On Data Loss and Backup Solutions

Ever had your entire book and all your notes disappear before you finished writing it? I really hope the answer is no--and let's keep it that way. My day job is as a programmer, but that also includes a lot of systems admin work. I'm used to being required to have lots of redundancy and backups in my computer systems, but this isn't an issue that your average home user is going to think much about. Total losses are rare enough that you don't hear about them too often, but they do happen, and perhaps more than you might think. Here's some information on how you can best protect yourself.

Most users know deep down that they should be doing backups, and the most common solution is CDRs or DVDRs. Problem is, this kind of backup is a bit of a pain to do, which makes users less likely to actually do it. Odds are that you've never lost any significant data, and so your perceived risk of losing any in the future is rather low. But computer problems of all bents happen all the time, and even if you religiously back up every week or every month, that may not be nearly enough.

How would you feel about losing a week or month of your work? I lost three days' worth of writing once, and that was terrible. Ideally you want a backup solution that works in something approaching real-time, so that whenever you do some work it gets backed up. This is what your corporate IT departments have been doing for years, and it's not as out of reach as home users seem to think.


Risks
Let's first consider the main risks, and then talk about solutions for each one.

Theft is probably the absolute worst thing that can happen to your computer, because there is much less chance of recovery. And while theft of PCs isn't very common, a surprising number of laptops get stolen. Corporate IT departments have all kinds of policies in place on what data has to be encrypted on employee laptops, and what data can't be put on those laptops at all, because they are aware of this risk.

Obviously you can try to take steps to make sure that your computer isn't stolen, such as putting cable locks on your laptop when you're in a public place (trust me, this is not as ridiculous as it might sound at first), but you also have to consider the fact that your computer might get stolen no matter what you do. And the only way to solve this problem is to have a copy of your data that is not on your local computer.

Fire is another extremely bad way to lose your data. Even if you've been backing up to CD, or copying data between your laptop to your PC, if your house burns down with everything in it, that's it for your book.

Even if you have a fireproof safe, most of those are only rated for 30-40 minutes or so, and don’t always do the best with plastic media inside. The ones that I linked to above are really expensive and do a better job, but they still won't last forever in the middle of a raging inferno. Having a good fireproof safe might well work, but it isn't guaranteed. Lowes has a wide range of prices for fireproof safes, as do many other retailers, and the smallest of these are affordable but relatively ineffective.

The only foolproof option in the case of fire is to have another computer somewhere off site that you can back up to.

Mechanical Failure of hard disks is a much more common cause of data loss than anything mentioned thus far. If your hard drive physically breaks for whatever reason, there are places that have forensics hardware/software to get that data off your dead drive. Prices range widely, as does quality of service, but you might be able to get your book back for "only" a few hundred dollars. The best solution here is to always keep your data on more than one hard disk.

Power Surges are even more common than mechanical failure. Hopefully everyone has their computers on surge protectors, but the quality of that surge protector is an important consideration. Some of the inexpensive ones just don't do a very good job. If a power surge does hit your computer, these days the most likely casualty is going to be your computer's power supply, which costs $30-$50 in a PC to replace and isn't going to lose any of your data. But that isn't to say that your hard drive is immune by any stretch. Make sure that your computer has an adequate surge protector.

Even better, you could get a battery backup unit (also known as a Uninterruptible Power Source or UPS). The smallest of these run for about $40, and will provide you with great surge protection as well as about 10-15 minutes in which to safely shut your computer down if your power does go out.

Improper Shutdown and Jostled Disks used to be a big problem for hard drives, but they aren't so much of an issue with modern drives. You still shouldn't hit or move your computer while it is on, and you should try to avoid turning off the power to your computer while the hard drive is actively writing, but the catastrophic failure that was common in such cases in the 80's and 90's isn't as much of an issue these days. Back then the armature inside the hard drive was more free moving, and so the above situations would cause the armature to skip across the data platters, scrambling little bits of data every time they hit.

That's much less of an issue today, but still not something to play with. Are you moving soon? It would be a good idea to make a backup of all the key files on your PC before you do. A battery backup unit/UPS unit is also a life saver when it comes to improper shutdowns, since it can help prevent those when your power blips or goes out entirely.

*If you'd like to know more about common causes of hard disk failure, there is a pretty good article on that here, and it covers some other topics like firmware corruption that I'm not going to touch on here.

