Suppose that you're building some custom storage backend that issimply too big to be backed up, so you have only redundancy; thisis probably common if you're building a cloud-style environment or are otherwise dealing with a hugevolume of data. This leaves you with the redundancy history problem , where you're protected againsthardware failures but any mistakes are 'instantly' replicated to theredundant copies.

Suppose that you want to do better than this; you somehow want togive your redundant storage system

Despite what I wrote last entry , I think thatthere is an energy efficiency optimist's view and that it has a lot ofvalidity. It goes like this:

First, there are an increasing number of interesting devices on whichpower efficiency really does matter significantly, either because itsaves you money (such as compute clouds where you are charged veryprecisely for resource usage) or because it lets you run at all (laptopson battery power, low-power netbooks, cellphones, and so on).

Second

There's been somewhat of a movement of late to make programs be energyefficient (both for applications that might be running on laptops andfor applications that might run in a cloud computing environment withdetailed charging). This all sounds very good and worthwhile, and hardto argue against.

But it has a problem (the same problem as always, really). Ingeneral, such efficiency is not free since development time is notlimitless. Saying that applications should be optimized to be 'green' inthis way