and building your workstations. This is because Redmond won't give you the same deal they give Michael Dell.
We provide a software product which we recommend the use of HP rp series Point of Sale terminals. Why? We aren't in the hardware business and when purchased as a bundle for an extra $250 per terminal they can buy a 5-5-5 warranty package on ALL the equipment and all the peripherals. Touchscreen goes bad in year 4, HP overnights a replacement. Receipt printer goes bad, they overnight a replacement. Barcode scanner goes bad, over night a replacement.
We have another company that sells a rebranded version of our POS for a niche industry and they elected to field cheaper equipment they built themselves for less $500 per terminal made up of dual core Atom boxes. Problem is, every couple months they go to order from Tiger Direct or EggHead, it's a slightly different box with different cases/psu's. Plus they order an extra box for every 5 they sell just to have on hand so they can ship it out overnight. And when they do it that, it costs them $100 - $150 in shipping costs. I don't know what they're field rate is, but I believe they are leasing out the boxes to companies for a monthly fee so technically they don't have to release their modified code.
Which I guess works for them, but I don't want to the hassle. And frankly, we've had a couple times where something wasn't working right. HP didn't mess around and just shipped a new unit. Client was happy, we were happy, everybody wins. Granted HP isn't the cheapest, but are competitive in the POS hardware market.
"The problem with socialism is eventually you run out of other people's money" - Thatcher.
and get poor PSU and other cheap parts better to pay more / build your own and get better parts and not some 2gb ram system with on board video (VGA only) that eats ram. Get a least 4gb and on board is fine as long it has DVI / HDMI.
I believe Microsofts EULA states you cannot transfer off that machine. And while with a good tech team, you can upgrade disk, GPUs, CPUs and memory quite easily. But a change in the motherboard or the case will often be cheaper to buy a computer from a supplier since they purchase in bulk and TEST alot better than you can.
Now, on the other side of the spectrum, open source has licenses that transfer to all systems, can interface with all Microsoft and Mac products (if the IT team on the other side knows how to properly configure for them). Need to interface with Ofice, use OpenOffice (still there are issues with the latest versions of Microsoft Office docs such as DOCX and XLSX). You can access calendaring and mail via the web or equivalent Linux mail tools that interface with Exchange servers. And for those who still need to use Windows, you can emulate it via virtualization without having to move to new equipment.
A third option would be to keep these machines all as 'thin clients' and move everything to a served architecture where they access all their applications from a central server/cloud/or clustered environment.
This is my sig. There are many like it but this one is mine.
I think the procurement process is probably the big problem you'll run into. You obviously won't be able to buy a bunch of stuff on an agency credit card, so you'll have to go through contracting. I could be wrong, but I don't think you can buy that much stuff off of CDW-G. And with around 1000 machines, it will be a pretty big contract, which means it will go through contracting even slower than usual. Not to mention you'll waste a lot of people's time going through contracting than just sticking with Dell.
My shop had done this a few years ago and we ran into the bad capacitor problem. The machines all failed within a few months of each other and the warranty on the motherboards was gone. I realize that was a worst case but shit happens.
It sounds like you could compete with Dell and that you should start a company. Maybe then you realise that 1kUS$ isn't that much for a system.
Don't make the mistake of not calculating the effort it costs you to assemble the systems yourself. Say you cost a modest 100US$ per hour to your employer and redo the maths.
You seem to know about hardware. Now consider how you will train co-workers to attain your level of expertise. Will you now be teaching as well? Think of what will happen when you'll leave the company. Don't worry, you eventually will move on to other challenges.
I myself build the systems for my own small business. It's costs me significant amounts of effort which I could put towards paying customers. I only do it because I like it and because I take the liberty to do so. But really, I probably shouldn't.
I hadn't the slightest objection to his spending his time planning massacres for the bourgeoisie... (P.G. Wodehouse)
Unless you are some intern making $5 / hour or something, the amount of time you will spend assembling these things will far outstrip the cost savings.
IE - say you save $200 / machine. How many hours will it take you to build that? Three? Four? Now figure in how much you make per hour. Your "savings" are out the window.
You can get entire Dell machines considerably less than $1K if you move over to the Vostro's, Quality of the hardware is less, but you still get them built to order and delivered as opposed to having to assemble the parts. They also use off the shelf components so you don't need to go through Dell if you don't want to for some of their proprietary parts. Personally I think they're crap hardware, but anything in this price range usually is if all the computers I get in for repair is any indication.
You're crazy if you think it's worthwhile building the PCs yourself. You can easily find an off-the-shelf PC for considerably less than $1000, probably less than $500, and unless you have a team of at least 6 people sitting around with nothing better to do then you won't save money building them yourself, and you'll just cause yourself a massive headache. Simply commissioning 100 pre-built PCs (presuming you're planning to replace 10% of them at a time) is plenty of work for a support dept., even if you're not making massive software changes.
Good fucking luck. If your self-built computers fail, it's your butt on the line. You may say "but $COMPONENT has a 3-year warranty!", but vendors are great at pointing fingers at other vendors unless you test enough to prove it's their component's fault. So there's some time wastage.
Then you have to learn to deal with support departments from n different vendors, rather than just the one OEM's.
But before you even get there, you have the enormous time outlay of building each computer by hand and (presumably) testing them to make sure they work, and possibly dealing with warranty replacements right away. Also? You have to buy retail Windows licenses to make them transferable, which is a few hundred dollars per copy.
It'd be a lot smarter to set up a site-license agreement with Microsoft. We can upgrade any computer to any version of Windows, provided it came from the factory with a Windows license. Don't know how much that costs - not my department - but it's really worth it to be able to say "OK, we've tested Windows 7 and it works well, so here you go". Also, that saves us $50 per new Dell Optiplex because I can get it from Dell with Windows Vista Home Basic, then install our Win7 Enterprise image. On top of /that/, you might find that if you set up a bulk order with your Dell sales rep they could cut you a better deal - that's happened to me before.
Hail Eris, full of mischief...
E pluribus sanguinem
Do the math on the trade-offs.You have not mentioned what your organization does, but building PCs is generally low-value grunt work. (Built a couple myself because I wanted specific capabilities not inherent in a generic PC.) There a few questions you need to ask beyond just the capital cost of sourcing PC parts. How much is your (or your peoples' time) time worth? (There is an opportunity cost when using internal staff to build PCs.) Do you have people sitting around not doing anything else? (If you do then it's probably more cost effective to let them go - most costs are labor, not equipment.) Out of a 1000 PCs (as quoted) how many have you made a warranty claim over the past year? You'll have to write those claims off and factor in internal repair costs if you take on responsibility for the PC hardware. Spare PC inventory for possible failure is generally wasteful in that you carry the responsibility for capital purchasing and holding costs. Can you do a better job at predicting PC parts failure? Do you curently possess that information or can you obtain said information for your organization? I'm sure other slashdotters can come up with more questions to ask. I understand that my phrasing sounds negative but unless you know the answers I'd say you'll end up increasing total costs (CAPEX + OPEX). Good luck
I would recommend trying to get into the Western States Contracting Alliance, WSCA, (http://www.aboutwsca.org/content.cfm/id/WSCA) and get reasonable HP workstations for ~$600 and LCD's to go with them for ~$200. You get a major name brand and save $200 off what you're currently buying from DELL.
Even if you don't buy from DELL, see how fast they lower your pricing when you have a quote from HP.
Additionally, WSCA is pre-bid, which means most government agencies don't actually have to re-bid, they can just order off this contract that's already been run through the whole bidding process.
Awk! Pieces of eight. Pieces of eight. Pieces of seven... ERROR: General Protection Fault. [Paroty Error.]
I think the answer might just be to try renegotiating your price or specs. I also work for a government institution with about 1000 computers and we pay about $450 with Dell for what I would consider a very decent desktop computer (4GB of RAM, Intel Core 2, etc.).
Are the gov't discounts really that great? Last time I compared(summer of 2008) the gov't(state level) & education discounts available to me they were higher priced than what was currently available on Dell's website.
This situation arise because the prices & configurations are negotiated only every few years.
I'd suggest comparing the pricing of the same machine with & without the "gov't discounts". It would also be useful to know what kind of specs you are looking.
The Dell Vostro series for your basic office worker will you run about $500. As others have mentioned don't bother getting your windows license from Dell. At the volume you are dealing with it, you can a simple enterprise/site wide license from Microsoft.
If the govt becomes a lawbreaker, it breeds contempt for law, it invites man to become his own law, it invites anarchy
That's the current low-end Optiplex, and it's pretty good hardware quality for a pretty good price. IMO, its existence removes any incentive to go Vostro unless you positively must get the cheapest crap.
Hail Eris, full of mischief...
E pluribus sanguinem
I've had the opposite experience at 3 different companies. For the price you pay they are no better or worse than Dell or Lenovo. I did have some pretty bad experiences using a local shop that would build machines to spec. Sure they were exactly what we wanted and cheap, but no matter which local shop I used the quality seemed variable. With more DOAs than I would have liked considering we pay these guys to do burn-in.
All cheap computers are crap and trouble, it's all relative. I think your comment needs to give some contrasting situation for comparison.
“Common sense is not so common.” — Voltaire
If you build your own computers, you are still getting roughly the same parts that Dell/HP would slap in there for a considerable amount less. How is the failure rate going to be higher? It's not! It will be equal. The difference being that you need a tech guy who can repair them when they do fail versus having to call in a Dell/HP tech.
Instead of telling this guy "Oh, this is a bad idea, I tried it and it didn't work." Give him the reasons why. I'm willing to bet the guys that are saying this work for Dell or HP and are just trying to scare him. After all, every sale counts, right guys? If you build a computer right, it won't fail.
And what does a government agency need 4gb of ram and a PCI-e video card for? Check their emails? Search a server side database? Manipulate a spreadsheet? Give me a break. The computer that can get all of these jobs done can be built for as low as $275 each. Pick up an enterprise copy of Windows/Office and be done with it.
FoxConn make most of the world's computers. Seriously. Approach them for what you need and say you need x1000. They'll build what you want without the 3rd party markup.
Do not mock my vision of impractical footwear
- You need to have some very frank discussions with either your Dell rep, or whomever is speccing out your quotes. $1k for corporate-level desktop PC in this day and age is ridiculous; you should be expecting to pay more like $600-700. To give you an idea, I work for a state university, and we're currently giving about $550 for a Core2 E8400/4Gig Ram/160gig HD HP. Integrated video and no monitor of course, but a 3 year warranty. Sure you're not going to be decoding the human genome with that machine but it's more than enough for your average office worker. Don't be afraid to use HP as a club against your Dell rep; they're currently getting hammered by HP in the corporate world, and won't want to lose your account, assuming you're of any kind of size. I wouldn't recommend going to HP unless you absolutely have to though; service is horrible.
- Take some time to consider whether the time spent building custom machines is really worth the time of whomever would be doing it. Chances are, it is not. Either you're going to have someone making peanuts doing the work, or a skilled IT person who really isn't all that interested in doing what essentially is grunt work. In either case, you're going to see problems.
- If you haven't already, you should discuss this with your purchasing department before moving forward. Depending on the level of beauracracy that is entrenched in your level of government, building your own computers may not even be permissable.
You mentioned that you couldn't find anyone doing this on a large scale, this should be a warning flag. Lot of potential problems and pitfalls here, not the least of which is your cunning "transfer the OEM licenses" plan. There are a lot of better ways to save money on computer purchases.
You could go back to Ledger cards!
I'd advise against it. We tried it where I worked. It sounds good to build your own boxes on the cheap, but it rarely works out like that. You build your own computers. The cards (ethernet, video, etc.) you used a few months might not be on sale this month, so you now have multiple versions of cards. If buying them in bulk, the line probably gets refreshed so it's hard to buy the same model of hardware twice. Then, when you have to rebuild an older computer a year or two later, you have to remember where you put those drivers for that particular card that this computer uses. Since it was bought on the cheap, it probably isn't marked very well and unless you had the luxury of looking at the computer before it went down, what model it is might not even be known. Even then, since the hardware was bought cheap, the drivers might not be as easy to get online as one would think, especially if the company isn't around anymore. There is also all sorts of tiny details dealing with this or that hardware that has to be remembered. Then you need storage for all the bits, parts, and driver software. Trying to call in hardware warrantees for the products you buy will usually be much more time consuming than just calling your vendor and having them do everything based on the serial number of the broken computer. In the end, building and maintaining our own computers was way more trouble and man hours than just going with a name brand such as HP or Dell and using their warrantees. Whatever got saved in material costs in building our own computers got more than spent in extra man hours maintaining them.
There are two main concerns with moving into a self-created solution - standardization and support.
Standardization
Dell may load a lot of bloat-ware on their consumer level machines but for corporations or the Government the Dell X-Image process can be used to standardize the base level install. This process allows for an IT department to build a base OS standard environment applications on a single platform of hardware regardless of the equipment in use. After baselining the system the image is uploaded via a standard web interface on which you also select the hardware platforms being used in the environment - laptop, workstation, desktop, thin client, etc. Dell then takes your baseline and codes it back using their X-Image process encorporating in all necessary drivers for the models you have selected then sends it back to you. It's basically an outsourced slipstream of the OS made extremely easy by Dell. The nicest part if you are a corporation or Government entity? It's free. Contact your Dell sales rep and ask about the process. This allows for standardization not only for hardware via the same manufacturer but also for the OS and applicaitons in use on those platforms.
Support
Dell may be a P.I.T.A. for consumer level services but for corporate or Government they are right there with HP and other high-end channels. Next day replacement part shipment - or within 4 hours based on purchased support for servers - means sites do not need to keep on-hand stock of components except for maybe a few key resources for critical systems. There is also no need to train your support team on how to replace these parts as Dell will send a technician to your site with the part to perform the replacement, test the system following replacement, and take the bad part with them for return to Dell. Considering the cost of the support of the systems is built into the purchase price of the system the overhead support cost is lowered and the staff is allowed to focus on the more "fun" issues related to using Microsoft products in a large scale environment. IF you wish to have in-house parts and repair capabilities, Dell will supply on-site sparing of parts and offer training to your IT staff to perform the actual replacement of parts. The best part here? If your staff is trained and completes the warranty work in house, Dell sends you back a credit of x number of dollars per "call" - I put x as the last time I was involved in contract negotiation was 5 years ago and while it was $40 per incident at that time I would expect there has been some change in amoount.
I work for the Government as well and Dell while in my opinion as an IT professional is annoying and bothersome to say the least, and while I know I could build a better, more powerful, and more robust hardware platform for the same cost, in an environment where standardization, quick support, compatibility, and operational state of my users are all at a premium desire of the customer, I say leave the headache of those messes to upper management and Dell. It may cost up to $250 more per system to have them supplied by Dell but considering standard rate of a technician and taking into account the amount of time needing to be spent on building and deploying, then training and support, combined in with overhead costs for maintaining parts and stock, the cost difference is a loss not a gain.
We use http://www.equuscs.com/ when Dell or HP don't fit a customer's needs.
As others have mentioned already you will need a volume license of windows if you want to transfer it. The nice thing about equus is that you can send them an image and any system you order from them will come with it preloaded. Oh and equus will be able to continue shipping XP Pro downgrade if you should need it.
