Search the Knowledge Base:

 

Compression FAQ
(KB # 57645742)

What is the difference between hardware and software compression?

Hardware compression, in Tandberg Data tape drives, is done inside the tape drive by a chip that is designed to do lossless data compression and nothing else. All of the compression labor is done by that chip. The compressed data is then passed back to the drive for writing to tape.

Software compression is performed on the server side before the data is sent to the drive. It requires CPU time and memory resources from the server. Once the software has finished, the compressed data is passed over the bus to the drive where it is written to tape. The compression may be lossless or lossy depending on the software’s algorithm.


Should I use software compression, hardware compression, both or none?

It is important to understand that some files (such as text, html and xml files) generally compress very well, while other data hardly compresses at all. It is also important to note that compressed data rarely compresses at all, and attempting to compress data which is already compressed (including data in a format that is compressed by definition such as MPEG and ZIP files) can actually cause the file size to grow when put to tape. For this reason, it is recommended that either hardware compression or software compression be selected, not both. Generally speaking, hardware compression works more efficiently than software compression, and should be used instead of software compression. However, you should use whatever you have found works best for your servers. Some servers work best using no compression at all.

Unless your company has a corporate standard that specifies how you must do backups, the best way to find out what works best is to do a few backups using hardware compression, then software compression and finally no compression. Make sure that your server is not being stressed out during the software compression backups. Look at the number of tapes being used and the time it took to do the backup. You will need to evaluate these and determine what will work best for your backup needs.


How can I tell if my drive is using hardware compression?

Hardware compression is on by default for all Tandberg Data drives. However, if you choose, through your software, to not use hardware compression, a SCSI Mode Select command is sent to the drive to turn it off temporarily.

Drives that have an LCD panel (including drives in libraries with an LCD) and VXA-320 drives have methods to display their hardware compression status on the front panel while doing a backup. Please refer to the manual for the drive or library type you have to see how it reports the use of compression during a write.

VXA-1 and VXA-2 drives will require that the tape be left in the drive after a backup and a log file gather from the drive using the VXATool. A Technical Support Engineer will be able to look at this log and tell you if compression is on.


Why don't I get 2:1 compression?

Quoting native capacity and twice the native capacity (i.e. 80 Gig / 160 Gig) of a piece of media is a tape industry standard. It is intended to give the user a potential capacity range for that piece of media.


Many things affect compression. Here are some of the major contributors:


Data being backed up:As explained above, different types of data compress to different ratios. Most data that is being backed up is mixed and therefore the ratio will be an average of the compression ratios of all of the files being backed up. Smaller files also don’t compress as well.

Throughput: If your throughput drops below the native throughput for the drive, it will write "gap" file markers to the tape. Most drives will, after a certain time, perform a backhitch. This is when the tape drive stops, rewinds a little and when the buffer is full again, reads to find where it left off and begins writing again. When this occurs over and over, the capacity of the media is reduced as well as the life expectancy due to the increased number of passes on the tape.

Aging media: As media is used more and wears, there will be more re-writes performed by the drive to compensate for spots that are wearing.

Debris: Dirty heads can cause the drive to attempt to rewrite the data further down tape.This effectively doubles the amount of tape used for that block of data. Clean the drive per the recommendations in the drive's manual and replace the cleaning media as required.


Also see:

KB 794 - How can I increase the system transfer rate to my tape drive, autoloader or library?



Rate this

Send Feedback

Send Us Feedback

Questions, comments, suggestions? Use our
feedback form and let us know what you think.

Feedback Form