Last time, we talked about some basic measures in fixing a computer and keeping it reasonably happy. In the course of that discussion, an incredibly important topic came up: data backups. It's okay if you don't know what that means; just stick with me past the jump.
I would say that the vast majority of computer users -- even highly technical ones, software engineers and the like -- simply do not bother with this. Our modern lives revolve around computers, and more importantly around the things we store on those computers. Documents, photos, music, work, memories. Everything is digital now, and you know what? Digital is simultaneously incredibly robust and incredibly fragile.
The non-technical overview: computers and many other devices include components to store data. There are hard drives, which typically store large amounts on high speed spinning magnetic platters (in some ways it's just a really, really modernized record player). There is also 'solid state' (silicon transistor based) storage which powers things like USB flash drives, digital camera memory cards, and so on. We've also got optical systems: CDs, DVDs, and now Blu-Ray. There's even old school magnetic tape systems. And all of these various storage systems are complete ticking time bombs. All of them.
I cannot emphasize it. ANYTHING you own that stores data can fail at any time, for no reason! Our computers, our drives, our memory cards aren't just appliances that come and go. They preserve our whole lives, and losing that can be utterly devastating.
Backups are redundant copies of data. They ensure that if you lose some or all of your data, it can be recovered from somewhere else. That somewhere else may or may not be readily accessible. It may be in your house, or it may be in a server located anywhere in the world. It could be a single copy, or it could be a chain of them.
Local Backup Devices
Probably the most common amongst people who have any kind of backup in place at all. The idea is that although any given computer and data storage device is fragile, it's not very likely that more than one will fail at a time. By storing the same thing in multiple places, you protect yourself in the event that any one goes bad.
USB Flash ('thumb') drives
These things are what we're talking about, in case you're confused. These little portable things don't hold a lot, but it's more than enough for many people. They're easy to carry and store, require no power, work anywhere, and are generally useful to have any time you need to work on multiple computers.
Unfortunately, these drives also show very, very high failure rates. Most are built down to a price, don't deal with environmental hazards, and sometimes they just up and die for no apparent reason whatsoever. Not only that, but they're remarkably easy to lose. The convenience is undeniable, but that's practically their only strength.
Oh, and they're completely identical to camera memory cards in every way except physical shape. Those photos you never pulled off your camera? Yeah, they might not be there tomorrow. This actually happened to me. Today. An hour ago.
Portable Hard Drives
Same basic idea as thumb drives, but much bigger and typically based on magnetic platters. These are very similar to the drives that live inside your computer, with the caveat that most of them are now exposed to the environment. Most of them don't like being banged around, dropped, bumped, and handled. Of course that's mostly what people do to them, because that's the whole point. There are shock proof, weather proof variants. Not a lot of options are built tough, and they're pricey compared to normal ones, but for back-ups it's worth it.
Even without all that, they (and the drives in your computer) are built around very high speed (5400/7200 rpm, usually) magnetic discs that spin while a tiny magnetic head skims over the surface at a distance of microns, moving individual atoms around as a digital representation of your files. What could go wrong? The platters can shift. The magnetic head can shift. The atoms can get shaken out of alignment. When things go awry in these drives, odds are good that you'll lose everything at once. And with today's modern size drives, we're talking about losing a terabyte of data. That's one thousand gigabytes, one million megabytes. Keeping multiple copies of your files is a really, really good idea.
Conveniently, nearly all portable hard drives now ship with backup software on board. The software is configured to automatically duplicate your files across both your computer and the portable drive, so that there's no opportunity to forget something important. It can be a bit of a hassle to set up, but once it's working you're protected in the case that either one fails.
Optical Discs
This has been a pretty popular approach since the Napster days. If you want to preserve files like photos that aren't really edited, just burn them to a CD/DVD and keep it somewhere. There's practically no risk of accidentally removing those files, since these things are usually one-shot writes that can't be modified. They're also ridiculously cheap nowadays.
The bad news is they're also prone to very, very high failure rates. (Are you beginning to see a trend?) Home-writable optical discs work by embedding a dye in the disc that is then modified by a laser in your computer. Unfortunately the dye isn't very stable and neither is the disc. Excessive heat -- a sun lit car is more than enough -- will cause it to deteriorate. Direct sunlight alone can do the trick. Scratches can render large amounts of it unusable. Excess humidity can destroy them. They're much, much more fragile than commercial stamped CDs, and all of us have had plenty of music CDs that skip. Since DVDs and especially Blu-Rays store much more per square inch, they're susceptible to far more damage, far more easily.
Burning discs is okay for short term use, but at the end of the day they're not particularly robust and the data on them is far more exposed than a portable hard drive.
Local Backup Guidelines
Okay, so I've painted a grim picture but things aren't
that bad. I generally recommend you buy a couple portable hard drives for local storage, the beefy robust reinforced kind. Set a few up to keep copies of your important files and don't abuse them. If you do lose one, or your computer, make new copies quickly to make sure that the redundancy is maintained. You might even consider using optical discs and thumb drives to spread out the risk further.