There are two main concerns with moving into a self-created solution - standardization and support. Standardization Dell may load a lot of bloat-ware on their consumer level machines but for corporations or the Government the Dell X-Image process can be used to standardize the base level install. This process allows for an IT department to build a base OS standard environment applications on a single platform of hardware regardless of the equipment in use. After baselining the system the image is uploaded via a standard web interface on which you also select the hardware platforms being used in the environment - laptop, workstation, desktop, thin client, etc. Dell then takes your baseline and codes it back using their X-Image process encorporating in all necessary drivers for the models you have selected then sends it back to you. It's basically an outsourced slipstream of the OS made extremely easy by Dell. The nicest part if you are a corporation or Government entity? It's free. Contact your Dell sales rep and ask about the process. This allows for standardization not only for hardware via the same manufacturer but also for the OS and applicaitons in use on those platforms. Support Dell may be a P.I.T.A. for consumer level services but for corporate or Government they are right there with HP and other high-end channels. Next day replacement part shipment - or within 4 hours based on purchased support for servers - means sites do not need to keep on-hand stock of components except for maybe a few key resources for critical systems. There is also no need to train your support team on how to replace these parts as Dell will send a technician to your site with the part to perform the replacement, test the system following replacement, and take the bad part with them for return to Dell. Considering the cost of the support of the systems is built into the purchase price of the system the overhead support cost is lowered and the staff is allowed to focus on the more "fun" issues related to using Microsoft products in a large scale environment. IF you wish to have in-house parts and repair capabilities, Dell will supply on-site sparing of parts and offer training to your IT staff to perform the actual replacement of parts. The best part here? If your staff is trained and completes the warranty work in house, Dell sends you back a credit of x number of dollars per "call" - I put x as the last time I was involved in contract negotiation was 5 years ago and while it was $40 per incident at that time I would expect there has been some change in amoount. I work for the Government as well and Dell while in my opinion as an IT professional is annoying and bothersome to say the least, and while I know I could build a better, more powerful, and more robust hardware platform for the same cost, in an environment where standardization, quick support, compatibility, and operational state of my users are all at a premium desire of the customer, I say leave the headache of those messes to upper management and Dell. It may cost up to $250 more per system to have them supplied by Dell but considering standard rate of a technician and taking into account the amount of time needing to be spent on building and deploying, then training and support, combined in with overhead costs for maintaining parts and stock, the cost difference is a loss not a gain.
We Are Stardust...We Are Golden...
I had the idea of building our own PCs for considerably less.
Seriously?
You may very well be able to source the parts for less... But then you're going to have to build and support them.
Dell has an army of minimum-wage employees. If you're a government agency, you probably don't have anybody who makes that little. So the cost for you to build one of these computers will be more than what it costs Dell. And it won't be as simple a build as reloading some Dell box because you'll have to grab drivers for each individual component.
There'll still be a warranty on most of the parts... But you'll have to identify which part actually failed, figure out the appropriate number, and call them. This will require more time and effort than simply calling Dell on anything that breaks. Good luck getting anything even remotely resembling one of Dell's 4-hour warranties.
And any real problems you have are going to come back to bite you in the ass. Get a batch of bad motherboards? That's not Dell's fault, that's your fault. Get some funky driver conflicts? That's your fault. Can't get replacement parts in a timely manner? Your fault.
You might very well see some up-front savings... The sticker price of the box you can build, compared to the sticker price of the same hardware with a Dell logo, may very well be lower. But once you start spending time building and supporting them I think your savings are going to vanish very quickly.
Honestly, I'm not sure what you're complaining about as far as Dell's prices go... We typically spend $1,500 for a machine. The hardware is generally more than sufficient for our current needs and there's always room to upgrade the RAM and CPU at least once. We typically order them with the 3-year warranty. We usually get Office and Adobe bundled, plus whatever monitor they're throwing in. The machine will generally last us 3-5 years depending on which bits of hardware fail in that time.
"Work is the curse of the drinking classes." -Oscar Wilde
So you have some free monkeys, a large static free build room and lockable large supply room?
You must have a _lot_ of free time. Have you catered for your staffing costs? Security costs? Insurance? Do finance want to bring all those costs back onto the books that the Dell single number approach removes?
Yay me!
use amd chips. they're a fraction of the price of intel chips, and there's really no difference in performance.
Sure, you'll save a few bucks...for a few months. Then, things will start failing. You'll find there're no hardware drivers for many of the parts inside, and when people start finding mechanical parts, like pushbuttons, sockets and controls, falling off, you'll discover that the caseworks maker doesn't sell spare parts. Also, as mentioned above, you'll have to pay for Windows licenses (unless you're moving to Ubuntu, too). I've got lots of happy clients, because I keep replacing the crap they have (e.g., the computer with a touch-screen display that overheats every Summer day afternoon) with brand-name products. They pay me my comparatively higher prices (considering the local dolts who call themselves my "competition") because I deliver stable, reliable systems that they never have to worry about (until, like yesterday, a UPS went up in smoke...quick to fix that one). Stick with reputable makers. Avoid the small storefronts that will "build yours" and put in everything cheap, but charge you somewhat less than brand-name products. Heck, I've only been in this computer business for over half-a-century, so my experience probably will be punished by others with a more "home-brew" bent...but you said it's a business, not a bunch of students. You should take a good look at whether, over the entire life (which will inevitably be more than 18 months!), you'll have made a good bargain. I'd wager that "white box clones" will end up saving you nothing.
Try the Dell Outlet. Great discounts and the very same warranty as new systems.
You can't order more than 5 per order, so it sucks to shop, but the savings can approach 30-40% on current model systems.
First of all, if one of the two (or however many) servers and physical switches goes down, they've still got the same problem. There's some large fraction of their workers unable to get their work done. Maybe it's not 100% like it used to be, but it'll still be costly.
Second of all, if the core router goes down, they're probably just as fucked as they currently are.
Third of all, now you're starting to get more and more backend equipment to maintain. This requires far more specialized (and costly!) technicians than the low-skill technicians it takes to repair common PCs.
Look, we tried the centralized model throughout the 1950s, 1960s, 1970s and part of the 1980s. It failed. That's why we used standalone PCs for part of the 1980s, the 1990s, and most of the 2000s.
Unfortunately, that was just enough time for an entire generation of technicians, admins and executives to move on. Now we have dipshits like yourself advocating centralized networks, even though you don't know it's an experiment that we tried, and it's an experiment that failed. So you call it "cloud computing" instead of "mainframe computing", but it's the same crap in the end.
Here there are a wide variety of white box PC vendors that beat Dell on quality and price for desktop PCs, and the only reason people used to buy Dell is that Dell used to offer support. Now that they offer less support than a dodgy dealer selling from the back of a truck there is no reason to buy their overpriced gear.
Assembling your own would work on a medium scale with the large number of decent single board solutions that even give dual monitor support (1/2 hour each for hardware setup) but you'd have to work out the scale where it is viable yourself. Software installs typically take longer anyway.
Personally I assemble desktops and buy servers and nodes, simply because I don't have many desktops. Where I am there is also a large margin and time cost on assembled machines, so it's cheaper and faster to put a single desktop machine together from parts than waiting for one to be put together by a vendor.
Try going to a local computer supplier, or a good local computer shop, and order a few 100 as a trial. It'll at least cut your dell costs, and at worst you'll be stuck maintaining a few 100 instead of a 1000. And any computer store worth their salt, would offer you a good rate on a few 100 - especially if your willing to do it 20-50 computers at a time. Which might make sense, if your upgrading a team at a time.
Check out daktech.com. Based in Fargo ND, this place has a 7 year warranty on their computers. $640 - http://daktech.com/build/28
Have you considered off-lease machines? At the school where I'm admin we get Core 2 Duos with 2GB RAM and a 1 yr warranty for $280 each.
This kind of thing is hit or miss; sometimes you end up with a support nightmare, and other times you end up with something that Just Works. Unless you're buying the exact hardware and bench testing the hell out of it before you before you do a roll out, you're asking for problems.
A couple points/questions:
* Are you planning to replace all 1k systems every 18 months, or are you planning on replacing 1/3rd every 18 months? The later is sane, if you've got support man power. The former is crazy; you'd be better off rolling out new systems every 4 years with an upgrade to RAM after 2. (There are different ways to dice this, but consider: you have a lot more time overhead due to the need to design the systems yourself.)
* Standard hardware installs are easier to manage with only a single system to base it on. Three different images/deployments is about as much as I'd want to mess with, though I suppose you could make it easier with RIS, to some degree (never messed with it, always used unattended, which isn't exactly finessed).
* Intel sockets are changing too much right now for me to feel comfortable buying them for "upgrade CPU later" purposes. Honestly, I'd not even bother upgrading the CPU. Getting an AMD board with an ATI or nVidia onboard graphics controller (preferably the later) seems a better bet regardless (reduced cost, better generic desktop performance than an all-Intel solution).
* Do not skimp on your PSUs. Get good ones, and you will see significantly fewer hardware failures in general. (Keep in mind that most OEM stuff is built for the limited-lifespan corporate deployment or the lifespan of batteries in mind.)
* Assuming proper testing, I'd keep around 5 PSUs, 5 "memory units", 8 hard drives, and 2-3 motherboard/CPU combos in stock for every 100 systems. (Keep in mind that up-front cost estimation is difficult, but necessary - money managers like things such as warranties, which you will not have aside from on individual parts. This will be another cost due to management overhead.)
* boards with Realtek ethernet are the way to go, due to the quality/availability of their drivers.
* I'd avoid ATI due to the headache of their drivers.
If you spec your parts right (ie not skimping where you shouldn't for quality), your workstation builds should last you 5+ years with only minor replacements or upgrades. You might get in at around $700/machine in parts for a fairly decent system. Keep in mind that your time is costly, too , so $100 for an hour of your work for assembly is not unreasonable.
In my mind, the break-even for time/money/effort in a large environment for each roll out is probably right around 1k PCs. With 300, you're getting a fairly substantial sunk cost with each cycle - and 18 months replacement is somewhat insane, since the hardware is almost invariably still able to run the OS and common applications without a hitch. (I say this while typing on an 18-month-old system which is, in my mind, still 'new' - and it does a lot more than just 'common' operations. Though I'm about ready for a RAM upgrade.)
Anyway, I think it's fairly moot. Your premise is all sorts of fucked up: you can get decent version-locked models for $500-600, and quite suitable ones for $300-500. It's difficult to build systems at those prices due to needing CALs. $1000/machine is crazy, even if you're getting them with OEM CALs and then layering site CALs on top.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
I agree with several posts here . "homemade" computers tend to be the best performing and over all best investment if you know what you are doing when you design the system. But all the overhead costs in such a large environment is pretty prohibitive. You have to remember that "end users " are using these machines. They will not treat it as nicely as you would. You may be thinking "I build a better computer than those darn cheap Dells and I never have problems with the ones I build." That is probably true, but you will probably have the same amount of problems in the end with you home built ones since the end users will abuse them. And why does a cubical worker even need a higher end machine anyway? All they do mostly is crunch numbers in Excel and shoot off emails. The cheapest computer out there with enough RAM should last you 18 month turn around. Only reason I would say build your own is if you have some high end users doing rendering or something that needs some extra power. Then you may have prob 10 -20 "special" machines to build and maintain. The others can just be the vanilla Dells. And one of the other posters was probably correct in that you probably got "sold" at $1000 per PC. $450-$550 sounds more like it.
I feel your pain, as I face a similar situation
I've been building PC's for over 20 years, and we all know how finicky building machines by hand can be. That said, the morass and mess of Government Regulations make it difficult to procure cheap, effective desktops for government workers. Anyone who thinks you can go out and buy a $400 Desktop and slap it on a government desk is completely ignorant of the facts and rules.
Quite simply: Any computer that the Government buys HAS to be 508 compliant. It's an E&IT device, and the purchaser HAS to prove that the device has met the given accessibility standards (or face the legal repercussion). And let me tell you, those $600 dollar machines from Dell and HP don't come with VPAT's (Voluntary Product Accessibility Templates). For those you WILL pay $850 or more, even considering the government bulk discounts.
Of course there are a few good WhiteBox vendors out there who are on the GSA schedule. However, try to find one with VPAT's, and you'll be looking a frigg'n long time. I've been looking for months, and have yet to find any. One I liked, AvaDirect, has great white box products, customizable and is on the GSA schedule. Unfortunately no VPAT's.
The simple economics of the issues are that you can buy decent components, from reputable vendors, and assemble generic white box machines for a heck of a lot less than you can get from HP/Dell whomever.
Components ARE exempted from 508 compliancy rules, and we all know you can build a decent i3 530 system, 4G of Ram, and a $320 gig drive for well under $500 a pop. Even assume a 20% failure rate, and $100 an hour for labor, and you are STILL hundreds of dollars ahead of the robbery the big vendors are charging. OS? I'm more than willing to bet your agency has already payed Microsoft for an OS licensing, so why pay for something you already have? Warranty? Not only do you have the ones on the components, but you buy spares. Even assuming 20% failure rate, you can easily stock up on spares (i.e. Corporate Stable Models), and save. Storage? Sure it's an issue, but so is spending the $$ to pay for shipping back and forth from a mega vendor. Drivers and stuff? Look, you'll have to tweak those machines to meet FDCC (or USGCB as the case may be) guidelines anyway, so adding a few more drivers to a WDS deployment image isn't any big deal.
Listen, I know (and honestly believe) that building machines for such an endeavor, is a terrible way to go. However as terrible as that option is, the alternatives are even worse. As a government employee I would not feel comfortable knowing I'd thrown hundreds of dollars away for each machine we purchased from one of the big mega vendors. Our custom whiteboxes would be better constructed, more efficient and overall a much better deal for the employees, the IT staff, not to mention the taxpayers.
It's your choice, but we're going the component way, and I hope the taxpayers appreciate the efforts.
for a mediocre computer, your idea of what qualifies as "mediocre" is different from mine. For typical office tasks you should be able to get the job done for about half that. Sure, your power users (CAD, software developers who need to run multiple VMs, etc.) may need a much more expensive workstation ($2000 or more), but your average office PC user does not need a $1000 system.
I also question whether building your own is going to save you anything over the long run once you factor in the costs of building and supporting those systems yourself, and the fact that you're unlikely to get the parts as cheaply as the big OEMs do.
Someone in your company, somewhere, probably someone with 'Strategic Sourcing' somewhere in their title either has a free MP3 player, tablet, Mini, or sports tickets, courtesy of Dell. If you aren't that guy, you've already lost.
I want to delete my account but Slashdot doesn't allow it.
If I read this right, my suggestion would be to look at frye's
http://www.frys.com/template/computerspc
From then (ownership on) you can use the old dell hardware for spare parts or add them into the new pc (memory, CD /DVD/ NIC / VID CARD, POWER SUPPLY, etc. when it can be used.
I had the idea of building our own PCs for considerably less.
Dell in its prime was absorbing the entire annual output of its Asian OEMs. When there was a dock strike in L.A. it hired fleets of air cargo planes to maintain just-in-time production lines.
Parts are cheap when you purchase them in the millions.
If you assemble and maintain your PCs in-house, you will have to pay US wages and benefits. You will need to maintain parts in inventory. You will need to hire someone to keep your home-brewed systems in repair. All of this costs money.
Listen to the advice in this thread. You cannot do better building the PCs yourself unless you have a pretty massive support infrastructure, space, time, and staff.
I work in the higher education field, and we've been building computers for the last five or six years. We spend about $600 per computer (that was the last build), for a decent, just over the tip of the average computer. We add in about a 10% margin, and keep those parts for repairs.
I suggest that you look at what Dell, HP, Gateway, and the other pre-built PC companies are selling, aim for the medium range, and spec-out what it would cost to build the exact same thing yourself. Remember to add in the cost of time/labor, otherwise you'll throw off your figures when you present to the boss.
Dell IS a generic PC.
Honestly, though I'm pretty steadfast about building my own for use at home, I'd never do it in a corporate environment. The support and warranty will be a nightmare tracking down manufacturers for each component when they fail (because with a whole fleet, many WILL fail).
If you want to save money, my advice would be to go through the regular small business store, and buy the budget systems Dell offers for around $300-400. They'll still have a single source for warranty service, and in the event that you need to CYA you can still honestly state that you bought Dells.
Just personally, as another person in small/local government IT, I'd say you can probably save a good bit in transitioning to OSS on the server side. We for example pay nearly $25,000 per year for our Lotus Domino support contract. Now, it's got a lot of features that would be hard to duplicate, but from the standpoint of having a completely functional email server, I could setup a Postfix/Dovecot system that would run at a fraction of that cost. We even ran that as a LONG time as a backup email setup for IT for when the Lotus system was down. If your website isn't using ASP, transition from IIS to Apache. Look for things like that. I won't currently recommend hitting that up on the desktop yet, but on the server side, people care much less. You could possibly save quite a bit without the users noticing that much of anything is up.
"People who think they know everything are very annoying to those of us who do."-Mark Twain
Lots of heartache and pain.
You're looking at short term savings to inflate long term expenses.
Sure, it seems like it'll save you money in the beginning, but then components start to fail, and you can't find exact replacements. Then components that you bought to work with the old system don't work with the new motherboard/videocard/something. Then you think "we'll have two disk images". Then, a few months down the road, something else becomes hard to buy. Three disk images.... and so on. And you have to remember the quirks of each system as you set work with them.