Viruses are the very last way to lose your data that I am going to discuss. Certainly the average user should have antivirus software, anti-spyware software, and all the latest patches on their OS. This doesn't have to cost you an arm and a leg. AVG Free is free antivirus software that a lot of people swear by, and Spybot Search & Destroy and AdAware SE Personal are both free spyware prevention/removal tools that I love. Use all three if you're worried that you have a virus or spyware on your computer, because they all catch slightly different things. And remember that even then, they might not catch all issues on your computer. Browser Hijacks are notoriously hard to catch, for instance, but fortunately they generally don't pose a risk of data loss.


Solutions
Still reading? I hope so. Because now that we've covered the main risks, it's time to talk more about the various solutions available to you.

Hard Disk Redundancy in a single PC is an excellent way to protect yourself against all the various forms of hard disk failure, and it is generally convenient because it keeps your data on multiple hard drives in a single computer. It's unlikely in the extreme that two hard drives in the same computer will fail simultaneously for mechanical reasons. This obviously doesn't help at all when it comes to theft or fire, however, and probably won't be much use in the case of viruses or power surges either.

Simply copying data between two hard drives in one computer is a valid approach. You can do this by hand on any computer that has multiple hard drives (and if you only have one, it's generally very easy to install a second one).

If you have some extra money to burn, you can always set up a RAID mirroring array on your home PC. If you build your own computers, a lot of motherboards on both the Intel and AMD side have this built right in already these days--with the rise of SATA, these sort of controllers are much more prevalent than they ever were with IDE. However, there are advantages and disadvantages to RAID even if the cost isn't an issue, so it's something to consider.

Odds are fair that you could stand to lose a few hours' work if it actually came to that, so just copying your data between hard drives when you finish for the night might well be enough protection for most people (it is for me).

Data Recovery Software is a valid option if your hard drive is physically fine, but has had some sort of software-initiated data loss. For a virus to truly, permanently wipe your hard drive would take hours--and most viruses don't try to do this in the first place, because that would make them more fatal but less virulent (same as human diseases, they won't spread much if they immediately kill every host they infect).

If a virus was to go for your data, it would probably just try to corrupt or delete individual files. Corrupted files might be hard or impossible to recover from (this is why you should have antivirus software), but deleted files can often be recovered relatively easily if you act fast. If your hard drive is still working but you lost your data because you got a virus or accidentally deleted some files, software like File Rescue can get your files back much less expensively than sending it to a forensics service--but you must use this software as soon as you realize you have a problem, because your OS will overwrite deleted data eventually. This gradual overwriting process could take months or years if you never use much of your hard disk, but by the same token it could take minutes or seconds. Best not to tempt fate by waiting.

This software is also ideal when the FAT index on your hard drive gets corrupted. Essentially, all your files are in a central index on your hard disk, and when your computer "deletes" a file it just removes it from the index. Software like the above ignores the index and just looks for files on the main part of the disk itself. Improper shutdown of your PC can sometimes cause your FAT index to be corrupted, just as if a really nasty virus had attacked you, and so that's another situation where this sort of software comes into play.

CDRs / DVDRs are the most common way that home users choose to back up their data, but I don't think a whole lot of this approach. I've already mentioned how this sort of a backup is time consuming and something of a pain (thus making users less likely to actually do their backups), and I've also noted how these are just as vulnerable to fire unless you have them in a fireproof safe (and arguably, even that might not be enough).

But the other issue with CDs is that they don’t last as long as people thought they would. When CDs first came out, they were touted as likely to last 1,000 years, and CDRs up to 100 years. But these days people are already starting to have CDs (especially CDRWs, and to a lesser extent CDRs) that are failing because the plastic layers are coming slightly apart, allowing the thin sheet of metal between them to oxidize--thus losing your data. Gold CDRs are reputed to last longer than the regular CDRs (gold doesn't oxidize), but your mileage may vary. There just isn't enough of a history with CDs and DVDs to know what their life span will really be. Check back in a hundred years.

Other Removable Media besides CDs and DVDs can also be used. The most popular at present are USB flash drives, because they're just so darn easy to use (and these days they're pretty inexpensive, too). These are solid-state memory and don't require a battery to retain the data, which makes them ideal for semi-long-term storage. I'm not sure exactly what the range of expected life for these drives is, and I certainly wouldn't trust one as my only backup of anything (let alone my only copy), but as a quick, portable backup option this is a good one.

