Real back-up is expensive, whether it be media or time. How much you spend on back-up (time or money) should be based on the value of your data, and the cost of the time that the data may be off-line (or the value of that data in some other measure).
When specifying a server for a small company I’ve been told that the price looks expensive. I explain the hardware in the server and it’s cost, and when I get to the back-up device that’s when they choke. No one ever *expects* their hardware to fail, that is a user’s biggest error. On a long enough time-line hardware *always* fails. What I say to the bean counters is this: All of your data, details of everything anyone owes you is gone, you could get no income from any debt that you currently have, and you wouldn’t know that you should how do you recover? *This* back-up device will ensure that that doesn’t happen…. That tends to persuade most people.
If, for example, you’re Amazon.com would it be unreasonable to spend one million dollars on back-ups? I think not, and I’d not be surprised if they spend ten times that. But, for me, an average Joe with a very basic web-site, and not much that is really critical by way of data stored, I copy my servers to each other once a week. I back-up my laptop (excluding music and photos) once in a blue-moon. What would I lose if it all went tits up? In the scheme of things, not much; personally, I’d be upset – but it’d not be the end of my life.
Most people are, I expect, somewhere in between. What people have to realise is that the greater the physical distance they can put between the original hardware and the back-up the better. The more regular you carry out your back-ups the better. And not verifying your back-ups is no better than not backing-up at all. *You must test your back-up* failure to do so will come back and bite you, and prove that you were just wasting your time.
Most of what I’ve written is applicable to business, what about real people? It’s *so* difficult to put a value on photos of the kids – there aren’t many of me as a child or adult, there are hundreds of my cousin; I was (am) an only child to a single-parent who couldn’t afford a camera but my cousin’s father is a very good amateur… What about the photos of the child that died? Or the grand-parent, or parent? The data doesn’t seem so valuable when there is the opportunity to get another photo, but each opportunity is a one off – birthdays, weddings, etc. People eventually die, and then you can’t get another photo. It’s not possible to put a monetary value on this personal data, for the generation that captured the data it may be priceless, it will be valuable too to the next generation (providing we were good at editing out and deleting the crap), but to each subsequent generation it will have less value… You have to face it, once you are a “great-great” relative, interest in you has pretty much gone. So, how much do you want to spend backing up this junk? It really isn’t easy to decide.