Assuming you don't buy 1000 groups of parts immediately. (You won't, no accountant out there will approve it. Basic business needs say that's a bad management of cash flow.), You'll be spreading that purchase over 6 months or a year. I don't know if you've noticed how often basic components get refreshed, but by the end of the year, you won't have 1000 identical PCs.
I know calling support for Dell is a pain in the butt, but try calling tech support for Asus. In Taiwan. During business hours EDT. Plus, you can't expect them to keep spares for the time you need them, where most OEMs keep spares on the shelf for 3 years. Do you have the warehouse space for that? What if there's another run of bad capacitors, and all components manufactured during a time period are bad.
And... then there's your time. At my peak, I could assemble and build a PC in about 3 hours. Multiply that by 1000. That's 3000 hours. Non-stop building. There are only 2080 working hours in a year. When are you going to have time to do your actual job, that of system administration? Yes, you can hire someone, but so you want to hire a person who assembles PCs, or do you want to another system admin, which will actually make your job easier?
We're not even talking about the government's needs to track where money is spent. How are you going to stick asset tags on a random bunch of assembled components? What happens when most of the guts of a PC get put into a case where it already has an existing asset tag?
Man... I've beendown this road. We got about 18 months down it and we went back to the OEMs. Dell, HP, etc.
If you're going down to the local PC store, or buying direct from Dell's retail side, STOP. Look into Dell's corporate line of PCs and the HP's corporate line. I just checked HP's government purchasing site, and you can get a small form factor PC for about $350. I'm sure RAM isn't that expensive, and the three year service contract in bulk won't be that bad. If you're in the state of Virginia (and if you're working for the government, there's a high likelihood that you are), consult an eVA price list. Or go off the VITA contract. The amount of PC you can get for very little money at government pricing is somewhat ridiculous.
Just, please, no, don't go down the path you're going. If you really really really want to, do it in one department only, for 18 months. See if it's worth the hassle. I'm betting it's not.
These thoughts are all just dashed off. I'm willing to go into more specific detail if needed. I just remember how happy we were to get out last HP Vectra machine in to replace the custom built pieces of crap we had before. It was nice in so many ways. I could actually go back to administration and not construction.
Reeses
You are not accounting for IT administration costs. Loss of ability to use Dell (or other corporate PC manufacturer's) central management and support tools.
Difficulty (or impossibility) of finding reliable replacement hardware.
Lack of professional QA in regards to workstation configurations.
Proliferation of new workstation configurations, leading to inability to have a 'standard image' supporting hardware in all workstations, when every time you need a replacement, you are buying the cheapest generic parts that just happen to be available at the time replacement is needed.
The moment you try to start standardizing hardware in these "generic" workstations, you will quickly find that availability of exactly the same part is non-existent, as the generic part you used 3 months ago is now discontinued, etc, etc.
At least by using mass-produced workstation models, you are assured of being able to continue to buy identical or near-identical hardware in the future to minimize new compatibility or stability issues.
I sense a volume demand here. Either go to another supplier and negotiate an OEM volume deal, or do that and use the numbers to get better prices from Dell. If you build your own PC it's fun to do, because you do not have to worry about warranty, holding spares, assembly lines, testing, oh, and paying people to do all that.
In short, I recommend against it. Life's too short..
Insert
The GAO should be looking into this.
RIP America
July 4, 1776 - September 11, 2001
I worked for a DOD Contractor that did this for about 500+ users for a few years. What you save in upfront per-machine cost, you will pay back with extra in manpower and frustrated users. Keep the Dell Contract, negotiate for discounts based on your size and the fact that you "have a workable alternative"
Think of this, 1000+ Machines will need in-house IT to manage. You WILL have at least 5 on the bench at any given time. Just accept that. You will have to employ someone at each site to manage those machines, staff who will be idle at some of that time.
With vendor PCs and Onsite contracts, you will STILL have 5 on the Bench. But it's not YOUR Bench and the user community will join you in commiserating about the vendor rather than complain about you, at the same time, you will NOT be paying for staff to manage all those machines, you will simply call the vendor and have them come to the site as per contract.
No muss no fuss.
As a tech, 1000+ hand built machines with full control was absolutely awesome!
From a management standpoint, it doesn't make sense.
Growing up sucks
Google, Amazon do it - I guess any government could too. Make it pay off - order extra parts with your savings and you
are golden.
What could you possibly be running that needs upgrades that often?
Duh. He's running Mac OS X.
*ducks*
Boot Windows, Linux, and ESX over the network for free.
Until a member on the board of directors, a vendor, a lawyer, or anyone else on the outside world emails said manager a Microsoft Word document, your argument fails hard. But what are the odds of any of the former situations happening?
They're still stealing the 'W' off the keyboard. After about 18 months, they figure there are some 'W's inside the grey box, so the one who's bright enough to use a screwdriver opens the box while it's running and drags the screwdriver over the motherboard, frying the machine.
Contribute to civilization: ari.aynrand.org/donate
Either work with other vendors to get competitive quotes in order to reduce costs or look at desktop virtualization. The latter can be done with much less than the 1M you are currently spending. My bet is you can cut that easily under half. As others have mentioned use competition to your advantage. Getting competitive quotes from HP and IBM will have a nice effect on getting Dell to lower their quotes. I too think the 1000 per unit cost you are paying is way too much.
Building these yourself won't reduce cost since now the support team is also going to be building these machines. In supporting them you'll be building them for lowest cost which means the individual components will creep. End result will be tons of permutations you'll be asked to support. This is a recipe for disaster. In scale you want same hamburger everywhere and the cheapest you can get it that still meets the objectives.
It is a legitimate expense to buy a computer; it's tax-deductible on corporate level
Disclosure: I'm a certified accountant. It is not true that buying a computer is tax deductible. A computer is normally a capital expense. It is purchased and then depreciated over the useful life of the asset to emulate "using up" the asset over time. While this does reduce profits to the corporation and thus normally reduces their tax bill, saying that a computer is tax deductible is not true for businesses of any size under normal circumstances.
Aside from that nit, I agree. Building your own machines on any significant scale is a most likely a very bad idea. High chance of serious problems and it's only cheaper if you don't factor in the time required by the staff to assemble the machines. Unless you work for HP, thinking that you can build 1000 computers cheaper than Dell is delusional. As an accountant, I'm quite sure it would be cheaper, even at $1000 a machine to buy the computers from Dell or another similar vendor than to have my staff build the machines in house.
I'm buying decent spec'd Dell Optiplex 380 mini-desktops with 2 19" wide screens and the optional nVidia card to support them for under $1k in quantities of 25. Having come from a white box background then moved to HP at one organization and Dell at another I'll NEVER go back to white box....
And have they got a supplier for you that meets your criteria?
This is not a corporate purchase, it's a government purchase. Different rules. You can buy PCs from approved vendors, and they will be happy to give you a list. You will be buying the OS from a GSA vendor, Microsoft, probably, and the terms and conditions are not yours to decide. There are very, very different terms for GSA purchaes, and this can be done but not as easily as just buying from the conventional vendors like HP/Compaq and Dell, for instance. There are others.
I've tried to do business with agencies for quantities from 30 to 1000, and never had much luck. Even when I had an opportunity to broker a deal for >1000 systems, the vendor kicked the door in and took the deal over. Learning fast, I didn't bother.
You probably should just give in, but I admire your intentions.
deleting the extra space after periods so i can stay relevant, yeah.
Our product is, technically, little more than a computer in a fancy box. We started off wanting to buy a small computer and shove it in, but found that we needed way more power than anything available at the correct form-factor.
Being computer guys, we figured we'd just build it ourselves.
Truth is, around here (Toronto), OEM computer suppliers are everywhere. Good ones (Infonec) with reasonable inventory and reasonable access, Poor ones (Tiger direct) with huge inventory and no access, and remote ones with infinite access and no inventory. So we're covered from every angle be it some rare component or an immediate same-day requirement.
Do components break? Sure. Some hardware is defective out of the box. That goes onto the reject shelf. Some break when we drop it. That goes into the garbage. Some break after they are installed when it's just not stable and it takes many hours to figure which part is at fault. Those are annoying, but they go onto the reject shelf just the same.
The reject shelf gets turned over by mail with a few RMA phone calls every few months. The nice part is that if you wait long enough, you tend to get newer models from the manufacturer, so it's winds up almost being worth-while.
The garbage is, honestly, an easy thing to avoid. Wear cotton, ground yourself, and never put a motherboard onto a chair unless you atcually want someone to sit on it.
The nice thing about 1'000 is that while you can't get much of a discount on the components themselves, you do get more than priority service from the suppliers. And that can really be valuable when it means that your deployment schedule is uninterupted.
Yes you can save money. You should wind up saving about 40% over a dell machine. Of course, you'll lose the warranty service. And that's where the trick comes in. You get to balance something that you've never balanced before.
You get to say: "cheaper = more servicing = more expensive" while also saying "higher quality = less servicing = still expensive"
Here's the trick: "higher quality = longer life-span = re-use"
The real savings aren't on those 40%, because you have to service them instead of dell servicing them. dell's more efficient (money wise) than you are. But because of that, dell's cost-optimizing the quality, because they don't get to keep it. They'd rather take the risk that the parts won't break, and fix the 20% that do.
That doesn't work for you.
You want to spend more, only saving 20%, then you want to do minor upgrades at the right now, so really only wind up saving 10%, then you want the machines to last twice as long, and be able to salvage the parts for future machines -- repeatedly. This also has service replacements of broken parts and diagnostic repair fed for free.
In the end, you wind up spending the same 100% out of the gate, you spend only 80% the second generation, and then you spend closer to 40% by the third generation.
In the end, you have high-quality machines, top-quality parts, and very few break. Service calls are not only at a minimum, but you're just swapping out the possibly bad parts with known-good parts, then checking the possible bad parts at a later, more convenient date.
You're also providing the new guy with a better computer to get him started on the right foot, you're giving the guy with a lot of work to do this week that extra gig of ram to make it easier.
But yes, this presumes that you are comfortable running such a service. It's definitely easy to do, but it's complicated as hell to keep it organized.
If you can justify it, go for mini-itx boards with SSDs. You'll pay about $300-400 per machine.
Advantages:
- Very much reduced power consumption. Electricity isn't cheap anymore - consider $1/week for a desktop.
- Silent. (You really appreciate this when you get used to it). No moving parts => less to break.
- SSDs are really really fast. Even a value 64GB SSD, coupled with 4 GB RAM will make the machine fly, and
the D525 CPU is not at all bad (it's effectively a quad-core)
- Very easy to build
- Mount the PC on the back of the monitor for compactness. A VGA mount should cost only $10 (I had some made out
of aluminium angle for $5)
- Use the saving to buy 24" LCDs for everyone - this is a major productivity win.
If you really can't live with an Atom CPU, (though the D525 is a 64-bit quad-core!), you can get mini-itx boxes that take Intel Core2 (or similar) CPUs, though these do have fans.
Regarding software, turn off all the flashy desktop animations and the machine will appear much snappier. Also, consider whether Etherboot/Linux has anything to offer you - even if you just use a customised Ubuntu install usb-key in order to get the machine automatically set-up with your Windows environment. (Or perhaps every night, send a wake-on-lan packet, etherboot, and rsync the image across to update).
Lastly, go for one of the Mini-itx boards with its own internal 12V power-supply on the motherboard (uses a laptop-style brick) - this eliminates another step from assembly. You should be able to get the Mitx systems fully assembled in half-an-hour each, once you've built the first 10 (though you can get suppliers who will build them for you).
Make sure the employees like them: buy good keyboards and mice, + the gel wrist-rests. (Personally, I really like IBM ultranav keyboards)
We're buying Dell Optiplex 745 towers coming off corporate leases from a refurb reseller. It should be easy to find XP Pro, 4GB, 80+ GB, Core 2 Duo with 3 year warranties for under $350 shipped. Less if you can buy in quantity. Grab decent Samsung monitors from Costco locally for easy returns in case of bad pixels.
It's just not worth it to roll your own when you can buy a workstation that solid for that price, particularly when you don't have to be responsible for the warranty.
You purchasing people are idiots. My brothers small biz routinely orders well appointed Dell machines (mid-tier Core processors, 4-8gb RAM, nice monitors) for around $700. That's for one or two at a time. It sounds like your folks negotiated some "standard" buys which are crappy deals. This is not Dells problem, and it's unlikely white-boxing will fix a broken procurement system.
What a joke. A Mac costs less than that and is a far better computer with a much longer life cycle and lower cost of ownership.
Seriously, thinking you can build your own computers at a price competitive with Dell (I Thought Dell *WAS* Generic, btw...) is a bit silly. If you come out ahead, you are probably missing something. Sure they are going to lop on a 5% profit margin or something, but they also have a scale of .. millions of machines. Their price includes the OS license (if any) burn-in testing, quality control, warranty, assembly, etc. If you are going to in-source production of PCs, you have to include all of that stuff. If you can really do it cheaper, than you should open a new business and compete with Dell.
Even with Apple:s stuff, it's pretty hard to build or buy a *similar spec* laptop for the same price. With Dell, you may as well forget it.
Also, when you buy from NewEgg or whatever, what's going to be cheap (or even available) is going to be different each week, but what you want is the opposite - the same thing every time.
Building PCs yourself is like building your own power plant. It's probably just not economical on a personal scale.
Generally I would agree that building isn't - these days - more cost effective than buying. However support is kind of an interesting facet and not nearly as straightforward as one thinks. I have a Sony laptop with a hugely expensive warranty. Sony's on-site support requires diagnosis over-the-phone, (about an hour all-told) a technician is shipped the part and dispatched (1-4 weeks from the date of the call and at least an hour of downtime for the actual repair). The technician has no authorization to request parts. So if they bring the wrong part then you have to call again. As it stands my laptop currently has serious problems - only two-thirds of the screen is visible - but this support system is so costly to my work day that it's easier to simply use an external monitor and wait until the unit gets replaced by IT at which point it can be sent to a Sony depot. While this is an extreme case, I would contend that most support contracts inflict a huge cost on my workday. Dell, for example would ship me a part which I could install myself however when something serious went wrong (motherboard, screen) I found myself borrowing laptops from the loaner pool. The only thing that made this low-cost was that they had some units which were almost the same model. Effectively I could pull my drive, insert it into the loaner, reboot twice (first time Windows would install drivers) and I'd be working again.
The point? The thing that made support low-cost to the worker was having similar spare machines commensurate with the number of machines likely to be failing at one time. Having an on-site, expedited warranty system was clearly and always secondary to having a cold spare So outside of having someone to blame (which sadly is more important than it should be) it's conceivable that, if you plan correctly you could end up with significantly better support than your average vendor could supply.
Interesting fact, as you move away from the desktop and into the server room - I'd argue that commodity hardware is even more cost effective. Having a cold spare on a 1U machine is at least conceivable and assuming you plan well would allow you faster replacements than even the most expensive warranty service. Not to mention that the number of hours a server is down is far more likely to inconvenience or enrage more people than a single desktop/laptop. As you move away from commodity hardware and into proprietary systems like blades - it becomes difficult to cost prohibitive to have a cold spare for everything. This also goes for hardware that is longer term. One time I noticed that I could purchase four of the controllers for our wireless networks on Ebay for the price of continuing one years support on the two we had in production. Last time one failed it took two days for the manufacturer to get one to us.
Sure, :D
I have your problem solved right here:
http://www.citrix.com/English/ps2/products/product.asp?contentID=2300325&ntref=prod_top
Basically, it's a bare metal hypervisor for your systems. Fire up an image and you're good to go. Upgrade to new hardware? No problem, you're virtualized. Remote users with laptops? No problem, remote reimage works too.
Yay?
Little more complicated than you make it out to be. To be truly redundant that a switch going down won't take out a bunch of systems you have to have the systems themselves plugged in to two switches, and then every switch down the line. That can work but not only takes a lot more switches, but more complex clients. Most thin clients aren't going to do that. You need not only 2 NICs but the understanding of how to handle failover. Also if the failover is to be fast and reliable you need expensive switches. Maybe not a problem, maybe you use those anyhow, but something that has to be considered. At every level the switches need to be high end such as Cisco to be able to quickly, reliably, handle rebuilding the span. No Linksys stuff that may freak out and create a switching loop (which they do with RSTP sometimes, as I've seen).
