Backup Basics
Planet Fox > Computer World > Backup Basics

Backup Basics

I am absolutely fanatical about taking backups. I have several hundred gigabytes of data that's really important to me, some, like photographs are irreplaceable, while other things like source code, artwork, purchased software, etc... can be replaced, but not without considerable effort or expense. Since most computers store their data on hard discs, and since hard discs have moving parts, failure is inevitable, at some point.

I have a lot of computers, and since I don't want to spend all of my time moving files between them on USB drives, I've setup a central fileserver I can access from anywhere in the world. This simplifies backups a little, since all of the important files are in one place. If you plan to do this, I suggest you build in as much redundancy as possible and take regular backups, since the fileserver becomes a single point of failure.

Redundancy is important

In my system, I have two very reliable (server grade) Hitachi Ultrastar 1.0TB discs mirrored together in what's called a RAID-1 configuration. Both discs behave as a single virtual 1.0TB disc, and everything that is written to this virtual disc is copied to both physical discs. Even if one physical disc fails completely, no data is lost. This still presents a problem, since you're still using hard discs, and there's still a chance that both could fail simultaneously. Some circumstances where a mirrored set of two or more discs could fail include natural disasters like floods, tornadoes, and earthquakes, electromagnetic pulses from solar flares and damaged power lines, power supply failures, fires, and water leaks. Even if both discs continue functioning indefinitely, there's a chance that malfunctioning software or a mailcious intruder could delete sensitive files. It's also highly likely that a user will accidentally delete an important file at some point.

The best solution to all of these problems is to use some form of removable media. If your data is really important to you, I suggest you make multiple sets of media and keep one set offsite. What type of media you use, what format you store your backups in, and how often you need to make backups are all things that you will need to decide for yourself, but I'll try to help you make the right decision.

Choosing media

Optical Discs

If you only have a small amount of data, say less than about 40GB, the cheapest and easiest way to go is to use optical discs. They're cheap, relatively reliable, supported on every platform, and don't take up much space. Since the media isn't part of the drive, it can be more reliable than a hard disc. The biggest advantage is their ubiquity, you shouldn't have a hard time finding a machine capable of reading optical discs for quite some time.

Please keep in mind is that, for the most part, recordable optical discs are not an archival storage medium. Optical discs are made from several layers - a polycarbonate disc, a laser-light sensitive dye, a vacuum desposited layer of aluminum, and a protective coat of lacquer. Over time, the lacquer layer can separate, the dye, which is the layer containing the actual data, can fade, and the aluminum coating can oxidize. Commercially prepared discs are immune to some of these problems, since the bits are pressed directly into the plastic rather than written by laser.

A standard grade CD-R or DVD-R will usually have a rated archival life of around 5 years. For system snapshots this should be OK. In terms of commonly available discs, I recommend Emtec, Fujifilm, JVC, Pioneer, and Ritek. Stay away from brands manufactured by Infodisc, which included Memorex the last time I checked. If you have data that needs stored on a longer basis, but you still want to use optical discs, I suggest you buy a finer quality of disc, such as Mitsumi's MAM-A formulation, which is rated for 20+ years. Regardless of what type of disc you use, you can prolong their life by keeping them in black plastic cases, in a low humidity environment, and away from sources of heat and UV radiation such as sunlight and fluorescent lamps.

Good quality CD/DVD drives are relatively inexpensive, typically costing under $200 depending on options. I suggest you go with the highest quality model you can afford, since you'll need it to be there for you when all else fails. I would highly recommend TEAC, Plextor, Samsung, and Panasonic drives for their highly reliable service and good performance. Compression tools like WinRAR can be used to increase the amount of data you can fit on a disc library, and can even split compressed archives into 700MB or 4.38GB for writing to CD or DVD.

Tape

Tape is sort of the gold standard for backups. Like optical discs, they have the advantage of the media being physically separate from the drive. They're small, fast, highly reliable, and are much more practical for large volumes of data than optical discs. Tapes can be used for archival purposes, since most tapes come with a lifetime guarantee and are rated for a 50+ year archival life. The only drawback is initial cost, which can be higher than optical discs.

Wading into the deep pool of  tape drives can be kind of intimidating, since there are a variety of tape formats. In my opinion, the best system for workstations is DAT, which is also called DDS or 4mm. It uses the same type of cassettes used by high-end audio gear in the 80s/90s (DAT stands for digital audio tape), which are very small, about half the size of a compact cassette and at the time of this writing hold up to 80GB.

