Are there any recommended minimum standards for a "mission critical" computer?

Question

By "mission critical", I mean that it would be dangerous or expensive if the computer failed, eg: a computer being used during a surgical operation, or to serve an important website.

I'm looking especially for "official" or well-recognised minimum standards, rather than just people's opinion about what would be a good idea to have.

It can be assumed for the purposes of this question that there is no risk of a network attack. I'm looking for physical requirements to stop the machine failing (eg need to have redundant PSUs, etc).

Answer 1

I don't know if you ever read the license agreement when installing Windows, but it says not to be used if failure will result in loss of life or limb, so depending on what type of business you want this for you may have a bigger task than you thought. For what it's worth I've worked with servers that have multiple CPU's, multiple Network Cards, mirrored hard drives, and dual power supplies, but they still had single points of failure. For a whole lot less money but a whole lot more reliability you can get two identical PC's and set them so that files stay in sync. Anyway, it worked for me. I attached some web links for your viewing pleasure. Good Luck!

Answer 2

Probably the easiest way of addressing this would be to have a 'cluster' - two linked machines that act as one. If one of the machines (called a node) has a physical hardware problem, the other takes over. It's hardware intensive (since you'll need two machines), but it's very effective, and is what gets used by most large businesses for their mission critical machines.

In terms of general concepts, you'll want to prevent there a single point of failure (SPoF). This goes beyond the computer itself - having two power supplies, for example, does no good if the electricity goes out and you don't have a generator.

The bare minimums that you'll want to look at are dual power supplies, and hard drives in some sort of RAID configuration (RAID 1 or RAID 5, typically). Depending on budget, dual CPU/memory/system board is good too - the more redundancy, the lower the chance that any single hardware failure will knock it out.

Answer 3

If it is important enough, you would go with redundant systems. Two complete systems which automatically back each other up continuously or at preset times. Of course you would require external battery supplies for power losses... It's possible and not all that cheap but it depends on how important the 'mission' is. You might look at the big systems such as used at the trading houses such as the stock exchange or the military. Most of the best systems are kept simple as possible to keep possible failure points to a minimum.

Answer 4

Well u said before ""I mean that it would be dangerous or expensive if the computer failed, eg: a computer being used during a surgical operation"" ......HUH this is really mission critical (if computer have a hardware/software failure in to middle of medical intervention and pacient die ......we have a critical situation here) .In order to prevent this happend you need to have one computer/server AND one or two BACKUP computers / servers .If main computer/server FAIL the backup must INSTANTLY take all tasks form main computer and take it to the finish. Now if u want more details.......
Feel free to contact me on yahoo messenger for more details.

Answer 5

hmmm, really thinks on that sugery thing, lol. well if its THAT IMPORTANT id have 2 power suppilys incase of one failed, Battery Backups, atleast like aGB of Ram and for a web page u better get it as a server, cause there designed for that, but for your webpage thing i wouldnt be soooo careful, ONE Power Suppily, Battery Backup, 3GIG processor, 1Gb or Ram, and mabe a 80-200GB Hard Drive, nothing else too special