Bandwidth needs will also go up substantially. If you go a little heavy on the oversubscription in a normal office setup, no big deal all it means is file transfers to the servers are slow. Do it in a thin client environment, and you are talking interface lag which is really bothersome. So you'll need to have plenty of bandwidth to the switches, probably 10gb instead of gig, and maybe more to the distribution switches.
Then you also have to do redundant power for the switches. If both switches on a floor are on the same breaker it doesn't help much, you need separate circuits, all the way out to the grid/generator if you want real independence.
Of course there's the servers also. If one server runs 50 machines, well then its failure is a major outage. So you'll need backup servers. How many depends on how much depth you think you need, but you need to have servers ready to take over if one goes down. Probably fairly beefy servers too. While you can stack low-impact servers (like DNS or license servers) pretty heavy on a VM, you have to be more careful with interactive systems. Get too many, they'll get sluggish. You'll want lots of CPU, lots of RAM, and still won't want to load clients on them too heavy. You'd have to test your specific setup to find what works but I'd bet no more than 2 clients per server core and probably less.
That also means everything has to be on a separate, high speed, disk system. You can't use local storage or they can't be migrated to new servers. So something like a NetApp. Disks need to be high performance too, since they are going to have a lot of random access put on them. IO is also the biggest problem for multiple VMs. No large cheap SATA arrays, you'll need 15k SAS most likely and SSDs would be a good idea, except real expensive.
Well that needs to be backed up too. If everything is riding on one NetApp, reliable though it is, that's a massive failure point. So you need two of them, running in sync, so that if there's a failure there's no problem.
Ok this is all doable, no question about it. I could design an implement such a system... However I'd have real questions as to if it would save any money. You weigh all that high end gear with service contracts against the cost of a bunch of reasonable desktops. Is it really worth it? My guess is not.
Also remember you aren't saving any money on other server costs. You still need all your other server infrastructure. Maybe you could get rid of your central storage and just use the storage the VMs are on, but I would have to see that in action to be convinced the performance would be ok.
The thin client idea isn't a money saver I don't think, unless low performance/reliability is ok. Maybe a school lab situation or the like. I think it is more the sort of thing you'd do when you need portability (like no matter where someone physically is in the building, they can get to "their" computer) or for security (for whatever reason you want all systems physically in a secure room).
Get other vendors involved. With hp, when was working as a reseller,we would go after volume big deal pricing, and drive the costs way down.if you have not been buying as a big purchase at one time, you don't get on the vendors radar either. Dell isn't necessarily the cheapest vendor if they don't see competitors at the table with them....
I work for a local school district, and we did just this. Laid out our specs, and sent it out vendors. Our current hardware is:
Slim desktop case (has 3.5" and 5.24" external bays, 3.5" internal bay)
Asus motherboard (I forget the model number, but lots of USB and SATA)
AMD Sempron 2.0 GHz CPU (64-bit capable)
1 GB of RAM (supports 4 GB)
onboard nVidia graphics
onboard HDA audio
Without a harddrive, floppy, optical drive, or OS, they're $200 CDN in groups of 30.
What are you buying that's $1000 (US I'm guessing)?? That's insane for a business desktop PC.
For instance, you could try just doing this with 5 computers over a period of 6-12 months and see what the costs are relative to the Dell computers. Immediately trying to do this with 1000+ computers is a terrible, terrible idea.
It really depends on what kind of volume you expect to come through your shop. I work for a university where we have an on-campus shop that builds our desktops and it definitely has its ups and downs. One significant problem we have is in build quality we see during summer when our orders ramp up. That's when lab managers across campus plan lab replacements and the problem is that if we don't put our orders in starting in March/April they just can't build fast enough to deliver the machines in time for us to have them installed and setup for fall semester (secondary problem is that they delivered 70 machines to me in June, when I wouldn't be able to put them in place until mid-August, I had to find a place to put 70 full desktops with monitors, mice, etc). The reason the build quality slips is that the shop manager hires extra help during the summer to try and cope with the additional demand. Warranty-wise we're typically okay and we've got about 1,500 of our on-campus built machines deployed.
The other issue we have is that we often can't sit on hardware for too long. Hard drives, processors, and RAM aren't typically a big deal but we have real issues with staying on the same motherboard for more than about 6 months. Hardware gets revved, or something else stupid happens and we can't get the board we've standardized on, then it takes us about 6 weeks if we fast track our testing. Motherboard changes sound like they're not a big deal, but we've gotten boards in that we couldn't image with our imaging software, or other strange issues that are specific to the model (as opposed to a one-off bad motherboard). Also don't forget that you're going to have to start handling your own RMAs versus having Dell do it for you.
So, it's generally worked fairly well, but some of the lab managers (myself included) have sworn off any more mass orders from our on-campus shop. The main reason being that they can't keep up with the demand when we switch out our labs. We're looking at keeping 1 to 10 machine orders on campus, but anything larger than that we might divert to Dell as they have a lower failure rate on the few large orders we have done with them, and their next business day service is straight up better than anything our on-campus shop can hope to match - that's a logistics and manpower issue for them, they just can't respond as quickly to repair requests as our local Dell depot can.
This space for rent...
o Pick Intel chipset (You benefit from larger customer base) Research for potential bugs in SATA, GPU and NIC. Get one that is compatible with Linux as even Windows drivers are more mature then.
o Build system with low thermal footprint. Heat kills components. Don't try to stuff too much in a small case...or use a lot of fans...
o Do not skimp on components. Try to save money by buying cheap stuff and you'll fail.
o It's best to purchase PSU separately from the case. PSU is usually the first to fail (maybe right after fans). It also makes more noise than CPU fan so get a good one.
o Oh, btw, it's a bad idea to get a motherboard with PSU build in...I'm sure you can figure why.
o If you don't need a lot of storage use intel SSD for storage - they save money in a long run (faster virus scans, etc) and make users happy (try that with Dell and you pay premium)
o Don't put a lot of RAM in if not necessary but do qualify that the mobo can support the max. Sometimes 1GB is enough (HD is the bottleneck). You can always upgrade later.
o Don't install CDROM drives if users don't use them on daily basis. Buy a few external ones.
It's best if you can do it in batches- if your offices/users/accounting support it: move fast but old computers down the chain while speed hungry users get the new batch with better components. Computers with good components can last 5 years and longer.
If you don't standardize on all parts for every machine, you'll start having a support nightmare when it comes to imaging your machines.
So if you buy 1000 of each item to build 1000 machines, you need to account for having a bunch of spares of each item (motherboards, videocards for the most part), even under warranty, because you have to send in bad parts ot the manufacturer, wait for them to determine the issue, and them to fix/replace and send back. And, if the part you sent them isn't current, they may replace with a different, but comparable part.
And if you're paying over $1000 per computer, then your agency is doing something wrong. My mid-sized company buys dell desktops, including acrobat and office, at around $850/ea, including monitor. And they are decent office-type machines.
Most of the posts so far are mentioning the obvious reasons why "not" to do build your own. However, I have another couple of reasons, which is more powerful yet.
Customized system (BIY) often include "Hidden" costs of emotion. They are "MY COMPUTERS" ... and people tend to try to hold onto them much longer than they ought to, just for that one reason alone. They spend more time trying to revive dead computers, fixing hardware that should be tossed or parted out and so on.
One of the things I've also noticed lacking is the how easy most "corporate" computers are to "fix". Most Dells, HPs and so on all have quick/easy access to things like RAM, HD, and slots. Often not needing ANY tools whatsoever to gain entry, and repair. Just swap and go.
The cost of tech support is twice the wage, and if you take 30 minutes opening a case and swapping a hard drive vs 10 mins with a Dell or HP, you're saving money every time it needs fixing. It isn't worth whatever savings up front you might see.
If you really have 1000 computers you need to install/manage, you're better off building out a stable support platform to provide first line support than spending the time building them in the first place, one that plans upgrades and replacements as part of the process.
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
I work for IT that supports all of the County Goverment's various departments and agencies, about 500 computers. We build most of the computers. It has worked out as big cost savor. Fortunately we setup our volume licence to be transferable because we have used Windows XP and won't transition away from it a department at a time for another year or two. Probally won't compleatly stop using XP until 2014 or 2015. 12+ years for one OS purchase isn't bad. We don't save much up front, but in the long term to cost are great. Many of the parts have 3 to 5 year warranties. The companies we buy parts have a easy RMA process. We do have a 100 Dells, and I hate jumping through their support idiot's hoops when I have just to replace a dead part. Dells non-stadard parts make it expensive to repair when out of warranty. More than half of the computers I support are 5 years old or over. There are still a few 8 year old computers! I had to put a 3 year old Dell out of service, because it would cost $200 to just buy the motherboard on a computer half way though it's life cycle. If it was a custom computer, it would cost $50. I wish we would go all custom computers, but we still buy prebuilt computers for Library Public access computers, for a few reasons. We couldn't build 20+ computers very timley with our small staff already loaded with other tasks, it is harder to put custom built computers into the State and Bill and Melinda Grants paperwork, and the smaller size works better on the Public access desks (Though I guess we could build mini-ITX machines)
It's the fleet maintenance that does you in. You pay for that as well as for things like keeping the fleet constant -- paying extra for downing the specs of replacements. That you're asking means that you don't know what you're really paying for.
What you could do, next to the usual negotiation and offers and stuff, is make a deal that allows for newer configurations within strict constraints. Say, you replace a quarter of your PCs each year (and keep a couple spares on-site) and each year sees a new model. That cuts costs both ways because you now need four sets of driver configurations (and while not one, that's not a gazillion either), allows you the latest with only a year delay at most, and doesn't require your supplier to keep four year old parts in stock. That's the sort fo thinking you should've thought of yourself instead of asking slashdot. What are you going to do with the oldest trance of PCs when they break? Replace parts? Of course not.
But you have to realise that on the level of "a thousand PCs" you oughtn't care about configurations and fleet maintenance. If you do you're thinking too small and not managing your job, which is to deliver a thousand working PCs, not tinker with building amd maintaining them. Your leeway isn't in the hardware; it's in setting the right constraints for your suppliers.
Actually, capitalism partly rests on "division of labour". Dell can build PCs and do it cheaply - and handle the issues of compatibility testing, etc that backyard/amateur operators (that's you) cannot do. You can specialise in your work that Dell cannot do. That is the most efficient arrangement.
You trying to be Dell is just setting yourself up to fail at their job and at your own.
Do you also propose to build your own cars? Construct your own offices? Mill your own paper?
I am anarch of all I survey.
Sure, you might know how to put components together, but do you have any idea how much work goes in to staying abreast of new products so you can be sure the stuff you're using is faulty?
What you want is a generic-winterm-like solution without the single point of failure.
1000 "generic" computers that store all user data and settings both locally and on a server will do nicely. If that's too much then at least have the user profiles and other directories that contain user data sync up with the server nightly. I say "all" - if you allow removable media you don't have to mirror that to a server. If you allow users to store personal files on company computers as a courtesty then be sure to tell them which directories will and will not be backed up. Company policy or no, unless there's a strongly enforced edict against it, some employees will store pictures, videos, and music on their workstations. You don't want that chewing up your backup servers.
Configure the workstations so if the server that stores the profile isn't available or a switch dies the user can still log in and access network resources he can reach.
Configure them so if his computer dies you can just "drop in" a replacement and when he logs in and re-synchronizes with the server he's good to go, even if the new workstation is a different brand than the old one. The only noticeable differences should be any applications or plug-ins the user installed - assuming you allow user-installed applications - would be missing from the replacement computer. That's okay, if he was smart enough to download his favorite web browser before, he's smart enough to do it again.
By the way, it's not required that all 1000 computers be the same make or taken from a short list of models, but it helps a lot down the road. I'd much rather support 1000 customers where all but a few of them share the same half a dozen models of computers and they are all the same make than 1000 who have a hodge-podge of different makes and models. "Half a dozen models" is realistic if the oldest one will be 4-5 years old when it is replaced.
I also don't want to be supporting more than 2 operating systems at a time - e.g. the current OS and whatever is new that we are rolling out, or the current OS and the few stragglers who haven't upgraded yet. I also prefer not to support more than 2 Windows service packs at a time and I don't want to support a service pack that Microsoft doesn't support. The same "no more than 2 major versions and no more than 2 major service levels of each version" rule should apply to other major software packages that are part of your "everyone gets this software" or "many people use this software" list.
In the long run applications will be supplanted by web-based solutions, either on your Intranet or run by outside providers. Your finance department may already use web-based solutions provided by your bank in its day-to-day operations, for example. For many shops, including perhaps yours, "it dream world" is where the actual data is stored and manipulated in a "glass house" somewhere, either in your back room or at a 3rd party provider, and the computers are just glorified dumb terminals. Unfortunately, this dream world can't happen until the networks are so reliable that a crash doesn't mean everyone stops working.
Knowledge is how to play a game, intelligence is how to win, wisdom is knowing what game to play.
$1000 for a mediocre computer? I work for a K-12 public school and we get very nice, corporate-level HP desktops for around half of that. We get Intel dual core PCs with 2GB RAM, a 160GB hard drive, DVD burner, and a 3 year warranty for around $430. A 19" LCD is about $120 and a keyboard & mouse is $15, but we reuse those as often as possible. A license for MS Office is $50. So, all told, for a complete PC with basic software, you're looking at around $615. That's a lot less than $1000.
Please excuse me for being blunt, but I suspect you are not really in a position of authority for this project. You sound more like someone who is in a large IT department for the first time and that you are questioning your bosses decisions. I find it un-likely that a 1000 desktop organization would hire a department or project manager who would even consider a white box solution.
Here are a few points:
First, just don’t do it, it is my tax dollars being wasted. Over 2-5 years the support and maintenance will cost you at least 20% more than manufactured business desktop, and this will not include the extra support staff you will have to higher just to manage warrantee issues. And believe it or not there is a significant increase in the time it takes just to service a large white boxes population because the cases are not built for production accessibility like manufactured desktops are. 1 screw VS. 4 times 1000 desktops just to open it up. Time IS money.
Second if you are paying $1000 a station you are either getting ripped off or you are not purchasing business desktops. You are probably buying entry level workstation class equipment which is overkill for 90% of users. You should be spending $600 - $800. Again please don’t waste my tax dollars
Third, I am an HP reseller and have noticed a lot of Dell advocates here. Dell is not a wrong choice but in my opinion is a substandard solution to HP. Most companies buy Dell because they think it is cheaper. My experience is that after you add a 3 year warrantee and upgrade the components to something actually usable which then matches an equivalent HP the price difference usually comes out to around 3% - 5% difference. Then when you consider that that HP has desktop stability programs that guarantee that every pc you purchase within the 3 year life cycle if a business desktop family will have identical components from year 1 to year 3. And that they offer driver and firmware updates long after a pc has been end of life. The long term savings far exceed the 3-5% difference. Dell generally has no interest supporting equipment after they are no longer legally obligated to do so, and cannot guarantee that you will not get 3 different component build sets in a batch of 300 PCs bought at the same time. Not good for imaging or t-shooting.
I do admit that I have a bias but I can offer this observation. I have seen several dedicated Dell shops switch to HP after testing their first HP, but I have never seen a dedicated HP shop switch to Dell desktops.
There is my 2 cents take it or leave it.
I once read an article here of a consumer-based data center building their own linux-based NAS boxes as they found commercial solutions poorly economical. If you want a bare system (w/o monitors, keyboards, etc) you can easily get the parts for under $200 on NewEgg, granted you sacrifice the optical drive (a $25 value!), switch to Ubuntu (they sell support licenses just like RHE), etc Assembling 1K computers, however, will be a cost in itself, monetary or not; just remember to factor that into your equation.
I have some experience with this, and IMHO, you never want to substitute a real PC workstation for a "thin client" running a session of some sort on a back-end server, whether you're using Terminal Server with RDP or Citrix, or whatever else. These technologies certainly have a place, but it's more of complimenting an existing infrastructure than replacing a basic part of it for "cost savings" purposes.
