On Fri, Aug 09, 2002 at 04:47:59PM -0600, Donaldson, Mark wrote:
> 
> Caveats: this was for NB client compression, not hardware compression at the
> tape drive.  I'm assuming that the algorithms are close enough to generalize
> to overall compressibility.

I have no hard numbers with which to challenge this assumption, but I
suspect it's not so simple.  Different tape drives use different
compression algorithms (NBU uses Lempel-Ziv, a la unix `compress`)
and I bet you might find some statistically significant differences if
you ran the above tests and used the HW compression.

Interesting info, nonetheless.  Thanks!

rob