software vs hardware compression

[heather]

Currently I'm using Hardware Compression with the following results:

A full backup with hardware compression takes 16 hrs and fills up 4 AIT tapes as follows:

 

tape 1 45.2 GB

tape 2 53.8 GB

tape 3 49.4 GB

tape 4 27.9 GB

 

(taken from the Backup set members tab)

 

I'm going to create new backup sets and enable SimpleCrypt.

 

If I enable software compression, how much slower is the backup going to be and will all the data fit on the currently available 4 tapes?

 

I'm running Retrospect 5.0.238 on Mac OS X , 10.2.4

 

The Autoloader tape drive is a Sony:

TSL-A300C

 

The spec for the AIT tapes is:

SDX1-35C

AIT Data Cartridge

35GB native and 70GB compressed, 230 m.

 

Thanks for your help,

 

HS

 

[mcswgn]

It's really impossible to say how much slower it will be, since that depends on too many variables. Certainly the faster the backup computer, the less of a hit it will be. If right now your system is having no trouble keeping the tape drive's buffer full and has a lot of idle time on top, then it may not affect it much. (Although, with an AIT drive, that's probably not the case. You could look at process viewer next time a backup is writing to tape and see what it says about CPU demand.)

 

As fas as tape capacity is concerned, I would guess (and a completely uneducated one :-) that you probably wouldn't see a dramatic difference. There are two issues. 1) How efficient is Dantz's software compression vs the built-in hardware compression? Certainly there will be a difference, but compression algorithms have been around for a long time. They are still being improved, but the marginal improvements are less and less. Therefore, I wouldn't expect the software compression to give radically different compression ratios than you are seeing now.

 

Fortunately Dantz compresses first before encrypting (as you can't compress encrypted files). So the second issue is: 2) By how much does encryption increase the size of the file? I've never tried to check this for their SimpleCrypt. All encryption methods will add some overhead, however, but I would hope that they chose this method so that at least for larger files it can do it fairly efficiently.

[heather]

I tried it and ran out of tape space.

I was short 44 GB

Each tape filled to 37.2.

Will finish the backup tonight, but am trying an experiment, one with software compression and one without software encrytpion.

 

Heather.

[Drishmung]

Which reminds me...

 

Does AltiVec make a significant difference in compression? (I would think it would, since MP3, MPEG & JPEG all gain large speed increases from AltiVec on the G4s).

 

And if so, does Retrospect make use of it? Anyone know?

 

[heather]

Ran part one of the test to try and figure out how tape capacity is affected by (1) hardware compression and (2) software compression. For Hardware compression I got

 

44.9 GB on tape 1

20.9 GB on tape 2

Time was 5 hours

 

Will run software compression test tonight.

 

If you password protect the tapes and save the password for scripted access, does that prevent someone from loading the tapes on a different machine and getting the data? In this scenario, the password is saved in MY retrospect config file, so I know that if the tapes are loaded on a different machine the password will not be provided by Retrospect for any operations.

 

If a hard drive is protected by a password you can't get into the hard drive unless you have the password. If a tape is password protected does it have the same behaviour? i.e. no one can access the data without the password?

 

Heather.