I love things like Citrix or Terminal Server to provide remote access to people, and even when you need to run an app that's not installed/configured for the environment in one office, but IS all set up for one in another office that has WAN connectivity between them. But most of the thin client desktop replacements don't really come out any cheaper than a basic workstation when you factor in the cost of licensing AND the limitations and headaches it creates down the road. Chiefly, there are too many issues getting printers working in those sessions. Say someone has a cheap all-in-one inkjet printer on their desk. Maybe they scan or receive occasional faxes on it that not everyone is supposed to be seeing, and they may as well be printing to it as well. Chances are good it won't function right via Terminal Server or Metaframe, where they expect all your printers/drivers are "enterprise class" units. Additionally, it removes flexibility. A lot of sysadmins claim to love this type of environment, because they retain all the control. Well, that's only ok until you've got hundreds or thousands of users who start requesting special software on their computers and you realize that now, YOU have to do all of those installs yourself and retain full responsibility for them. (Truth is, if you give people standard Windows workstations with enough rights to install things, most of them will just go download and install what they want/need and you'll be none the wiser. Sure, someone *could* infect their box with a virus or something that causes problems. But most people are still afraid of doing that sort of thing and getting in trouble for it, so it shouldn't be that common of a problem. Meanwhile, you've saved all the hassles of people asking for Firefox or other popular and harmless apps. And if/when they try out things that are defective/buggy? They'll usually just figure that out, uninstall them again, and move on. No point in getting I.T. all worked up over the "problem app".)
I know there are exceptions where you just can't allow this much "freedom" for your users (banking, military, etc.) .... but those people know their specific scenarios. I'm talking about typical corporate settings here and what works best, vs. someone's ideal "controlled environment".
Something, somewhere, is wrong...
1.) If you're ordering a THOUSAND machines from Dell, widdling that price down should be easy enough. Get a quote from Dell. Get a quote from HP. Go back and forth between them (ideally with the same sales reps), and before lunchtime, one of them will have shaved at least $200 per unit off the sticker price.
2.) Stick with an OEM for that kind of volume. What you're paying for (or at least what you SHOULD be paying for) is a 3 year, kitchen sink, no questions asked warranty. We've got them on the Dell machines at work. I had a question about a server that was two years out of warranty, and I got an English speaking support rep. I had a power supply die on an in-warranty desktop, and a new power supply was on my desk the next morning. If I wanted a Dell tech to swap the part out, all I had to do was say so.
2b.) If you build yourself, you're stuck supporting everything, and dealing with the finger pointing crap. If there's a mobo issue, you're going to be on the phone dealing with Asus trying to convince them that it's not the RAM, the processor, the GPU, or the SATA drives, and in order to make the argument, you'll be stuck swapping all those components out to prove it. If you go with an OEM, it's Dell/HP's headache to isolate the issue and replace the faulty part.
3.) like has been said before, about the only thing that can make your job a bit easier on yourself is to do some sort of disk imaging. You'd be retarded to install Windows manually unit by unit. The only problem with that plan is that you'll need identical hardware each time. Acronis' Universal Hardware Restore is supposed to deal with that on paper, but I haven't had success with it. If you can purpose one of those machines to be a server and throw Server 2008/R2 on it, you can use Windows Deployment Services on it, but even that requires a certain degree of planning. In any case, procuring identical hardware in two years will be near impossible shy of eBay, but with an OEM that's a given.
4.) It's also prudent to order a few extra spares and keep them on the shelf. Either image them before they go out, or use them to swap out bad parts to get the users' machine back into production before you call for warranty service.
5.) If the computers end up with capacitor issues or overheating problems, it's obviously not good. The difference is that if Dell or Toshiba built them, it's your fault and you get to keep your job. If you build them, even though 95% of Slashdot would agree that it's not your fault that the motherboards shipped with bad capacitors, don't even TRY to explain that to your boss - just post your resume up to Monster.com.
Building machines in the office works well on a smaller scale. If you had a 5 member team of engineers doing the AutoCAD thing in their own section, then yeah build those - you'd spend plenty less building over getting Precision units and could build them all out in less than a week. Don't kid yourself into believing that assembling hardware day in, day out, for months at a time, is a good idea. You'll spend enough time designing an image and building out the software stack (corporate Windows licensing and Windows Deployment Services come in VERY handy for this), but your best bet is to whittle the OEMs down to a more reasonable price, pay for the kitchen sink warranty, and give yourself as few headaches as possible. Even if it takes you two days to get the best possible price from an OEM, there's no conceivable way that you'd come out behind.
"I work for a government agency that is having a one-day conference for 1,000 people. I looked into having the conference catered, but then I figured out that if I just buy 100 loaves of bread, sliced meat and cheese, and condiments to make 1,000 sandwiches myself, it'll be much much cheaper. I just need to buy a case of paper plans, some chips and cookies, and we're set. It's at least 10% cheaper! Has anyone tried this?"
.sig withheld by request
I'm Linux user for the past 8 years and can tell you that my 6 years laptop still runs fine.
I've upgraded RAM, changed 2 hard disks and burned 2 fans.
The good thing with Linux is that it does not slow down with time.
The advise: list down all necessary software, find free and open version or just buy one for Linux, get a vendor with good support (Lenovo would be in my list) and you will be just fine for the next 4 years.
Tiger direct inventory is in naperville IL.
And there you can get same day stuff from there inventory.
Are you a federal government agency? You won't be allowed to pursue this foolishness of building PCs out of spare parts. You'll need an enterprise architecture plan and you'll need to document your proposed IT investments on Exhibits 300 and 53. See OMB Circular A-11. Uncle Sam doesn't want to wind up with a bunch of computer parts.
On the corporate side they'll really gouge you. As TFA says, they're spending $1,000/machine and you can easily put one together for a few hundred. Corporate buyers are generally more concerned with having a guaranteed, set price, regardless of any possible hardware failures than they are about what that price actually is.
Large infrastructure needs reliability. Reliability is no better for generic-built than for name-brand, and you give up easy replacement. Costs will *SKYROCKET* purely in the man-hours needed to build and maintain them.
If Dell is giving you a crappy price, turn to HP, etc.
Another non-functioning site was "uncertainty.microsoft.com."
The purpose of that site was not known.
The should have a redundant network, regardless of whether they use PCs or windows terminals. Switched networks are extremely reliable if you buy the right ones and put them into the right architecture. For example, as the windows terminals are more reliant on the network than PCs (somewhat debatable though because of all the web based apps these days), then a better architecture would have been to multiple 24 or 48 port switches uplinked to two separate aggregation switches, such as Cisco 6500s. If one of the 24 or 48 port switches dies, they're cheap to hold in spares, although if you buy good ones (e.g. Cisco, Juniper etc.) they'll be very reliable, and be able to swap it out in no more than an hour, and only have 24 or 48 windows terminals down at once for no more than an hour (more like no more than half an hour). If one of the upstream aggregation switches goes down the other one will take over all the traffic until you replace the failed one.
The Internet's nature is peer to peer - 20050301_cs_profs.pdf
Check out Tigerdirect.com.. Get in touch with them. They sell a lot of PCs and have the ability to fill larger orders. Not only that but that have sufficient assets such that if they sign a contract that contract will be enforcible. I feel certain that they can do better than Dell for you.
If Dell is expensive to you, then the problem is that you haven't talked to the sales rep. If you get 1000 PCs from Dell/IBM/HP (more than 5 actually), then you can give them a HDD and they will image all machines with your own setup, so you can just roll them out at no extra charge.
Excuse me, but please get off my Pennisetum Clandestinum, eh!
My God! I bought Dell for years for far less for a network ready plug-in. Out of the box and onto the desktop. You want to BUILD them??? Good luck with that. You'll spend a lot more in staff time putting them together. Time is money. Space is money. Procurement of parts is money. The idea is not to provide jobs for IT. The idea is to get those PCs out to the clients as cheaply as possible. Think this through. And even Dell is less expensoive than that. You're not doing Alienware for the desktop, are you?
How about a moderation of -1 pedantic.
We get these off of Tiger Direct. Less $$ than Dell. Easy to work on if you had to for example swap a power supply. This is as opposed to some of the stuff from HP and Dell, which for example can have odd star screws and assemblies just to get to the CPU heat-sink and blow out dust. Essentially the Systemax units are as in-expensive as you can build but without actually using your time to do the building, and as easy to service.
If you are buying 1000 of them, I would expect the price to be *MUCH* cheaper than that. I work for a local council - and my last PC refresh I got the PC and a 22" monitor from our supplier for just over $1K/system. I only ordered 60. All components come with 9-5/5 3 year warranty, got access to (thank godness) the ENGLISH HP Government Helpdesk. I only need to do basic diagnosis of hardware faults - HD Dead, loud fans, PC/Monitor dead. We pay for warranty, we let our HP techs figure out the problem. One or two spare desktops is all we have ever needed. I haven't crunched the numbers, but rolling your own is gunna cost you more down the track. We also get our integrator to do the heavy lifting of swapping out desktops.
I build my own PCs for home use since it's easier to swap out parts here and there. I'm not sure if I've bought a completely new desktop in a decade - usually it's an HDD here or a motherboard+CPU+RAM there.
But, for offices, no way.
Stick to the brand names, and even better, the volume mainstream models from the big brands. A couple hundred dollars isn't that much - and an OEM Windows Business license isn't much either - less than 1/2 the price of a retail (not sure what a volume license costs these days). Why? It's in Dell's best interests to minimize repair costs, so they do real testing of the designs. Open up a Dell box and take a look. Only the ports and slots that are necessary are there, and the BIOS is quite limited - for geeks, this isn't so good, but for offices, that implies less stuff to go wrong. It's just not worth the risk that you get some weird incompatibility between your RAM or CPU or MB or PSU that shows up in a reboot every week.
I've heard that Dell used to be highly proprietary with their RAM and so on. Has that changed? Will you be able to make replacements from the standard peripherals market?
I'm not a lawyer, but I play one on the Internet. Blog
Agree with not building your own. If you want to avoid Dell, use a local whiteboxer.
Anyway, does anybody have anything to say about Dell vs. HP vs. anybody else (Lenovo?). Better warranties, better build process, more open architecture?
Does Dell make network PCs (X or Citrix terminals) like HP?
I'm not a lawyer, but I play one on the Internet. Blog
Where I live, Southern California, there are a number of wholesale suppliers offering used, brand name PCs at a steep discount. I buy and resell them to a number of local businesses.
Generally, I will pay $90 for a Dell Pentium 4 2.8 GHz with 1 GB of DDR RAM and a 40 GB hard drive, plus another $50 for a 17'' screen. Keyboards, mice, and cables are included. The machines come with Windows XP (including COA license sticker) and MS Office. The supplier carrys all replacement parts, has a 60 day warranty, and offers very good pricing on repairs after that.
These machines are completely capable of browsing the web, doing email, working in Microsoft Office, printing, scanning, etc without seeming sluggish. Modern desktops with multicore CPUs and gigabytes of RAM have far more computing power than most business users actually need and that excess capacity, and the money spent on it, are essentially wasted.
I have found that Pentium 4 systems are the sweet spot because they have the speed to keep up and are available for an incredible price. Additionally because the systems are brand name (Dell, HP, IBM) I can download the drivers, specs, and manuals from the manufacturer's website and replacement parts are easy to find. I would recommend that you find a similar local supplier. For that same $1000 you would be able to equip 6 desktop users.
My Company just did a hardware refresh. We got Core i5 machines with 4 GB of ram and got them for about $850 each; and that included CDW imaging the computers for us and a 3 year warranty.
Thats a bit less than the $1000 you are paying, and is much easier to support and deploy than what you have proposed. While it may be cheaper to build the computer yourself in some cases, building 1000 computers is a huge time commitment; whereas it takes almost no time at all to order 1000 and have them imaged and delivered to the appropriate locations. Then you just have to install them.
All you noobs claiming 3-4 hours per machine in build time should blow dick and get fired. I can have a system built and imaged in less then an hour and a half. Not to mention for that extra half hour of imaging you could already be building another system and be half way done. L2multitask noobs.
If you're saving money by building machines yourself, you're paying too much for your computers. Which you knew, having asked the question in the first place.
Find a local computer store that builds whitebox machines. My workplace is lucky enough to be next door to such an outfit, and they build our machines for us. Basically, for the Newegg price of the components, we get an assembled, tested (they have a burn-in suite that runs for a day) machine with a 1-yr system warranty and our custom drive image installed. Customer support is great - walk over there and talk to a real person who isn't following a script and knows what he's talking about.
Your local computer store may be crap, but it's easy to tell, and if you have a good one it beats the hell out of buying from Dell.
High-speed Road Trip (18.000KPH)
(Dell warranties may not like you having 3rd part ram)
Not true. If you put in 3rd party RAM, they just expect you to take it out or put in the OEM RAM before they troubleshoot it. And it makes sense-- you go and buy some "high density" crap RAM from PriceWatch that isn't guaranteed to work with the chipset on the machine, it's just standard practice to make sure that it isn't causing the problem.
I don't moderate anymore. Karma penalty for 90% fair mods? Can I mod that unfair?
The problem with many of these people who claim that it was a bad idea or it failed at their corporate office is that they they didn't do their projections properly. I remember a hospital i worked at a while back where they tried to get the staff to use tablets to take down notes and such, the problem there was that they didn't get proper feedback from the staff that had to use said tablets. For one the screen resolution was too high and touch based input made it nearly impossible to interface with the software they already had (EG: the software buttons were too small to account for the random factor of touch based input)
If you really want to do this I'd first talk to the manufacturers find out what their techs recommend. Sales representatives are idiots who just read what the corporate offices tell them to read, and end up being a menace to your staff more than a help to you. It's the designers and builders that can tell you what to expect from a particular piece of equipment maybe consult their returns department to find out what the situation is like there.
Second I'd hire a tech consultant (preferably one that is familiar with your situation) who can talk your guys through initial problems, train your techs, and even give you a better idea of pricing.
Third never start anything like this on mass scale, always start small. For example start with one or two PC's get them working find as many bugs as you can, then work up to a small department that can handle being down for a few hours each day, work out the bugs there then move up the chain. And i can't stress it enough always consult your staff at each step of the process, and keep a back-out clause in your contracts so that if need be you can drop the whole process and go back to your previous state with only minor expenses involved. (Contracts referring to Motherboard suppliers, case suppliers, etc...)
Other things to consider that may be helpful, consider using open source software, such as Linux and Open Office. These two programs can cut your expenses drastically if implemented correctly. Some distributions even come with excellent support (of course you have to pay for the support but it may be less than 10k licenses for win 7 and MS office). Plus if done right you can allow users to log in to their desktops from home, and work from there if needed and Linux has probably the best corporate control system I have seen. You can control a system down to the level of saying which executable can and cannot be run on a system, thus reducing or in many cases eliminating the need for anti-virus software. In the case of Open Office your staff is probably already trained in how to use it they just don't know it yet, it's like MS office 2010 with the MS office 2003 interface. It's easy to use, and it can save files in just about any format. I'm a full time student and many of my professors don't even know that most of the documents i turn in were written in Open Office. The only feature that open office lacks that MS office has is Open office has a slightly less advanced equation editor, which many people will never even use.
To be honest one of the big attractions Dell has for me is their service and warranties. If anything on notebook or PC dies, you call their support line and then the next day you have a replacement part in your hands. Especially in your case, where you have hundreds of machines. Hardware failures on whiteboxes where you have to swop out the component, interface with the differernt suppliers, and sometimes wait up to a week for a replacement is going to eat up a LOT of your time. Eventually, I would speculate, the amount of extra time you spend on hardware support would be much, much more (in $) terms than what you saved by building cheaper PCs.
We keep on looking at this stuff and we just standardized on mac minis without applecare. If a computer does not fail in the first year chances are it never will. If we replace 1 out of every 10 machines out of pocket we are still ahead in warranty costs. Windows volume licenses in corp/education/government can hop machines. The big part is Mac Minis are basically mill spec intel boxes. Standard drivers good speed etc. For 90 percent of the button pushers they are always powerful enough. It seems an odd choice but apple seems to get more reliable machines out of china than anyone else, we simply spend less time with dead boxes with the macs. The big money cost is internal staff assembling your own machines. Even if you think you are fast you are not and after 100 machines you will never want to build another computer again.
What you're suggesting is nothing new. There are many parts in a PC. Motherboard, 2-4 DIMMs, powersupply, case, DVD drive, hard drive, monitor, keyboard and mouse. That's 10-12 parts for a simple PC.
While you'd be able to claim "I got it all from one place, so keeping track of where I got things is not a problem". You still need to consider when you bought the specific item. This means keeping a detailed database or spreadsheet for all thousand machine with service record for accounting purposes.