Tapes and their drives are usually referred to by their capacity, the original DDS-1 tapes from 1990  were the same capacity as the audio tapes, 2GB. All DDS drives have hardware compression, so when referring to a tape's capacity, the manufacturer usually lists the tape's capacity for text data, which is twice the native capacity. Data that's already been compressed, like JPEG images, MPEG videos, and zip archives won't benefit from the compression, and will take up about the same amount of the space on the tape as they do on disc.

Here's a list of the currently available generations of DAT/DDS tapes and drives

Type
DDS1
DDS2
DDS3
DDS4
DAT72
DAT160
Native
Capacity
2GB
4GB
12GB
20GB
36GB
80GB
Compressed
Capacity
4GB
8GB
24GB
40GB
72GB
160GB
Speed
0.3MB/s
1.5MB/s
3MB/s
4MB/s
8MB/s
10MB/s

New tape drives usually cost a few hundred dollars, fortunately older drives are usually available for very little. I bought a DAT72 drive on eBay a while ago for $30. Tapes for DAT drives are usually fairly inexpensive, less than $20 each at retail prices for the current generation, and significantly less if you're a bargain hunter. In general, DAT drives are backward compatible, so if you upgrade you won't need to re-write all of your archives, and if you only need to store a small amount of data you can use a cheaper, lower capacity tape.

To get the maximum lifetime from your tapes, try to store them somewhere away from high temperatures and humidity, keep them in their platic case, and do not expose them to strong magnetic fields. Other than actual magnets, sources of magnetic fields can include electric motors, soldering irons, CRT displays, and electrical transformers.

Hard Discs

I personally would not want to depend on a hard disc for a backup. The main disadvantage is the reliability. Since the media is part of the drive, if the drive fails, the only way to recover the data is to send it to a company that specializes in recovering data, which is almost always an expensive proposition. The advantages of using a hard disc include low cost and speed. If you have very sensitive or important data I strongly suggest that you save it to removeable media, although it's perfectly acceptable to use a hard disc for taking system snapshots.

If you plan on using a hard disc for your primary backup, consider using a RAID configuration. Even if you don't have a hardware RAID controller, most operating systems will let you setup a virtual RAID. In Windows you do this through the Administrative Tools section of the Control Panel. The backup disc should be physically separate from your primary disc, never use another partition on the same disc as a backup.

Using an external USB or eSATA drive removes the hazard of a power supply failure causing damage to your backup disc, but only if you leave it unplugged while you're not actively using it. Handle the disc very gently, do not expose it to high levels of moisture, heat, or vibration. Hard discs should be kept away from strong magnetic fields. Other than actual magnets, sources of magnetic fields can include electric motors, soldering irons, CRT displays, and electrical transformers.

Online/Remote Backup

This has been an emerging trend in the last few years, with several companies offering free and paid online backups. The biggest advantage to this is that all of the data is stored offsite, so natural disasters, local hardware failures, fires etc... will not destroy your data. The biggest disadvantage is that you're entrusting your data to an outside entity who may or may not be reliable. When shopping for a remote backup service, there are several things you should consider. Where are their servers located? Are you comfortable with your data being stored overseas? What kind of hardware do they use? How regularly do they backup their hardware? How is the data transferred from your computer to their servers, and is it securely encrypted? Who will have access to your data once it's stored on their servers? Since their servers are connected to the internet, they should be rigorously secured, since the consequences of a malicious third party gaining access could be astronomical. Most importantly, will the company still be around a few years from now? Tech company buyouts happen constantly, so it would be best to stick with one of the larger companies like Microsoft, Amazon, or IBM that is unlikely to disappear overnight.

The most reliable companies will have privacy policies detailing who and under what circumstances a third party may access your data. Regardless, if you must use online backup, I suggest encrypting any sensitive data before you upload it to their servers. I would also suggest you avoid Google, since they have a very loose definition of privacy.

One last thing you may want to consider is that in the event of a natural disaster you may not have reliable internet access. Keep local copies of anything you might need in such a circumstance.


Powered by FreeBSD
Valid HTML 4.01
Site Map
©MMIX-MMXIV Planet Fox