NT and Unix Server Siege

Keith Dawson (dawson@sitaranetworks.com)
Mon, 17 Nov 1997 14:46:10 -0500

> Your growth figures for NT versus Unix are misleading in
> that Unix servers generally scale to more users thus they require
> fewer licenses purchased. Add on top of this the number of
> 64-bit unix licenses versus the number of NT 64-bit licenses
> and your numbers just don't mean all that much. Is there any
> way you can correct for these confounds?

It's even worse than you imply, as I found out after publishing similar
figures early this year (see http://www.tbtf.com/archive/01-21-97.html).
Such comparisons ignore the proportion of "NT" or "Win95" boxes that
upon delivery are immediately converted to Linux, FreeBSD, etc. I'm
perfectly willing to believe that the number exceeded 58K, which would
mean that NT did *not* outsell Unix in 1996.

I'll echo Greg's question for Tom: do you have any way to find figures
for the free-Unix-on-Intel phenomenon?
Keith Dawson dawson@world.std.com
Layer of ash separates morning and evening milk.

PGP id.1FDA25D1 38 C7 13 05 F3 EF C9 EF F3 87 15 27 65 BF F4 4C

= TBTF alerts you weekly to bellwethers in computer and communi-
= cations technology with special attention to commerce on the
= Internet. Archive at <http://www.tbtf.com/>.