Additionally, if you're going to build small form factor systems (which means less materials to build and less to throw away) and you want to stay at least somewhat "green", you'll need to spend a great deal of time researching and identifying reliable components (less throwing away) which use minimal power. It's not good enough to just buy what's cheap. You need to consider what the environmental cost of each component is. When you toss a beige box, can you reliably have it recycled? Consider the energy costs in melting things down. How much epoxy is in the system etc...
Then you have accountability. With Dell, you have an American organization (using all Chinese parts, but that's besides the point) which supports their equipment in America and build machines as a whole. Instead of dumping them, you should instead demand you get more for your money. Before buying a new series of machines, be sure to demand they provide detailed environmental data regarding the machines you're buying. Make them work for the money.
Now, if you're going to go the cheap way. Small tossable machines are best. Look for machines like the Shuttle XS35GT (http://www.newegg.com/Product/Product.aspx?Item=N82E16856101099), it's a simple little machine that works well for Windows 7 and Office. For another $150 you can get a 22" screen. Total cost $500 for the purchase + about $300 or so for the software (if you're lucky, Microsoft licensing shouldn't be too bad).
So, when all is said and done, you've save $200 a machine and now, you have to support each machine yourself. Dell will come and fix the computers for you. Now you've got to toss it and replace it. I have no idea what the failure rates on the machine is, but you'll find out quick enough.
Somehow $1000 per machine from Dell actually sounds like a good deal now. And I hate those guys. Out of the 30 Xeon based workstations we've bought in the past year, 10 have had major problems with running high performance software. Dell has now spent a month and sent 4 different guys several times to repair one of them. It's pathetic. The last guy, we almost died laughing at since we saw him take the new motherboard out of a box and hold it in one hand and actually pick his nose and scratch his ass with the other while staring at it for a while.
Lately Dell seems to be going downhill, with higher and higher prices for junkier and junkier computers.
I have done this on a smaller scale... about 150 computers, not 1,000. We built for reliability, so an Intel board with an Intel processor. We haven't replaced a single MB or CPU yet. Now we did make a mistake and went with Antec brand cases, which seemed great for the first year to a year and a half, but then we had what seemed like a 90% failure rate in the power supplies by the 3rd year, It seemed like I was going out twice a week to replace these power supplies, and we just knew what the problem was when the customer called and said "It wont turn on". But the benefit here is that most manufactures carry 3-5 year warranties, unlike Dell's typical 1 year. We sent big boxes of Power Supplies into Antec at a time for replacement... the first two batches they sent replacements at no cost, but by the 3rd batch we never heard back from them after attempting to contact them for months, and I threw a ton of power supplies in the garbage and vowed never to buy Antec products again. I should also note that the 120mm case fans that came with those Antec cases also experienced ridiculous failure rates as well. In Fairness I should note that none of the replacement power supplies Antec has sent us have failed so far... so perhaps we just got a bad model year or something.
So I think the moral of the story is... don't buy fancy brand or super cheap motherboards, just stick with Intel MB and Intel CPU and they will last you forever. And you can built exactly what you want, how you want, for a a good bit less.
And also remember heat is the enemy of longevity... if you want your equipment to last, make sure to have good cooling with front and rear fans if possible, I have found over the years that Hard drives that do not have a fan directly blowing on them have a much higher failure rate than those that are kept cool... same goes for the rest of the hardware. I typically try to get a case with Front and Rear 120mm fans, as they run quiet and keep things nice and cool and will greatly reduce your failure rates.
The bottom line is the labour cost, never mind the parts. Either you value your time or you don't, and if you value your time to you charge it out at more or less than Mr Gou's jumping employees? And, anyway, WTF does anyone think Dell are cheap in the first place? One of the great marketing triumphs of the past twenty years, that is.
unless you plan on dealing with all of the hardware problems and issues that happen when you can no longer get component X and have to substitute, or some user finds that hardware device X will not work with your hardware for whatever reason - just don't.
Dell or other OEM hardware may be mediocre, but it is good enough for business use, and you can get the same platform for x months. also, if the hardware breaks, it is NOT YOUR PROBLEM. You've got better things to be doing than chasing down broken hardware issues yourself. The time you will spend rooting around trying to prove to your hardware vendor(s) that it is a warranty issue, and which component/vendor is to blame for a particular problem is just shit you do not need/want.
this is why big organisations go with a mediocre box from dell/hp/lenovo/etc. you can just diagnose stuff as "not a software problem", log the hardware fault and get on with something else. maybe re-image a spare box you have laying around and hand it to the user.
If you're looking to save a couple of hundred bucks per machine, or get a little better nerd-spec for your users - don't bother it is not worth it for generic desktops. The cost saving is false economy (you'll pay for it in increased workload) and the performance benefit is just not there for office use.
I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
1000 PCs, with a four-year life (that's optimistic), and suppose you work a 50 week year, so 250 days. That's going to be 1 PC per day, every day, built, tested and out-the-door. I hope you're not just one person supporting this 1000 PC organisation, because you'll be spending all your time catching dropped motherboard screws.
OK, so the OP asked about 1000 desktops, but I thought I'd say something anyway. Our server team decided to do this. It started out with a few in-house-built custom servers because we couldn't get what we wanted from Dell. We settled with Supermicro as our MB supplier as they hands-down have the best selection of server/workstation MB's, and are much quicker at including newer tech that the likes of HP, Dell, IBM, etc. As it came time to replace more and more servers I identified a vendor that would build our machines with the parts we wanted burn them in. We kept a handle on the different types of motherboards we used as we were stocking spare parts, and had other support benefits. Plus, if we needed something from Supermicro or our server vendor they both were only 30-40 minutes away. I want to see Dell or HP provide that ;-). It was great.
Another benefit came from the fact that we worked in an area where classified work went on, and not having to escort Dell techs was also a plus. Of course we had actually stopped doing this years ago because it was almost impossible to get someone into the classified area in four hours, so we ended up changing our Dell purchases to four-hour parts only support; but I digress.
We also could consistently beat Dell pricing even for our Windows servers (due to an M$ enterprise agreement the lab had). Dell charges an arm and a leg for memory and hard drives, so this was generally where the big savings came from.
OK. The point is that what we did could be scaled up easily. All it takes is a team that knows what they are doing, doesn't go config-crazy and use every motherboard under the sun, etc. Identify what your users do and what their needs are, create your configs from this trying to keep as few as possible, create images as a vendor won't want to deal with install scripts, buy some spare parts, and you're done. Shampoo, rinse, repeat at whatever interval meets the requirements of your environment.
We actually tried to do this with our Linux desktops as our server team handled all the Linux support short of actual desktop/user-facing support. We got shut down because of the byzantine rule that all non-Apple desktops and laptops had to be Dell.
I just built 20 PCs at approximately $1200 per machine (including my markup/labor and Windows 7 pro licenses) for a company in St Paul, Minnesota.
The specs:
Core i7 870
8GB G.Skill DDR3-1600 (9-9-9-24)
320GB 7200rpm Seagate 7200.12
Gigabyte GA-P55-USB3
XFX Radeon 4870 1GB
Antec 650W
DVD+-RW DL
ATX mid tower
Windows 7 Professional
I beat the hell out of Dell's quote for their required specs (they had a mid-range Core 2 Quad, 8GB of DDR2, and dual Radeon 4550s for that price).
They're very happy. I make about $150 profit/labor on each, and it only takes me 45 minutes to build each machine, image it, get Windows activated, and join it to their domain.
I warrant the parts as long as their built in warranties (ranging from 1 year to lifetime), and my labor for repairs is free the first year (and on-site - better than Dell's standard 1 year ship-it-in warranty).
Granted, 1000 machines is different - but if you get a small army of people that can build as fast as I can, and if you choose good components and make a good base image, it's certainly do-able.
WHO NEEDS SHIFT WHEN YOU HAVE CAPSLOCK/ DAMN1
From the perception (aka reality) side of things, every little thing that goes wrong with one of the computers you build will be blamed on being cheap, and because it was your idea, you'll be the laughingstock who thought he was smarter than the entire corporate world. Even if you ultimately vindicate yourself on the bottom line to the CFO, you will be loathed everywhere else in the company.
Whats your chargeout rate for weekend work?
Minor correction: what's your chargeout rate for 200 consecutive weekends (you did 5 computers in a weekend, the guy needs 1000 computers done...) ? Still having fun? Wife and kids fed up yet? Boss getting angry because 4 years is too slow a turn round time for the 1000 machines to be done?
Could shorten that time by hiring in 10 guys and doing it in 20 weekends but now the company needs to set up systems for managing 10 people, health and safety, insurance, admin for payment etc.
Not scaleable methinks....
well, if your a sysadmin i would suggest planning a good strategy of hardware in similar to this:
1- Most important is motherboard - would suggest selecting a recent, stable and good motherboard combined with 'affordable' if possible letting you upgrade new processors and ram in the next 18 months.
2- Processor type - this will be chosen after you know exactly what users really need to run on their machines. If you have a machine running OCS-Inventory and GLPI you will have the most described data you need. At this time i would go for a mid-end Core 2 Duo as the average user wont be doing much more than office applications (this will depend on each department is running) that run on server side, so they will be running at most client software that connects to servers. stick with a socket that can be upgraded at least by double processing speed.
3- In case you have developers teams, i would suggest the same hardware layout with top processors and plus more ram. i believe the average user would need like core2duo @ 2.13Ghz and 2 Gb ram, just change the processor for the development team like core2duo @ 3.2Ghz and 6Gb ram plus and fast hard drives (this will be a must for running SQL servers and stuff..)
i honestly would assemble a hardware/support team, choose a really nice vendor for a no-brand pc's. The costs will be much lower and you will have what you need for that period of time and more. you just need to have a good plan and really know what users really need to run on each different department. use GLPI and OCS together.
4- OS - Just use Microsoft or any kind of proprietary OS if you really can't substitute it for a Open Source solution, like infra-structure services, file servers, etc.. this will cut your costs a lot to.
good luck :)
...and get Sun Ray thin clients plus a decent server.
It really is. You can undercut Dell there, because you are already on site, plus you can keep a stock of spare parts or PCs. But there are hidden costs associated with it, and you have to face them.
For example you can get all PCs equipped with removable hard disks. If one falls, you just put a spare on in place, pop in the hard disk, and then you have time to sort out what is broken. (Mind the software, some licenses don't like motherboard IDs changing etc...). That is a major cost saver.
Plus once you get rid of the Dell service, you can buy upgrades such as RAM or CPU for (trusted!) third parties. Dell uses long term contracts, you can access the spot market, and use the differential to again save money.
But you still need a good source for new PCs, and Dell is pretty much as good as any. Just make sure that the specs are what you want - big companies have a tendency to shift what they have, not what you want.
Would I want to build 1000 machines for less than the Dell cost - no way. Would I like to build a single machine for myself and damn the cost? - Now we're talking!
IF all that must be run is on the web (in your own servers or i.e. google apps) the desktop becomes close to irrelevant. Enables to use cheaper desktops/terminals/notebooks, and almost any OS
How about BYOC and VMs?
Setup a BYOC (Bring your own computer) policy and subsidy program to help employees pay for work computers. They have the luxury of choosing what platform and hardware they prefer.
In regards to IT maintenance, you can build and maintain a virtual machine image, and deploy it to a bunch of dissimilar computers (VMware Player comes for Mac, Linux, or Windows).
If anything goes wrong, you can provide a temporary computer while the original is being serviced or replaced by the employee.
Going cloudy?
If the infrastructure is very cloudy, and browser based, all you really need is a browser to access information. Why buy a $1000 machine for browsing? You can always go for thin-clients.
This really depends on the infrastructure available.
Windows is too expensive in support costs and training costs. Building-your-own just means more support costs and training costs. Again and again I have seen companies and individuals save money by buying Macs and now iPads because the users take them and go off and do productive work with them without I-T support, without training. And a ton of money is saved by reduced need for I-T consultants. I've seen places where 1 I-T guy manages hundreds and hundreds of Macs because AppleCare is cheap and Apple Stores are nearby when components fail. A $999 MacBook and $149 AppleCare is 3 years of guaranteed computing, day-in, day-out, for $380 per year. Then you sell the MacBook for $400, reducing your cost to $250 per year, and go again with another MacBook and AppleCare for the next 3 years. No viruses, self-patching, automatic backup, reliable Unix, and world class software all built-in and ready to run.
Rewrite this article about phones ... would you save money by building your own BlackBerry? It's an archaic mindset. It would be great if more than one PC maker knew what the fuck they were doing, but that is not so. Only Apple is shipping 21st century systems right now. And users go more easily from XP to Mac OS than they do from XP to Windows 7.
For corporate users, just having Keynote alone is an improvement for them. Their presentations will be a hundred times better, more like Steve Jobs or Al Gore, they will communicate better, sell more. And Keynote is a fraction of the price of PowerPoint.
But I'm wasting my breath because I-T people are very, very rarely involved in doing what's best for their clients. Instead, you will take advantage of their naiveté and keep them as powerless as possible and that means Microsoft.
Whatever you would save in hardware costs you would lose in labor costs having to build the PCs and track parts and systems for potential RMAs.
We currently buy Dell computers and even with our government discounts end up spending about $1,000 for a pretty mediocre computer...We'd also be able to transfer Windows licenses because the Dell installs are non-transferable. However, I couldn't find anyone on the large scale doing this.
I find it rather difficult to spend $1000 at Dell and get anything less than a kick-ass corporate machine. Granted, it isn't going to play all of the latest games at the highest resolution, but it will be more than adequate for the daily corporate grind for several years. As far as transferring the licenses goes, as long as you do your licensing correctly (either through a Microsoft Select or Enterprise agreement) then that is a non-issue.
Most large companies don't build their own for many very good reasons, among them are:
1. You can't get the volume discounts on components than Dell does, therefore you are very unlikely to be able to get component prices comparable to what you would pay for an assembled Dell system. In other words, it's not likely to be cheaper.
2. If you have to build them yourself, then your company/org will have have to pay someone the time to assemble and test the system. This costs more.
3. If you build them yourself, then your company/org will be responsible for warranty service. That means paying someone to repair them when they are broken, but also means having to deal with component manufacturers for warranty service. This can be very time consuming and frustrating, and it costs more.
4. Most component manufacturers have varying time periods for warranty service. Your motherboard may be warranted for a year, but the hard drive three years, the memory may have a lifetime warranty, the PSU may have a one year warranty, etc. On top of that, the warranty usually starts at the manufacture date, not the initial service date, and parts may be sitting on a supplier's shelves for months before getting to you. Having to plan to accommodate that is a complex pain in the rear. With Dell your warranty starts when your machine is assembled and you can get extended warranties for 3+ years that cover the entire PC. That makes it easy.
5. If you build them yourself, you lose out on the corporate management capability that is usually included with corporate PCs.
6. Does your employer want you to focus on building PCs, or focus on something else that is a core competency? In other words, it might be cheaper if your company manufactured their own desks from wood that they bought at Home Depot. But they don't, do they?
On the surface it seems to make sense to build your own, but when you consider what benefits you are giving up and the extra expense that you will be incurring, it becomes very clear that in the majority of cases it actually would end up more expensive to do it yourself.
BUoT is a crock. It's based on the idea that you buy fungible units in the work marketplace, use them for 30 years, and discard with pension.
BUoT leads to extreme specialization and job standardization. "Me? I replace left handed flibbits. I don't have the training to do bertwigs."
You WANT your people to do a variety of tasks. You want your people to do as many different tasks as they are capable of.
I worked for 20 years in IT. Always for small departments. It probably wasn't the best use of my time, but I did cable drops, spec'd switches, built servers, did hardware troubleshooting.
If I was building computers from the component level, I think I'd try to get all my waterfowl in alignment, and hire a couple of summer students to do most of the actual assembly.
***
Business has this idea now that it is most efficient to have zero cushion. Everyone runs full speed from the time they get to the office to the time they go home. Sure that means you have an IT department of 6 instead of 10. But your people don't have time to learn new technologies. Don't have time to have coffee and BS about the interesting problem they just solved.
And when a real crunch happens, they don't have the loads of different experience that allows them to uncrunch the issue.
***
Don't outsource. Insource. Take the attitude that the feudal manor houses did: Try to be as independant of outside supply. In those times it meant that each manor tried to have it's own grain mill, looms, fuller hammers, blacksmith...
