Tonight I’ve been perusing “How to reuse your old virtual hosts”. It’s an interesting thought, applied to the data center.
As the data center is upgraded with faster and more powerful equipment, what should be done with the old? Where is the line between value of work from the equipment and cost of support and operation? Is it greener to scrap a piece of equipment with potentially hazardous materials or continue to use it for something constructive, despite it’s lack of energy efficiency?
I tend to follow some basic guidelines when considering this. Dump all retired CRT based monitors, it’s greener to deal with the e-scrap than to suck up the amount of electrical power required to operate them vs far more efficient monitors such as LCD. Retire any hard drives larger than the 3.5″ half height form factor. I just scrapped out a server running a 22GB 5.25″ full height SCSI hard drive, the amount of power to run such a beast is phenomenal by modern standards. Dump any motherboard and CPU combination old enough to not have ANY power saving features at all. Scrap older low efficiency power supplies. Dump extremely small capacity hard drives of any type. Scrap anything using older than DDR type memory.
On the network side, dump anything for networking other than Ethernet (that means you, token ring diehards 🙂 ). Scrap anything that isn’t twisted pair based. Dump anything with less than 100mbps ports. Check routers or switches for security vulnerabilities and scrap anything that is vulnerable and not being maintained anymore.
So, now that we have some decent older hardware that still has some life in it and some degree of efficiency, how do we put it to work effectively? The first idea is to put it to use in a lab. For most lab uses such as compatibility testing or software familiarization, performance of the equipment is a very low consideration. Also, energy efficiency isn’t going to be as big an issue for lab equipment, since it isn’t generally going to be worked as hard as a production server and also not going to require 24×7 run time.
The second option is to create a cold or warm back up facility with it. Once again, energy efficiency is less of an issue, since the back up facility is not getting the use of the main facility, especially if the back up facility is cold, ie not powered up until needed. Of course, the older equipment relegated to back up duty will probably not be as powerful as the main equipment, but it may be adequate for emergency service.
These are just a few ideas for weeding out retired data center equipment so you don’t throw out things that have practical life left or try to use things that will end up costing you more than they could possibly be worth.
Strike the right balance and really be green!
Vern, SwiftWater Telecom