RedHat vs. the rest of Linux

rwotten@checkfree.com rwotten at checkfree.com
Fri Jun 14 15:51:00 CDT 2002


This is an interesting thread that I've been following pretty closely
because I've been thinking about similar issues for some time now.

Here is my take on it -

  There is a lot of hype and expectations regarding Linux amongst the
senior (non-technical) IT management in many of the Fortune 500's right
now.  They are under pressure to reduce infrastructure costs due to very
lean economic times.  There is a pervasive perception (and an incorrect one
IMHO) that the TCO (total cost of ownership) is much lower for Linux than
for pretty much any other Operating System.

   However officially adopting Linux is a scary thing.  It is a new type of
platform requiring staff training, new tools, and many unknowns.  One
(perceived) way to reduce risk is to try to choose a version of Linux "that
they've heard of".
   It is that simple.

   9 out of 10 senior IT managers have heard of Red Hat, and I'm willing to
bet, maybe only 1 out of 10 could name any other 'brand' of Linux.

    The second perceived way to reduce risk is to have some form of
official vendor support.  Red Hat and Dell offer a nice integrated support
package which is designed to give senior managers warm fuzzies.  (How
effective that support really is, is irrelevant.)

    Unfortunately for us in the field, Red Hat may not be the best version
of Linux for every application, every server, and every situation.  The
thing that annoys me most about Red Hat is their tendency to branch from
the rest of the industry on critical components - such as gcc.
    (Red Hat has their own gcc branch which is not endorsed by the public
gcc development team.  The first thing I have to do on _every_ RH server
that I set up is replace gcc with a universally accepted version.)

   Red Hat also is heavily bloated, overloaded with junk, and somewhat
leaning towards the desktop.
    It is a bit of work to go through and clean up the packages you don't
need and will never need and lock down a server that starts with a base Red
Hat install.
     Personally, when I'm building a server, I'd rather start with a bare
bones system, then install what I need, rather than start with a bloated,
overloaded server and remove what I don't need.

     United Linux doesn't make any sense to me.
     If you want to compete, you've got to have differentiators from your
competition.  If you and most of your competition have an identical
product, you've got many fewer options to differentiate yourself.  This
makes it harder to compete.
     Additionally I don't know how they will ever come to an agreement on
what 'United Linux' really is.

     I think Linux is a powerful, flexible, cutting (bleeding!) edge tool.
I think there is a lot lacking when it comes to the kind of maturity and
product cohesion the Fortune 500's have come to expect and demand.  It'll
get there.  Sun's hybrid Solaris/Linux may be a step in that direction.
IBM's forays into the Linux marketplace may help us get there.  And Red Hat
is helping us get there.

   Suse and others really need to build some brand name recognition in the
non-technical manager minds.  They then need to bring their product (and
service) maturity to the level, not just of Red Hat, but to that of
Solaris, HP/UX, IBM AIX, and the other big boys,  in order to be a viable
(and visable) option in the Fortune 500 mission critical (revenue
impacting) infrastructure.

--
Rick Otten
O=='=+








More information about the Linux-PowerEdge mailing list