Working for the bottom line is a mistake. Oh sure, it's necessary to pay attention to it, but getting that last 2% should not be the main goal.
Too many people and firms think "I do X in order to get Money." Instead, think "I make some money in order to do X well."
Third Career: Tree Farmer Second Career: Computer Geek First Career: Teacher, Outdoor Instructor, Photographer.
http://cs.brown.edu/system/hardware/desktops/ used to be Sun workstations, not anymore with assembled PCs
(Disclaimer: I'm an HP Employee)
But let me use an example wholely within the HP environment. I can buy new, refurb consumer or used refurb'ed business PC. I never go the business PC route for personal use as because I get less horsepower for the buck. Theoretically what I'm paying for in a business machine is higher reliability and lower support costs. However when the support cost is personal time that costs $0 to me, I pay the lower cost.
So I posit is that if cost of support labor and downtime is low, then the self build self support model makes sense. If the cost of labor is higher, then paying for a higher qualtiy, more reliable, extrenally supported solution makes more sense. Those labor costs need to include cost of down time of idled employees as well as support costs.
The brand PCs come with pre-installed c*rpware, which is rather difficult to remove.
I would say, even if you keep buying DELL's PCs, learn to build PCs on your own anyway. Learn it on small number first.
You may have, say, 990 DELL's PCs, and 10 of those, which you built yourself. And then compare. I am not sure that your PC would be worse.
I resell computers that you are looking for. Here is my experience. ASUS has a corporate line that promises that the mainboard will be avaliable for purchase for 2 years, they call it corporate stable. They typically have a 3 year warranty and have fast turn around on the boards. I would probably switch motherboards every year. Other motherboard manufactures may have similar deals. As for hard drives, I always buy Western Digital because I only want one tool to test hard drives and they have been good as far as reliability goes. I would recommend picking one hard drive manufacturer and go with it. Go with Name Brand memory, I always use Kingston. Do your research and pick good memory, never buy generic. I know Kingston has a life time warranty. Cases are tricky but cheap. I have had trouble with bad(noisy) fans and bad power supplies. Spend at least $50 on a case and you should be Ok. Warranties on cases are shorter and are the weak spot for me. As far as video cards go I try to go fanless. I prefer ASUS video cards with Nvidia Chips on them for now. Warranties for these are 3 years as well. All in all, you can build machines with the same warranty or better as Dell offers and with spare parts on hand you repair machines faster than Dells. If you can find a "corporate stable" mainboard then you can control your variations of system configuration. As far as windows goes the OEM versions that I work with are tied to the PC. Not sure what version is transferable. Building a few systems is a lot different than building 100+, unlike a Dell or other brand if you make a mistake in picking a part then you will have 100+ machines that are not right. Do the research, build a few spec machines, test with your apps, buy name brand components. Will you save money, probably but you will control your hardware changes. One area where you may save money is in building the exact machine you want. Dell has a quite a few holes in their line up. I like to save a little on CPU and spend more on memory or other accessories(SSD). Dell doesn't usually have options for things like this or if they do then they are in the high end configs. Others things to think about. Does Dell re-image the machines for your when you have a bad drive and get the system back to a working state or do you have to do this yourself? Does someone have to escort the Dell person around you location? Depending on your answers then your TCO could get better without Dell.
I have experience with this situation and know that building your own pc's becomes a serious waste of time, effort and ultimately money.
The problem with building your own pcs:
Support becomes difficult,
what driver is causing the problem? video card driver causing problems with another driver? Who will fix the issue - no one.
Imaging pc's becomes more difficult
Dealing with the suppliers of components becomes difficult and time consuming.
The invisible cost in in your time and effort in maintaining it all, far more than you can plan. There is no economy of scale when dealing with problems, it does not become easier it simply wastes time.
Imaging pc's should only take 3 minutes of effort: boot up, pxe boot, name pc and it will automatically image. configuring ip, joining to domain, configuring drivers, all of that should be automatic.
We used to use self-built PCs but ran into quite a bit of problems.
1. The biggest problem, warranty service. If a part breaks you could be waiting for weeks before it comes back. If you bought a bad batch of parts you could be buying more to replace them so everybody has a working computer.
2. Technical support. When you buy from an OEM you have one place to call, a self-built computer could have multiple places to call if you can't figure out why a computer is not working.
3. There's no way to guarantee every part will work with each other. You buy your stuff, it all appears to work, then a month later all the computers randomly BSODs because random part A does not like random part B.
4. Mangement meddling. You've got on paper exactly what has to be the same and what can be different. But wait, it says right here on Newegg we can get this other processor for cheaper so we need to get that, and why are you getting such expensive RAM we should get the cheaper RAM!
a) Get a deal directly with Microsoft for Windows Licenses. Volume licenses deal with this. I've seen contracts for the desktop OS and MS-Office for $26/seat, but these were very large deployments. Microsoft made their money on server licenses and CALs.
b) If you are buying corporate PCs, most should cost about $550 ea. When I worked for the US Gvmt, we didn't get new computers very often, so we always assumed a 5-7 yr lifespan and bought the best that we could, which usually cost 3x more than we really needed at the time. Most of the computers were UNIX workstations, not PCs, so the costs were higher.
c) I suspect you don't really have the power to negotiate with Dell for better pricing, but you may be able to get bids from Lenovo and HP and Gateway and choose a better deal for a lessor PC. That may get Dell to offer you more options.
With the corporate model PCs from vendors, you don't have to deal with different drivers, different hardware most of the time. Don't underestimate how difficult it is to do this yourself. I bet you really can't maintain enough stores unless you purchase 10-20% extra parts and have a way to get those parts where they are needed. I've worked with over 20k hardened PC deployments used in commercial vehicles and seen where 15% extra was not enough.
The vendors must charge a little more for the extra hassles in managing components over the 5 yr life of each model, probably $100+ per unit since it isn't cheap or easy to do this. I truly believe you shouldn't undertake this yourself. You aren't prepared for it or the complexities that having more than a few models in you environment will bring.
OK, it is a great idea, on paper. How much will it cost for you to build the PC (parts wise). Let's say $500, how many hours will it take you to build the PC? How much will it cost to have spare parts on hand? Remember, now you must count for purchashing the parts, storage of said parts, etc. I also agree with another poster, I think you should renegotiate your contract with Dell, or look at HP, or some one else. $1000 for a desktop PC is high, especially in a large quantity. Also how many of your users really need the mid-range PC?
..but I wouldn't want to.
Years ago, I worked as the service manager in a high-volume white-box PC shop. I was in charge of the guys who built boxes, the guys who troubleshot them, the guys who supplied inventory, and the guys who had to ship the parts back to the wholesaler for RMA.
Building 1000 boxes is six man-months worth of work, minimum. You need to figure in a 2% failure rate on finished machines, so you need to order parts for at least 1020 finished machines; you should also figure in a 3% failure rate during the build, so order enough parts for 1051 boxes. When the builds are done, RMA the bad parts and keep everything around for in-house spares.
Don't buy white-box based on warranty, because the warranty is useless after 90 days or so. You'll probably have to send the part back, it will get swapped with one that's been "fixed" and sent back to you. Half the time, that means you'll get one that somebody else returned and wholesaler's tech can't replicate the problem with, so he sends it to you, hoping you'll be okay.
As for building 1000 boxes at once, the way to do it is in partnership with a wholesaler. You'll need to rent some real estate, about 3000 square feet. Nothing dusty, and no carpets. You can probably get it cheap for 30-45 days, look for stuff that's been for sale/lease for a while. Hire 8 guys for a month who "like computers". Have the wholesaler ship you a tractor-trailer full of parts. Get a bunch of locked cabinets, lock all the RAM, CPUs, harddrives up. Stack the motherboards in a locked room. Stack the cases in a corner. Your 8 guys can unload a 40' trailer, count the parts, and put them away in ~12-14 hours.
Every morning, each guy who shows up gets his parts for the day. A low-output worker can build five boxes. A high-output worker can build 10. No drills allowed unless they have clutches. At the end of the day, they can demonstrate a working machine running windows, and start your burn-in suite. The next morning, for every box that passes burn-in, you give them $25. Then they pack the machine back up in the box the case came in, stack it in the room slated for deliveries.
Parts can be swapped 1:1, don't allow floating parts on the floor or they will never get put in a machine. Also, only allow clear garbage bags on the floor. No food or drink, either.
Oh - the reason for so much real estate? The most productive way to build machines is to use about 4 feet of table each, and to do them all at once. So, if you're building eight boxes that day, you need 32 feet of table. And 8 mice, 8 keyboards, and 8 monitors. Don't unbox mice/monitors/keyboards for your build, it will cost you time and not increase your reliability.
Make sure you provision these 1000 identical machines with removable HDD trays. That way, when one fails, you rip the drive out, stick it in a spare, and send that machine to be either fixed or pitched. Fixing might be expensive, though. Remember your assembly crew? They're long gone, and probably not very good technicians.
Do daemons dream of electric sleep()?
Buying 1,000 desktops should give you a lot of leverage.
In my experience buying 1,000 desktops gives you *no* leverage with the top three hardware vendors.
I just bought a million bucks worth of machines from Dell a few months ago. Dell's annual sales for the year ending January 2010 was 52.9 billion dollars, so my little million-dollar purchase was less than 0.00002% of their annual sales.
That's not a whole lot of leverage.
we see things not as as they are, but as we are.
-- anais nin
It's been about five years, but I was involved in such a project with a state government. At the time the cost of purchasing a computer with discount was about $1,000 after discount (albeit with lower specs than what that $1,000 would buy today). We assembled and deployed about 500 computers at a cost of $525 (including windows, office and labor). It should be noted, however, that being a state agency, we received really large discounts on the windows and office.
The point is, however, that it's not just the hardware, but the labor involve, too. Since we had an image of the hard-drive, setup was actually very quick. We cloned the drives after they were already installed. It did take about 20 minutes per machine, but other than booting the system, the process was automated. While 20 minutes times 500 machines seems like a lot of time, it actually took less time than cleaning and configuring factory pre-installed machines. Although it would appear that you could only setup 3 machines per hour, since it was automated, a tech could actually setup about 15 machines per hour. Even cutting that to 10/hour and figuring $20/hour in labor costs, it comes out to be $2/pc to install the software. On the assembly side, we had techs assemble and test two machines per hour so that added another $10 per machine.
Of course, your labor costs and pc specs might be different. These were basic office type of machines for clerical work (heaviest graphics work was powerpoint). The real question is whether you have a large enough staff to do it. At 2 machines per hour and 500 machines, it takes 250 man hours to assemble and test. Only one tech and that is over six weeks, plus their own work has to be accomplished. We spread it over 10 techs, so the time was much shorter.
Ultimately, we saved about a quarter of a million dollars by doing it ourself. YMMV
Tender out the contract with long warranties.
We had Dell, thousands of them, and after the tender another company had came out with bid that was really cheap - about 1/2 the price of Dell. Blank Generic PCs - all the same parts configuration.
They where not a small company and sold lots of computers but our order for 20,000 PCs must have been a bit much for them to handle. The failure rate was massive, I am talking about 60%+ PC failure rate. Everything from missing wires/components to component failures. It took twice as long to roll out the newer PCs due to this massive failure rate. Luckily they did stick behind there product and fix/replace the defective PCs eventually. Their internal Quality Control was so poor it cost use dearly in our own IT budget to role these PCs out. After calculating our human costs for these bad PC it cost over 50% what Dell would have cost. Who do you think wanted to eat this extra cost - the departments got the PCs? Imagine our budget sheet and trying to explain why the PCs role out budget was at least double?
You must remember your internal people costs... it can add up quickly if there is any issues.
Rule of thumb: Stick to a reliable supplier of any large order.
The key words in this Ask Slashdot question is "government agency." That means any purchase contract that doesn't go through pre-established "preferred" vendors setup by the administration have to go through a bidding process. Therefore, the OP does not have the option to build the computers him/herself unless they can push for bidding out every component of the computer system. Government agencies also don't pay anywhere near $1000 a computer because they normally get at least 20% off list price before even negotiating a final price. Also AFAIK, government agencies also have a site license for Microsoft software so that wouldn't have to be included in the system cost.
Ignoring that, to achieve good performance for low cost and maintain a stable chipset/memory standard, you most likely want to go with an AMD-based system. They have been using the same socket technology for a few years now and their boards are very flexible with memory chips within a certain family (like any DDR3 memory chip). If you need to purchase from a Tier 1 vendor, Lenovo has a great AMD line-up and is very flexible on their pricing. A channel reseller would be a great option as well and you'd be helping a small business (which in turn helps all of us through jobs and the economy).
With Dells your just paying for a name.
If you don't want to build 1000+ machines yourself, I might recommend looking into a company called Equus out of Mn I beleive.
They will build bulk machines to specs with a 2 year warrantly for much less than you are paying now I bet.
The downsides:
* _You_ will be responsible for catastrophe, not Dell
* Requires more work for staff ($$)
The upsides:
* You can customize the systems so they only fulfill what you need them to (big cost savings sometimes)
* You can select the highest quality components (Dell does not)
* Initial cost will be lower
The questions:
* Does having higher quality components reduce failure rates enough to balance the extra work involved?
* Are you willing to risk your job to give your employer potential cost savings?
* Is the IT staff willing to work more?
Get training on procurement process through whatever level of government you are with (e.g. US Federal, Ontario provincial, etc.). Then you will know what purchasing options are available. The problem is that many departments and agencies have cut their procurement & supply staff, and those remaining tend to be at best amateurs in their knowledge about IT purchases, and most IT departments staff avoid dull training like procurement, and thus their purchases are done in a very ad-hoc manner, and often reflect personal bias (like only buying from Dell).
In the USA, at the federal level the government U.S. General Services Administration runs a program GSA Advantage.
In Canada the federal government Public Works and Government Services Canada (PWGSC) runs a Standing Offer Index including a specific one for microcomputers. For generic office PCs, these are starting points that are "pre-approved" purchases so you don't need to create and advertise and evaluate and have contested a RFQ (Request For Quotes), and as well the purchasers can also do LPO (local purchase offers) or RVD (Request Volume Discount) and likely other techniques I don't know of in particular if the Index is out of date, or a large volume purchase may warrant a discount.
Also learn and understand TCO - Total Cost of Ownership, that is the overall cost of the purchase (i.e. a PC in this case) including repairs / replacement components, down time due to failure - including the cost of have the employee unable to fully function at their job, plus cost in time & labour of IT staff for deployment (roll-out) including disk imaging (by vendor or IT), as well as overall satisfaction of the agency with the IT department's perceived value. This is standard IT management (CIO and below) knowledge, so talk to an intelligent IT manager for any help you need in this regard.
Unless you have scores of idle IT staff, assembling 1000 PCs is a lot of labour and time. And that's assuming you pick a hardware configuration that works, is stable, and is reliable. Most IT departments are staffed with strictly software oriented system administrators as their rank and file, with reasonably few if any computer hardware technicians and tools. And "at-home builders" experience don't cut it in a professional setting; it's a starting point for new hires, not a substitute for fully trained technicians. I'm sure that will cause flames, but I don't care, and to wit, I'm not a computer hardware technician, but I have respect for the few I have worked with, they were excellent at their job, and knew which end of the soldering iron to hold when making custom serial cables.
http://www.aboutwsca.org/content.cfm/id/WSCA
big gov discounts on that contract.
We did this in a small scale for a while, until Enterprise IT put a stop to it. Was making them look bad. What we did was partner with a local PC vendor who took are specs and assembled the machines. This all occurred about 12-13 years ago and back them we could get a PC with recent technology, for that time, for ~$800. Which was less than half the price Dell charged for an equivalent PC. We also did our own support with the local vendor supplying replacement parts for those items under warranty.
Bring in a couple of other hardware vendors like HP and IBM. Tell them Dell is charging you a fortune. Let them fight it out. I can tell you that I would be surprised if one of the other two wouldn't do it for much much less.
DoD Acquisitions has a focus on 'performance oriented' contracting. They suggest you buy things the way you buy managed servers at a server farm like rackspace.com. So, you could try to let a contract for the use of 1000PCs for 2 years and specify downtime, replacement, and repair measures. Offer the contractor rewards for exceeding the minimum requirements and penalties for not.
I think the idea applies pretty well here. I wonder how it would work. I'd call it HAAS - Hardware as a Service.
there are 2 kinds of people. those who divide people into 2 kinds, and those who don't.
