Slashdot Mirror
How Do I Put Unused Servers To Work?
olyar writes "I worked for an internet start-up last year and during the 'we have plenty of money' phase, a lot of server hardware was purchased. Eight months later, there is very little money, but we're still plugging along — using only a fraction of the hardware. We just cleared out a co-lo and I now have a stack of 17, 1U servers in my garage. Each of those has 2 servers, each of which is a 2-processor, dual-core box with 8 GB of RAM. Add that up and I have 136 processors and 272 GB of RAM with nothing to do. The IT guy in me thinks that's a waste of FLOPS. The wanna-be businessman in me thinks its probably a waste of money as well. So I've been brainstorming ways to put all of that power to good use. Any ideas?"
Set up a Eucalyptus installation and compete with Amazon. OK, not really, but you can have an EC2 workalike without the usage charges. Use this setup as a sandbox to test migrating your current IT infrastructure to AWS. When it all works, hit the switch and actually make the move to EC2. Then, sell your no-longer-used hardware. You just converted Capital Expenses into Operational Expenses.
... is whot bwings os tugevza tsuzay.
Ive worked in non-profit IT and servers is one thing they always needed. They dont really need more hands at soup kitchens, they need equipment and expertise. I bet your local food bank would love that stuff. I also bet their existing servers are a couple of old non-raid desktops moved to a closet. You can probably just call someone at Feeding America and they would dole out the servers to deserving foodbanks via their grants system.
Also, if the businessman in you doesnt have a business plan then theyre just going to waste and will probably end up in a landfill. You might as well give them away to someone who needs them.
Have you thought about just selling the servers?
Let's say the 17 U1 servers cost 2,500 each. And let's say that selling them now would gain 1,250 each. That's 21,250 dollars available there.
Now let's say that Moore's Law continues to hold. And that you need the additional capacity when the economy makes a miraculous turnaround in 2 years. By that time, it should cost you less than 21,250 to get the same capacity back. And you would be doing it in half of the number of servers, which implies a space and power savings.
In that case, it is downright advantageous to sell now, buy back later. It all depends upon when you think you will need the capacity again. Too soon, and you will pay through the nose for selling. Too far away, and not selling now saddles you with old hardware.
Other options ---
If you're set on keeping them, I see only a few other options. One would be to see if any established small-to-medium sized businesses would like to lease the capacity of your servers. Perhaps those companies who sell time on private servers on video games could use them when the next one releases. Web hosting is probably a bust, but I wouldn't be surprised if a local university would be interested in leasing the iron for better rates than your garage pays. There is also cpushare.com and other cloud computing projects, but it doesn't seem like they're paying out at all.
The ______ Agenda
I wouldn't sell the equipment. If you have a colo you already do business with and a lot of extra server hardware, try subleasing it to someone you think might need some extra server capacity.
Sure, it's a lot of work to find customers, but with that much hardware sitting around you have a lot to offer.
You can't legislate goodness. Let each to his own destiny, by will of his freely made choices.
Create rainbow tables and charge a small fee for access. If you target M$ Office passwords, specifically the password to open, 40-bit RC4, target the possible keys because there are less possible keys than are possible passwords. See Ophcrack office, Rainbow crack office and Elcomsoft AOPB.
I'd argue he should sell 16 and keep one for himself. While computers do rapidly depreciate in value, having a high-end server for home use can be nifty (so long as you keep it someplace where the noise won't bother you).
It takes a decent amount of electricity to run that much hardware.
A 2KW setup of machines (all crunching numbers) could heat up a room as efficiently as a 2KW electric heater, so why not use it in this way? You could even make a climate control that starts/stops @Home-processes to get a constant room temperature. Sounds fun (and a bit nerdish).
"It's too bad that stupidity isn't painful." - Anton LaVey
Twenty thousand dollars investment in a startup on overpurchased server equipment isn't all that bad. You should figure that a company with a dozen people should be pulling about a $million/year gross, against which $20,000 to ensure sufficient delivery capacity is a pretty wise investment.
I figure that having at least 100%-200% additional capacity on-hand at all times is a good idea - computing power is relatively cheap, while the cost of downtime can be astonishingly high. Personally, I start upgrades when load averages approach 50% at the highest usage part of the day. (For me, about 10:00 AM)
That said, if the servers really aren't needed, every day they lose value. Put 'em up on Ebay and get what you can for them before they depreciate any further. Space them a few hours apart over a week or so, so that you don't depress their sale price.
I have no problem with your religion until you decide it's reason to deprive others of the truth.
Free operating systems like OpenBSD and DragonflyBSD continously need to compile their ports tree in order to make snapshots available for download and testing.
Compiling the whole tree takes quite a long time, especially with piece of software like OpenOffice.org. The currently can't build snapshots as often as they (and user) would like to.
Some other projects like Drizzle and GCC are also looking for remote build machines for regression testing.
Your unused servers can really help the open source community.
{{.sig}}