Archivering data - more than backup

Some documents must be stored for a long time. The (Dutch) law has no specific regulations about how long data must be stored in archives, but it requires companies to state how long their data must be kept available.

This means that every company uses different standards for archiving data, and the type of information might also be different. For instance, the Dutch tax agency forces companies to keep tax records for at least 7 years. But the retention time of sales records might be 2 years, depending on the business. There are however circumstances where data must be kept much longer, sometimes even more than a human lifetime.

Some examples are:

• (Pension)Insurance companies must keep records of their history of people and claims;
• Hospitals store medical information during the lifetime of a patient;
• The Justice department keeps records of crimes (specially if they are not solved) for a long time;
• Newspaper archives, archives of television networks and government (like image- and sound archives).

Digital format

Data must be kept in such a way that it is certain the data can be read after a long time. This means the digital format (like a Word file or a JPG file), the physical format (like a CD or a magnetic tape) and the storage environment (temperature, humidity) must be such, that data can still be retrieved after several decades.

This is not a simple task.

Let's look back a few decades. In these years mainframes were common in the IT industry. Data was stored on reel-tapes, that are not readable anymore (using 8 or 9 parallel tracks, usually in EBCDIC format instead of ASCII). Data was also kept in propriety formats, like mainframe database tables.

Can you guarantee this data can be read and interpreted today?

After mainframes, PC's became the norm, using applications like Wordstar, Lotus 1-2-3, WordPerfect, MS-Word (how many versions?), Lotus Notes, PDF, etc. Furthermore, there are now audio- and video formats, like BMP, GIF, TIFF, MP3, WAV, MPEG (several versions). The list is too long to comprehend, and all of this was developed in the last 30 years.

Here is an interesting list with file formats and their specifications (47 pages long). It ill not be easy to make sure all these file formats can be read in 30 years time. And probably more file formats will be developed in the years to come.

I recommend to use open standards for storing data as much as possible. These standards should be not too complex, like the ODF format. The ODF format is a zipped file with XML text files in ASCII, that will be readable for a long time, because it's format is well described and not too complicated.

Physical format

Many physical storage format exists. Except the previously mentioned reel-tapes, these are also tape cartridges in many formats, DLT tapes, SDLT tapes, LTO tapes, DDS tapes, CD, DVD, Floppies (3,5", 5 1/4", 8"), etc.

Many archiving technology is storing data in optical formats. While this is much better than magnetic storage on disk or (even worse) on tape, is it not obvious that media like CD's or DVD's are still readable in many years time. How long will it take for Blu-ray- or HD-DVD's to be common media? And what will follow?

Therefore, it is advisable to transfer data that is to be kept for a long time to the latest storage media standard every 10 years (from Floppy to CD's) or to at least move the data to a new copy (burn a 10 year old CD on a new CD).

This entry was posted on Donderdag 05 Juli 2007