Hardware or software compression with Tivoli Storage Manager

Answer

Hardware and software compression have varying features and limitations which should be considered prior to implementation and usage in an active production environment. These pros and cons are generally outlined as indicated below:

Hardware compression:
Many tape drives include fast and convenient hardware based compression. Tape drive manufacturers, reasoning that tape is used primarily for backup, have found that hardware compression is a great way to efficiently pack data onto tape. Users like compression because it saves money and backup time. Tape manufacturers have usually identified two capacities for their tape products: the actual capacity (also called native capacity), and a second figure called compressed capacity.

The issue with hardware compression is that hardware compression is usually proprietary. Even if a tape compression standard is open, the hardware compression method used by one tape drive manufacturer may be completely incompatible with that of another. Therefore, tapes compressed on one drive may be indecipherable on another drive even though they may use the same form factor.

Software Compression:
Software compression is useful in that the compression is always the same regardless of which tape technology is used and does not vary between varying types of hardware involved. The operating system (example is UNIX) or software (Tivoli Storage Manager) performs compression on data prior to sending it to the hardware, which makes compression operating system or application dependent. Because software compression is relatively standard and can be coded into updated software, data compressed by one operating system or application can usually be decompressed by other operating systems or applications.

Software compression however is not as efficient as hardware compression, and performance overhead from software compression will be observed during operations. This will impact the performance of all server operations since the compression activity is performed at a software layer rather than at a hardware layer.

Independent Compression Considerations:
It is important to note that not all files compress equally well. This is an issue of special importance to users of imaging systems where scanned images are already in compressed image file formats like JPEG, GIF, Group3 and Group4. These types of files among others don't compress well at either a hardware or a software level, and can actually grow in size when compression is attempted (results as a negative compression value). This type of behavior can be seen by running sample files of the above types through a software compression utility like Winzip or Stuffit. In these situations, compressing a file that is already compressed is considered "double compression". Other file types such as database or text files compress much better.