| |
| |
| Tech Note 004 |
| Title: Drive
compression and tape capacity |
| Updated: April
2005 |
|
| |
What is compression?
When data is written to a tape device it can be compressed, so that it
takes up less room on the tape. This also carries the benefit of speeding
up the data transfer - the greater the amount of compression, the faster
the data transfer speed.
There are two basic forms of compression: software compression and hardware
compression.
Software compression is performed by the software that is controlling
the backup; the data is compressed by the software and then the compressed
data is passed to the backup device. Software compression can sometimes
be 'lossy', which means that it isn't actually writing the correct amount
of data. FlashNet does not use software compression.
Hardware compression is performed by the backup device. Most modern tape
drives are able to compress the data as it is written to the tape. Hardware
compression is 'lossless'; the compressed data is completely intact at
write time.
|
Native and compressed
data speed and capacity
When a drive writes uncompressed data, the transfer is known as 'native'.
This means that data is not being compressed, and the drive is writing at
its native speed. The amount of uncompressed data that will fit onto a single
cartridge is known as the 'native' capacity of the cartridge.
When writing compressed data, the drive is able to fit more data onto a
cartridge, and to transfer the data at a greater speed. When quoting speed
and capacity, most drive manufacturers use a compression ratio, often of
2 or 2.6:1. When purchasing new drive technology, it is always a good idea
to find out the native speed and capacity of the device as well as the compressed
data capabilities. This is because drives compress data at different ratios;
the guideline speed and capacity quoted by the manufacturer is for sustained
transfer at the quoted ratio, but in the majority of write operations this
ratio varies quite drastically from file to file. |
Not all data
can be compressed
The amount of compression that a drive can obtain (and therefore the speed
and capacity that it can write) is completely dependent on the type of
data that it is writing. Data compression is performed on a file-by-file
basis, and the type and size of an individual file defines how far it
can be compressed, and the speed at which the drive will write it. Some
file types, for example jpeg and mpeg files, are already compressed, and
the drive is unable to compress them much further. This means that a data
set comprising mostly jpeg images will not be compressed very much (usually
a fair amount less than the 2:1 ratio quoted by drive manufacturers),
and the amount of data that fits onto a cartridge is not much more than
the native capacity. Other file types may be highly-compressible; text-only
files, for example, can be compressed at very high ratios. |
How does FlashNet
set drive compression?
By default most drives always set themselves to the highest possible compression
level at write time. In addition, FlashNet instructs most types of drive
to set themselves to the highest level that can be achieved when they first
write to a volume.
Most drives have dip switches that can be used to enable or disable specific
functions. If you disable drive compression using the dip switch on a drive,
this will override any other settings and the drive will write in uncompressed
mode. |
| |
| |
| |
==END==
>>
top |
