Someone please explain compression!!!

[weescotty]

OK have an 80/160 tape so I assumed it was a 80Gb native/160Gb compressed.

Here's what I dont get - just backed up 167Gb to it and it says these 90Mb left. Tried restoring files from the start and the end of the sessions, both are there.

So I take it the 160Gb compressed could mean that say if 200Gb of actual data compressed to under 160Gb it would fit, or am I missing something?

Thanks

 

[jfrantz]

The compression ratio is an estimate. The ability to compress data varies by the type of data. Text compresses very well while say MPEG-2 video is already comressed and probably will not compress further. The manufacturer of your tape drive assumes that the average amount of compression is two fold.

For instance, I have a database which takes up over 7GB of disk space. The database is almost all text so it compresses great. I get an average 5.3x compression on this database when it is backed up. If I only backed up this data to your tape drive you could effectively say that the 80GB tape drive is capable of storing 424GB of my database.

If I were backing up compressed video to your tape drive, most likely I would only fit about 80GB onto the tape.

Hope that helps!
-Jeff

 

[weescotty]

Thanks.
If I understand correctly - You can't rely on the Mb shown processed by Arcserve as you dont know how well it is being compressed. So it's entirely possible for it to show say 250Gb processed and still be copying to a 80/160 tape as that 250Gb may have compressed to a lot less?

Thanks

 

[vschumpy]

This is a handy explanation in general about hardware compression :

http://supportconnect.ca.com/sc/redir.jsp?reqPage=search&searchID=QI63222

Media remaining capacity is actually a feature of the tape drive and not something which ARCserve calculates - hence why you'll see it for backups on some drives and not others.

 

[davidmichel]

and by golly right here in the faqs for this forum is:
How does ARCserve control compression? faq478-4399

 

[aistswc]

Scotty - u r correct n your analysis. Compressed data on a tape is th actual number of bytes that the original data compresses down to. Think of it as taking a big cloud, and compressing it into it's final size of a gallon jug. There be a lot of air there.

Note that not all data compresses equally - some data will actually be a BIT larger AFTER compression. Others may know better on rules of thumb for "average" compression ratio. Seismic data does not compress at all. Word docs w/o graphics generally do pretty well. You can get a good feel for compression by running zip (or UNIX utilities if you r on a UNIX O/S) on your data.

Your mileage may vary.

steve c.

steve c.
Systems Analyst
Superior Court, USVI