Which tape backup to buy?

Posted by: TigerJimmy

Which tape backup to buy? - 19/09/2006 09:32

Hi everyone,

A friend of mine has is own law firm and I've set him up with a network file server running OpenBSD. He is moving to a "paperless office", with everything scanned into searchable PDF, with hardcopies stored offsite as legally required but not in the office. He has been making backups to DVD-ROM, but expects to have significantly more data as a result of this big scanning process.

I have recommended tape as a good long-term backup solution, with 2 weeks of daily tapes in rotation and semi-monthly or monthly archive tape for offsite vault storage.

The last time I bought a tape drive for an office, DAT was still the answer. I'm not familiar with the current hardware/tape combinations and am hoping some of you can offer your opinions of the current tape solutions.

I'm expecting he will have more than 100GB but less than 300GB of data, and (as law firms generate a ridiculous amount of documents) that he will need some room to expand over the next 2-3 years.

The friend is a technophile, and understands that any computer infrastructure is a temporary solution. He would be fine with 2-3 years of use out of the system.

I'd prefer to have UNIX dump the whole filesystem to a single tape unattended overnight. Then all his paralegal (who is not tech savvy) will need to do is change the tapes in an orderly manner.

Thanks in advance for your guidance!

Jim
Posted by: furtive

Re: Which tape backup to buy? - 19/09/2006 09:56

I would recommend getting a NAS rather than a tape backup - much more reliable and faster.
Posted by: TigerJimmy

Re: Which tape backup to buy? - 19/09/2006 10:07

I agree, but not good for long-term archiving of data that the law business requires (unless you want to buy a new NAS every month).

Actually, the plan is to have a redundant server offsite using rsync. The archive backup is a whole different beast, related to never having anything thrown away.
Posted by: tahir

Re: Which tape backup to buy? - 19/09/2006 10:27

We recently bought a Sony AIT Turbo something, quite happy with it
Posted by: BAKup

Re: Which tape backup to buy? - 19/09/2006 11:13

I was going to recomend NAS as well, but since that's out, I'd go DLT, much more expensive than DAT, but has much greater storage per tape (160/320), and much faster on the backups, and through my trials of dealing with tape backups, the DLT tapes are more reliable.

http://www.quantum.com/Products/TapeDrives/DLT/DLTV4/Index.aspx
Posted by: caseyse

Re: Which tape backup to buy? - 19/09/2006 13:14

Have him take a look at LTO. I purchased an LTO-3 off of eBay, and have been very happy with it.
Posted by: tfabris

Re: Which tape backup to buy? - 19/09/2006 15:40

Quote:
The last time I bought a tape drive for an office, DAT was still the answer.

Eeeeeew. Yuck. Those were so unreliable.

I swear by SuperDLT technology myself. Not cheap, but ultra-reliable.
Posted by: drakino

Re: Which tape backup to buy? - 19/09/2006 16:20

If the price of the higher end DLT is a problem, consider the DLT VS160 (80/160gb) or DLT V4 (160/320gb). Those two drives should be cheeper then a full SuperDLT tape system, but is still built on the DLT tech.

LTO is also nice, but it usually is even more expensive then DLT. Reason being, LTO is a similar tech to DLT, but adds a RFID chip in the tape to store the catalog, allowing for faster file access. AIT does the same thing, where it adds a similar chip to DAT tapes.
Posted by: wfaulk

Re: Which tape backup to buy? - 19/09/2006 16:27

Actually, the tape tech for AIT is much more similar to 8mm tape, kinda like Hi8 camcorder tapes.
Posted by: lectric

Re: Which tape backup to buy? - 19/09/2006 22:13

Exactly how much data storage per month are we talking about? At my work, we have a paperless system tha thas been in place since 1995. We scan every piece of paper that goes through finance. We have 1 full time and 2 part time scanner clerks. They do nothing all day but scan. After 11 years, we're just pushing past the 60Gb mark. That's without ever purging anything from the system., even though we can, by law, drop everything over 3 years old not relating to personnel records and not involved in an active lawsuit.

My backup regime is simple, On the first of every month, I burn to DVD the last 3 month's images and do a full dump of the database. This way, I always have 3 copies of any particular image. All stored on different parts of the respective disks. Since it's running on a linux server, we have a clever little cron job that does all this automatically at noon on the first of every month. I just have to remember once a month to pop in a new blank DVD. It's kinda hard to forget since the tray with the freshly burned DVD is open the next time I walk into the server room.
Posted by: Dignan

Re: Which tape backup to buy? - 20/09/2006 09:46

That system sounds good, but I'm guessing you don't back up any email. That's what takes up most of my backups.
Posted by: TigerJimmy

Re: Which tape backup to buy? - 20/09/2006 19:16

Thanks, everyone this is really good information.

I don't know how much data we're going to be talking about. We did a few tests and the files are fairly large. He is scanning into PDF, which seems to make a bigger file than printing to a PDF does. How are your scanned images stored to get all this in 60GB? He wants the documents to be searchable, hence the PDF format.

Thanks again,

Jim
Posted by: Ezekiel

Re: Which tape backup to buy? - 20/09/2006 19:27

Scanned PDF's will not be searchable unless they are put through an OCR engine first - otherwise they're just jpg compressed graphics files (or .bmp depending on your pdf settings) wrapped inside a .pdf file. 'Printed' .pdf's consist of text data, embedded fonts embedded pixel graphics, and embedded vector graphics depending on the source application.

Since text, fonts & vectors are 'small' data compared to pixel graphics, the resulting file will be smaller than a purely pixel graphic of a scanned .pdf - even if they look identical to the eye.

OCR would replace the pixel information with text information, creating searchable text and shrinking your file size. No OCR on scanned documents = no search capability.

-Zeke
Posted by: TigerJimmy

Re: Which tape backup to buy? - 20/09/2006 19:34

Yeah, he bought a big expensive scanner and runs everything through OCR on a PC. I think I need to do a better estimate of the data size requirement before I can make this decision.
Posted by: Roger

Re: Which tape backup to buy? - 20/09/2006 19:42

Quote:
OCR would replace the pixel information with text information, creating searchable text and shrinking your file size.


Not necessarily -- I've seen OCR solutions where the original bitmap is preserved alongside the recognised text. This makes it searchable, but also means that you can read the text from the original image if it didn't recognise something quite right.

I don't know whether the system we're talking about here does this, though.
Posted by: msaeger

Re: Which tape backup to buy? - 20/09/2006 20:22

You could change the resolution to lower the file size too.
Posted by: lectric

Re: Which tape backup to buy? - 20/09/2006 21:01

In response, we're saving 200 DPI black and white tif's. With zip compression on. I'm thinking the biggest size-saver is BW vs Color. At 200 DPI, I couldn't tell the difference from a 600 DPI when printed, so it is more than enough for my needs. Your mileage may vary.