In the past, floppy disks and zip disks were the primary backup options for the home user. I used these myself, and in fact I still have some 15+ year-old floppy disks that still work fine and have all the data I put on them. But these disks have such a weakness to magnetism that their data seems less safe compared to any of the more modern alternatives. In a business environment I've also used the ever-popular tape backup setup, which can store massive amounts of data very cheaply, but I found the reliability to be too low for the purposes of my company and so we switched to the final solution I note below.

Remote Backup is the last solution that I'll touch on, and in my opinion it is by far the best. This solution has your data on another computer (or multiple computers) elsewhere in the world. That protects against any sort of hard drive failure, viruses, fire, computer theft--hell, it even protects against localized meteor strikes. Pick your non-global natural or unnatural disaster and this solution works, because no matter what happens to your home computer or laptop, your data is safely far away.

If you've got a large amount of data that you want to save in this way, you'll be hard pressed to beat the offering of Connected TLM (now a division of backup industry titan Iron Mountain). It'll cost you something every year, but you get a whole lot of redundancy. There are a lot of other solutions, as well, such as IBackup, but I know less about them. UPDATE: There is also now a program called Mozy, which I give more detailed information about here. This is now my far-and-away favorite Remote Backup solution.

One clever thing that you can do is email it to yourself with GMail or another free mail provider, so that you have a backup on their servers. I particularly note GMail not only because it is free, but because they give you so darn much space. Do note that if you have a POP account that automatically deletes off the server after processing messages, that kind of defeats the point--generally speaking your ISP's email services aren't going to work for this purpose.

Finally, if you have remote access to a computer in another location, you can just copy the files yourself. Have a friend or relative in another city? The two of you could arrange to swap files on some interval--either by setting up SSH on each computer (that's generally safer than FTP, which isn't usually encrypted), or just by emailing files to one another. It largely depends on how much you want to back up, file size allowances on your respective email providers (many have a 2MB or 10MB send/receive limit on individual emails), etc. If your employer doesn't mind, and you have a VPN connection to work, then you can also consider backing up a few key home files to your work computer. But please don't do this without your employer's knowledge and permission.

UPDATE: Free Online Document Management Systems are another way to go, but each one has its limitations. Google Documents allows you to upload as many files as you want, but they have to be HTML, plain text, MS Word, or MS Excel files. That's nice, but I also need to back up copies of my TreePad database. If you don't have any such files, then this might be an ideal solution for you. Just bear in mind that Google's terms of use for this service includes the line "Google has no responsibility or liability for the deletion or failure to store any Content." This is true of GMail as well, but it just bears pointing out that this doesn't have guaranteed retrieval like the paid services do. So it's a good idea not to put all your eggs in one such basket.

Yahoo! Briefcase is another such service, and it allows you to upload files of any sort, but there is a limit of 30 MB to what you can upload in total. You can pay to get more space, but you probably wouldn't need to until you had 20-30 full books stored on there. Their terms of service were not clear, but I'm fairly certain that they don't provide any retrieval guarantee as a true escrow service would, either.


Closing Thoughts
There are a lot of options here, and which ones you choose to take will depend on you and your own individual situation. Generally speaking it's best to use more than one approach to truly safeguard yourself. Just for the sake of example, here's what I do:

-I have three hard drives in my PC (for capacity reasons), and I manually copy my writing files between each one at the end of each writing session (I don't use RAID of any sort).
-I keep my desktop PC on a CyberPower CP425SL.
-I do have GMail, and very occasionally I'll send myself an email with a copy of my book.
-My employer allows me to, and so I back up my writing files to my work computer over VPN every few weeks. It would probably be better if I did this more often.
-Whenever I do work directly on my laptop, I copy that work to my PC as soon as possible.
-I don't use a cable lock with my laptop, but I also don't ever take if off my lap if I'm using out in public (which is rare in itself--mostly airports when I travel for work).
-I use both Spybot and AdAware to keep myself protected from spyware. This is a good idea in general.
-Trend Micro is actually the antivirus that I use, but it isn't free.
-I don't do any form of CD backup.

-UPDATE: I also now use Mozy.


If you've read all this and feel like this is just ridiculous over-protection and a complete waste of your time, consider this: you won't feel the same way if you actually do lose your data. Through my work at my day job, I've just seen too many computer failures to treat this cavalierly.