Do some reading before buying a portable backup drive. If you're not tech savvy, some are a LOT easier or harder to work with because of what software they include. This doesn't have to be hard core technical research; just check out what people are saying in Amazon reviews. Don't rely on professional reviews. It's every-day users whose experience is relevant.
If you really want to be intense about it, your could store one of the drives in a fire-proof safe. The key is just to not have a single point of failure, and make sure that your backups are up to date. How much work are you willing to lose when something goes awry? Make copies at least that often. It's a bit of a hassle but if you follow those basic guidelines, your data will be safe.
Right until your house is leveled by a tornado.
Remote Backup
There is quite a lot of information out there about disaster planning. Most of us buy a half dozen different kinds of insurance. How much of your life is on your hard drive(s), and how many of you count hard drives amongst the items you'd save from a house fire?
For that matter, how often do you actually dig that hard drive out of its fire proof safe and update the files on it? How often do you have an internet connection? Remote backup exists because most of us are on broadband, because even the best local storage devices you can buy are awful, and because your computer can remember to save things even when you don't.
Regarding Security
I actually plan to devote a whole diary to understanding computer security, because it affects everybody. Stay tuned for that. In the meantime, understand one thing: remote backup involves sending your most intimate and important data
to a complete stranger. We're talking bank documents, tax returns, legal paperwork. Remote backup only works if you can read your files again later, and if you can read them then who else can? You're handing over the keys to someone you ostensibly trust. Do you trust everyone who works for them? Do you trust that they're good enough to outsmart the hackers who are gunning for them?
There is a serious, real trade-off here between the safety of your data and the privacy of it. There are techniques to counteract these risks, but they're sophisticated and somewhat involved. As we move more of our computing into the 'cloud', it's a problem we'll have to deal with repeatedly. The choice is, at the end of the day, yours alone to make.
Why are the Big Guys different?
I spent a lot of text explaining why your data storage will probably fail. The big guys can't afford to have such fickle machines run their business, and so in the enterprise world there are vastly more sophisticated tools for saving data. They use the same building blocks, but in massively more complex ways. Just ask anyone who's ever tried to erase anything off the internet. In some cases, there are even privacy issues because these companies
don't know how to delete things. All the research was in preservation. Google probably isn't even entirely sure where all your email lives, because it's spread across their entire network.
That said, there are still regular incidents where Big Guys lose Big Data without a trace. Leaving thousands of customers in the lurch doesn't tend to make data magically appear once a mistake is made.
File hosting services
There's plenty of online services available completely for free that save your files. They're not intended for backup, but people do use them like that. Google Docs (or even Gmail), Microsoft SkyDrive, Dropbox, and so many more. They tend to include a lot of cool features like online editing, easy sharing, and synchronization between computers.
They also tend to be relatively unreliable, relatively insecure, and relatively awkward. Don't rely on them for regular backup. If Google can read your files in the browser, a disgruntled employee can read them perfectly well too. And your documents aren't preserved correctly either in many cases. Dropbox is cool, but it's also heinously insecure. These services were not designed to safeguard your files.
Backup Services
There are plenty of backup services out there, all with various pros and cons. Most offer a small amount of storage for free, and a very large amount of storage for a nominal fee. I've thrown my chips in with
iDrive because they have some cool features. For non-technical users, I've heard consistently good things about
Mozy. For $6/month, you get 50 GB of space; $10 buys you 125 GB. It's not as big as even the cheapest portable hard drive, but it's more reliable and more than enough space to store the things that really matter. Not a bad cost for an insurance policy, either.
Whatever you choose, the basic idea is the same. You install this software, and tell it what files on your computer need to be saved. At a scheduled time (or continuously), the software sends everything to their server to be kept safe. If things go wrong, you can install their software on a new computer and download it all again. Sometimes you can download from the website too, which is handy if you need to retrieve a file on the road.
Apart from the security dangers, there are some problems. If you're a dial up user, you're pretty much S-O-L. Even if you're on broadband, most home internet is optimized for receiving data, not sending it. The backup process can take a very long time, especially the first time you do it. Many internet services have quotas on sending and receiving data, and you're at risk of plowing right past the quota. Heck, sometimes these people go out of business and take your data with them to the grave. (But you kept a portable hard drive copy, right?)
When should you back up?
When should you wear a seatbelt? When should you buy insurance?
If you're using a computer for anything vaguely substantial, then you should be backing up. Period. Computers are incredible at storing, collating, combining, and consolidating data. Exceedingly efficient at it. They're even better at destroying it. Permanent(ish) destruction of important documents is a few key presses away. Typical hard drive failure times are measured in months. Tech guys and salesmen all have their horror stories about which brand is the worst, and they all swear by some brand. They're also all completely wrong. Every drive, of every brand, is ready to die without notice. It's not about buying the best things, it's about buying enough of them so that any one failure is just a hassle of a hundred dollar replacement and not a multi-thousand dollar data recovery service.