Merp. If you have 1,000 PCs, you should have a volume license agreement and NOT ordering software from dell. If you don't have a VL at that level, then you should be submitting resumes after being dismissed for incompetence.
Negotiate better. I work in an all Dell Shop, and we only have about 140+ PC's. We negotiated with Dell and got dam near gaming class systems for $1,100 each, and we only actually bought 38......(core i7 860 @ 2.8Ghz quad core, 8GB DDR3 memory with ATI Radeon HD 2450, 160GB 10K SAS disks......)
Microsoft and the vendor's that provide software for the same require new hardware about every three years. the system is planed that way to make revenue. Linux on the other hand is not out dated so fast. You can use PC for 6 to 10 years without needing to upgrade. So cut your OS bill to close to nothing, and your hardware bill in half.
A $300 beige box would have (slightly) better components, but more importantly they'd be standardized components. Buying an el cheapo Dell without their support is the worst of both worlds. You'll have to deal with any hardware failures yourself, but often won't be able to buy off the shelf replacement parts.
If we're looking at absolute price over any other concerns, I could certainly beat $300 with an Atom based system. For the kind of workload that a $300 Vostro would be appropriate for, that would likely be fine.
I don't think you can save money doing it yourself.
How to save money is forgetting about standardization. That's the big push for a big agency/company AND IS COMPLETELY IMPOSSIBLE TO ATTAIN.
You can buy all new pcs on day one, but when you hire new folks on day 200 or on year 3, you cannot get the same pc even though you wrote specs into the contract that raised the price of the original ones.
I think you're better off buying a bunch of ultra-cheap Dells or whatever and then in 8 or 12 months buy however many more ultracheap NotDells for replacements. This actually makes computer support easier because now the lame user doesn't need to tell you he's got a Dell 2310 with 4G Ram, he can say I've got a Dell or a Rembrandt or Inspiron and you'll _know_ what the configuration is. And if they're cheap, toss them when they break and get the next version from Best Buy.
As a government entity, you should not be paying base contract (GSA or State Contract prices). PCs are a commodity, so you should be developing a good spec for what you want in a PC and issuing a solicitation to all of the vendors on your PC contract (or at least the major players). Basically, you need to make it plain to the vendors that you are looking for best price, period and that they are competing for the business, and you'll start getting better pricing.
It is easy to get attached to different vendors for various reasons -- including things that matter like support experience and certification to do your own repairs. But at the end of the day, you need to be able to balance those things against the 20-30% price difference that you'll see in a product where there is little real differentiation between products.
Check out the NYS Aggregate Buy program (http://www.ogs.state.ny.us/purchase/pcpurchase.htm) for an example of the type of pricing that you can get.
I worked for a college that did this for a while. I inherited the situation. It worked OK for a while, but we got a batch of motherboards and power supplies that were substandard and it created a mess a couple of years down the road. I personally like hardware so I see the appeal but, as another poster pointed out, the person that comes in next may not be able to deal with it. Imagine if you bought 1000 video cards and find out that they have problems working with the motherboards.
If you do decide to "roll you own" be sure to weigh getting a three year warranty on the individual components versus the expected failure rate. Whatever you do, DO NOT buy "gamer" parts. Focus on reliability. Especially power supplies and motherboards.
A better approach, IMHO, is to buy from a major manufacturer, but spec them out with minimal RAM, etc and add RAM from Newegg. I personally like to pull RAM from every other system and put all aftermarket RAM in every other box. This keeps the pairs matched. It's also easy to swap RAM if Dell tech support starts complaining.
This can also be done with harddrives and works well if you're cloning images anyway. You can sometimes buy a larger drive outright for the price of a "upgrade" from the OEM. This way you get spares for older machines. Let's face it -- hard drives die.
Another cost saving measure is to skip the optical drive for some users. Most software installs and backups were ran from the network anyway and it keeps unauthorized installs down. It also saves power. Buy a few externals for checkout.
Even if you buy the computer from Dell, you might find a better deal on monitors elsewhere.
Competition Good, Monopoly Bad.
What would be really cool is if you buy 500 and build 500 and report back to us every year for five years.
Competition Good, Monopoly Bad.
Once a year, you count how many employees you have, you write MS a big check, and you're done with it.
Does this include employees who don't use a desktop or laptop PC in the course of their job, and employees who use a desktop or laptop PC running something other than Windows, such as all the Macs in marketing? If so, it smacks of the "per processor licensing" that got Microsoft in trouble with the US Department of Justice in 1994.
I used to work for a small college, and we would buy ~100 each desktop and laptops direct from dell per fiscal year (barring special CS etc projects). My manager was quoted around $1000 for a decent machine (core2duo when it was the high end 2 gigs of ram 19" square monitor) and I thought we were being gouged. I literally did no more than go to dell.com and look at the deals, they had an optiplex with the same processor, 4gb of ram and a 22" widescreen for $539. When we called our dell rep and pitched a fit, they gave us a further discount, and we wound up saving tens of thousands of dollars. Why shouldnt education and government be able to take advantage of "jope sixpack" deals?
Sauer
Done that last year, but on a smaller scale (about 250 hand-built PC's, for a 400 people's company). It was a very special case: acquisition of a failed hardware supplier business, with substantial unsold inventory. The management was happy to let me try to scavenge something from the pile of "worthless" materials that come bundled with the acquisition. This solved the most difficult part of the project; I bet you will not gain the management approval in normal circumstancies.
You can save a little, but only if you are really, really competent. Most consumer-grade, heawy-pubblicized components are not really suited for business use. You need to focus on some brands (Intel, Asus, Supermicro...) that specialize in professional-grade parts. If you are unsure, pick a barebone system from the mentioned companies. The barebone concept also relives you from the regulatory approval problem. Then, stick to the manifacturer list of tested memory and components. You will save al lot of headcache. Choose a long-term supported chipset and platform: the Intel OEM site will guide you. The last thing you want is to buy a ramdom collection of components, then spend the next months in a device driver hell trying to build a suitable Windows image for the mass deploy of your company desktops. In my case, I standardized on some Intel chipsets with integrated graphics and I was able to use a single Windows 7 image with little efforts.
For my experience, you can save about 20%: is not enough to justify the trobule, really. In my special case the saving was more like 80%, but the absolute saving value was unworthy (in respect to the annual IT budget). I made it mainly because it was fun and because for the company management was a very visible "saving". In reality, you make true savings in the software selection and licensing negotiation.
Don't build 'em yourself, but for sure, spend the money you would have spent on on-site service and warranty options on a good supply of spares. You can just roll in a new machine instantly, set it up with a standard image, and have your user up and running again. Worry about repair/replacement when it's back in the shop, and use repaired machines for non-critical functions.
And that dune buggy can go places it can't.
I don't think we disagree on any fundamental level. For most businesses, with most use cases Prepackaged Dell's are going to be a good choice. There are going to be cases where resources or requirements differ from the norm and that's not the case.
If you have a surplus of IT services, buying them from Dell can be uneconomical. Consider a university that has access to cheap student labour and a mandate to create part time work, or a union shop that has employees they have to pay regardless of actual need.
If you have specialized needs vendors like HP, and Dell (though to a lesser extent) may not be flexible enough to meet them. Consider a development house with finicky programmers who get snarky when their hardware isn't just-so.
For the record I'm writing this on a Studio 1558, my personal workstation is hand built, my personal mailserver is a cheap Atom board.
My employer provided workstation is hand built because I'm a snarky developer who can demand it. Granted that is effectively a perk along the lines of a nice chair or cola in the frige.
If I were putting 100 (or even 10, really) desktops out for general use, I wouldn't want to bother with that. I don't really blame our IT minions for grumbling at us.
Zino doesn't seem to be Atom or at it's price point. You can put together an Atom box for less than $200.
Even ignoring the Linux issue, you don't need to buy licenses if you have an existing pool of them. Especially for shops that haven't yet had a reason to move away from XP.
There's a reason why only fairly small operations do whitebox builds: it scales very poorly. Consider:
-How long would it take you to build 1000 PCs? Too long? OK, how long would it take your entire IT staff, if they did nothing else for 40 hours a week? Still pretty long, I bet.
-Maintenance: when a Dell has a hardware failure under warranty, it's on Dell to repair or replace it. When your whitebox fails, it's on you. The individual part is warranted by a vendor, sure, but replacing the part from inventory and following up with the vendor for cash or a replacement part for your inventory is up to you.
-Speaking of inventory, you'll want to buy a stock of spare parts for those whitebox machines up front. Getting those parts a year or 18 months later can be tough and may cost more. Sometimes, getting the same part even 6 months later isn't easy. When you buy from a big vendor, the vendor is the one who has to worry about maintaining the parts inventory to repair/replace those machines under warranty. I work for a big company with a mix of PC and Mac hardware. PC hardware has limited IT department maintenance support and vendor repair support. *All* Mac hardware support is outsourced to Apple. It just takes too much staff to do that all in-house.
-Inventory database - you'll need to maintain a DB of each of those machines and what is inside them so you know what parts to put in when they fail. With a large vendor, your vendor does that by serial number.
-Refresh. Where I work, we have a three-year refresh cycle. When my machine is three years old, I pick a new one from the available choices and procurement does the rest. If you go whitebox, you have to build each refresh machine as they come due (if government shops have regular refresh?).
All this stuff costs money that is not apparent upfront. Running a support organization. Order fulfillment. Parts inventory. Tracking. This is most of what makes those large-vendor machines cost more for the performance you get; much more so than the vendor's margin, which is fairly thin on a business desktop machine. Chances are you currently have neither the staff nor the budget for that now. If you do, you're overstaffed :p
I build my own machines for personal use, but that doesn't scale much further. Even in a small business, I'm not sure I'd do it.
I've built somewhere in excess of a hundred thousand PCs over the last decade, and I agree with you: the big box guys almost universally suck, especially if you're straying beyond the lines of what the average boring consumer wants.
The most important question to ask before starting on such a project is: how much time will it require, and what is the value of that time ? You have no one to fall back upon when you DIY, you _are_ the support guy. If a machine fails, it is your reponsibility to fix it.
The nice thing about shops like Dell and HP is you can call them up and say "I need 20 pieces of X" and a few weeks later you'll have 20 pieces of X on your loading dock. If you go DIY, you're the one who has to worry about sourcing all the parts for those systems, and sourcing can be a bitch. For one, hardware distributors are terrible at stock keeping - backorders are embarrassingly common and can take months to be filled. Two, hardware manufacturers are extremely fickle and like to discontinue products within a few months, in favor of the new hotness. Some manufacturers specially designate a handful of "long term" SKUs, like Asus' CSM label, but even the CSM is only promised for one year. If you're shooting for an 18-month lifecycle, you should probably buy 18 months worth of gear at the same time and "ride it out". Plus, you can put any extras to work doing other stuff, no one says you have to leave them in the storage closet.
The other great pain, as others have pointed out, is RMA. RMAs suck, they are practically designed to piss people off and waste your time so you end up buying a replacement anyway. Can you afford to lose a machine for 6 to 10 weeks while it gets shipped to Singapore and back ? If not, you need to stock spare parts to buffer that waiting period. Service quality varies widely from one manufacturer to the next. If you're building 5 to 10 desktops, get one spare. 10 to 50, I'd get 2-3 spares. Beyond that you can optimize your inventory, i.e. stock more of the cheap and common things like Ram, HD, power supplies, and fewer boards and processors since those are more specific to a given build and also less likely to fail. Don't forget about peripherals, people spill stuff on keyboards all the time, and mice love to die for no apparent reason. LCDs usually aren't a problem, you can buy them cheaply just about anywhere (yes, even Best Buy) and consistency is less of an issue here.
The main thing to bear in mind, whether you build in-house or contract out, is that in most cases the hardware you're getting from big brands is on par with the cheapest bargain-bin parts you'll find in a real computer shop. Foxconn boards, Hitachi hard drives, minimum-spec ram. That $1000 Dell can be built from near-identical parts for a third less, or you can spend the same $1000 on a much better machine. One thing is certain though: you can't possibly do worse.
-Billco, Fnarg.com
Google does this. It has its own PC support and builds up what it has as a spec. However, if you support it yourself, will your labor be able to support 1000 pcs when 100 at a time (or whatever your numbers become) start failing.
http://www.networkworld.com/news/2006/062906-google-builds-own-servers-for.html?fsrc=rss-datacenter
http://news.cnet.com/8301-1001_3-10209580-92.html This story tells you why they used each component.
Good luck.
That tinfoil hat is a bit tight. Its cutting off blood circulation to your last surviving brain cell.
This is my sig. There are many like it but this one is mine.
It sounds like you could compete with Dell and that you should start a company. Maybe then you realise that 1kUS$ isn't that much for a system.
The saving is in that he isn't starting a company. So he has no costs for inventory, advertising, shipping, distributor discounts, etc. If he is building a dozen systems he can't win on cost over the hardware lifetime, but if he is building hundreds, he says he needs that many, he probably can shave quite a bit of the cost. But how much hardware is going on those machines, to drive the cost that high? Dell sells a reasonable office machine for just under $600, without massive discounts. Companies like eMachines go lower than that, and have similar performance. There is something driving up the cost we haven't been told.
The next obvious question is how much of the cost is software, and how much of that (possibly including the OS) is required cost? The old "easier if they're all the same" argument is usually made by a salesman or lazy purchasing agent, and often doesn't match reality. Data entry jobs which are poking numbers into web forms or spreadsheets don't require proprietary software. That doesn't mean that there may not be some need for commercial software, just that there's a lot of tasks in most enterprises which don't. And the "retraining cost" FUD is just that, people doing data entry, or any activity where the browser is the computer, need to learn login and start application from a menu or icon. Just like Windows. And free software will read/write most proprietary formats, so the need for a proprietary data format doesn't mean proprietary software is necessarily needed. One size does not fit all, there is probably room for saving in software, too.
This might even be a case for thin clients and a few servers, and get the cost way down, not enough information to guess, but a possible large saving. The problem is convincing management that the best approach is finding the most cost effective solution, not in finding the best price on the "way we always did it."
Well, there are a few problems. If you're federal government, you have to only use GAO approved vendors.
There's the build time. Back in the day I worked in a crappy computer store. Three of us could assemble 30 to 45 machines per day. The assembly part wasn't so bad. The OS install was. We had 4 work benches with monitors. 3 would be populated with machines doing OS installs while we worked in any leftover bench. By the end of the day, we'd have a nice queue of machines waiting for the OS. There are faster ways to do it now. If you did an OEM preinstall, you still have to get that thing up and working, so either you do it, or you defer the time off to the employee who's going to complain that their machine is doing "one last thing" before it boots for 1/2 hour.
Still, if you were able to work the same rate, 1000 machines would take your team of 3 23 days to prep.
I'm not a mass produced machine fan, so I'm not trying to persuade you to purchase any premanufactured machine. I'm just warning you. The bottom line may sound good, but by the time you're building your 300th machine, you may wish you had gone another route.
Be consistent in your building. That is buy parts and spares that are identical. Hard drives and CPU fans are the most likely to fail, so it's very important to keep those handy. Power supplies (or at least their fans) are next. CPU's, motherboards, memory don't go so often, but it can happen. Finally, unless someone does a gravity test or has a bad day on "bring your gun to work" day, you won't be replacing cases.
Use name brand parts. Trust me, white box motherboards from an unnamed vendor in China will never treat you right. I lean towards Asus and SuperMicro for motherboards, AMD for CPU(s), Crucial for memory, Western Digital for hard drives, and whoever's cheapest for CD/DVD drives (frequently LiteOn, which have proven to be good)
When you start building, do it in assembly line fashion. One person installs motherboards. One person installs drives, etc, etc. After a while, it's mindless work, but that helps. When mistakes are made, they'll likely be made the same way each time, so you know what to fix.
If you can just write an image with the OS and applications to the drive, that'll make things easier than doing individual installs. Write the image, boot it up, shut it down, and box it.
Since you're asking, I'm guessing you or your department already does the support. You'll probably be pleasantly pleased that you'll always know the answers for hardware faults, because you're never wondering what's inside, and you won't have to call the vendor with a blind problem just to waste time and finally get an RMA number. It's easier if you can take your kit/cart to the desk, swap out whatever you need, and they're up and running in a few minutes.
Serious? Seriousness is well above my pay grade.