My History With Linux

OK, time for a little back ground...

In the late 1980's I was looking for a UNIX operating system to run in my office. I worked with UNIX, wanted some type of UNIX, but couldn't afford the products that Sun and HP were selling. IBM was too new and other products were very niche. For a time, I owned various UNIX systems from Morrow Designs (Motorola 68000), Masscomp (Motorola 68000) and even Nabu (8086).

So, when MicroPort (long since defunct) released ATT UNIX System V Release 3 for the Intel 80306, my long standing love affair with UNIX on Intel CPUs commenced.

Over the years, I have run not only MicroPort UNIX (defunct) but also Bell Technologies UNIX (backed by Intel but now defunct), Everex's ESIX (defunct along with Everex), Dell UNIX (defunct - not Dell, just Dell UNIX), Interactive UNIX (defunct, bought by Sun and abandoned) and also SCO (nearly defunct). Note the common word I use to describe these products (defunct).

All of these products had the same flaws which were:

They were expensive (usually about $1,000 US per license or more)
Hardware support was limited (you had to shop for the right hardware or it wouldn't work
Vendor support was limited to non-existant.
They all gave me headaches at one point or another over 'license issues'.
They were all lacking in polish in some way.

In the early 1990's, I became aware of Linux and was intrigued. Here was a product that looked, worked, acted and otherwise promised to behave like UNIX - but had no license encumberances. Otherwise, it was just like the commercial UNIXs I was using (in that Linux had all of the same annoyances as the commercial products, but without the pricetag!).

So, I started purchasing various Linux releases including Slackware, Yaggdrasil, Walnut Creek, SuSE, Debian and RedHat. My early experiences were not encouraging... Installs were difficult and usually failed. Hardware support was very limited and documentation was almost non-existant. I got into the habbit of purchasing a new Linux release every 6 months or so and trying it out. Invariably, I would chuck it because it was just too much bother. That is until RedHat 4.0.

I installed RedHat 4.0 on a Intel 80486 system. Much to my pleasant suprise, the installer worked, it did what it was told, the software installed correctly and the system booted and worked - A first for me and Linux.

I played with the machine but abandoned the software because it wasn't as functional as my SCO OpenServer 5 powered system. However, when I got my hands on RedHat 5, the story changed. RedHat 5 did everything my OpenServer system did. It was just as fast, supported just as much hardware, installed and ran just as smoothly and, required no licenses. My RedHat 5 system took it's place beside my OpenServer system as a learning tool for Linux.

I upgraded to RedHat 6.0 as soon as it came out. Finally, Linux was superior to OpenServer 5. More modern GUI, built in firewall, capable software development tools, great device support and symmetric multi-processing - all free!

Over a few months, I gently migrated all of my software and services off of OpenServer 5 and onto RedHat Linux 6.0. At the end of the migration, I re-formatted my OpenServer system and put RedHat Linux 6.0 on that as well.

I have been happily running RedHat Linux ever since. I currently have servers running customized versions of RedHat 8, 7.3 and 7.2. I intend to install RedHat's Enterprise Server 2.1 in the near future.

Linux Distributions

Below are my thoughts on the various Linux distributors. This is just personal opinion (not fact), but should serve to illustrate where I think the Linux market is going.

There are anty number of Linux distributors; some with commercial releases, some with free releases. Linux distributors usually differentiat themselves along the following lines:

Technology Adoption
Some Linux distributors release software as fast as they can. They are at the leading edge of the technology wave. Their stuff isn't always the most reliable, but it is sexy and new. Others (usually business focused) adopt later on, do more extensive bug testing and as a result, provide more stable releases.
System Focus
Some Linux's focus more on the workstation, while others are more server oriented.
Open Source Advocate
Some Linux vendors are staunch open source advocates while others try to apply proprietary license models to software they didn't write.
Long Term Stability
It is hard to make money when you give your products away. This has caused a great deal of trouble in the Linux distribution business and has forced more than one distributor to actually beg for money. If you are going to adopt Linux, you want to pick a distributor who is in the business for the long haul.

Linux Vendor Matrix

Here's my opinion on how various Linux distributors fit when measured against the above criteria:

Distributor Technology
Adoption System
Focus Open Source
Advocate Long Term
Stability
RedHat Mid Level. Minimize risk while adding features. Desktop, Server and Enterprise (clustering) Strong on Desktop, Commercial on Server. Good due to strong commercial ties and market share in North America.

SuSE Mid Level. Likes to be on the leading edge of Linux for business. Has the broadest range of CPU support on same releae. Desktop, Server and Enterprise (clustering) Strong on Desktop, Commercial on Server. Good due to strong commercial ties and market share in Europe.

Debian Mid to Late Level. Desktop, Server. The Best! Totally Free. Available for download. Debian is the only large scale Linux release that is 100% free! It is unlikely that Debian will ever go bankrupt.
SCO Linux Mid to Late Level. Very conservative. Very dead. Don't look to ever see this one again. Desktop, Server and Enterprise (support). No free versions, no downloads available. Entry level price is $599US. In my opinion - Poor. SCO abandoned their previous Linux releases (eDeskTop and eServer) in favor of the United Linux. And, their market share is poor (single digits). No longer shipping and will never ship again. Look elsewhere.
Turbo Linux Mid Level. Minimize risk while adding features. Server and Enterprise (clustering) Commercial on Server. No opinion. Regional distributor that focuses on Asia. Limited presence in North America & Europe. UnitedLinux partner.
Conectiva Linux Mid Level. Minimize risk while adding features. Desktop, Server and Enterprise (clustering) Commercial on all platforms. No opinion. Regional distributor that focuses on South America. Limited presence in North America & Europe. UnitedLinux partner.
Mandrake Linux Leading Edge. Always the newest features Desktop, Server, Enterprise and appliances. Both commercial and open source. Always makes releases available via ftp. Strong player in Europe and North America. Great software with lots of polish. However, they stumbled in non-Linux related areas that have cost them financial stability. Not sure if they are solvent so I can't predict long term viability.

Distributor	Technology Adoption	System Focus	Open Source Advocate	Long Term Stability
RedHat	Mid Level. Minimize risk while adding features.	Desktop, Server and Enterprise (clustering)	Strong on Desktop, Commercial on Server.	Good due to strong commercial ties and market share in North America.
SuSE	Mid Level. Likes to be on the leading edge of Linux for business. Has the broadest range of CPU support on same releae.	Desktop, Server and Enterprise (clustering)	Strong on Desktop, Commercial on Server.	Good due to strong commercial ties and market share in Europe.
Debian	Mid to Late Level.	Desktop, Server.	The Best! Totally Free. Available for download.	Debian is the only large scale Linux release that is 100% free! It is unlikely that Debian will ever go bankrupt.
SCO Linux	Mid to Late Level. Very conservative. Very dead. Don't look to ever see this one again.	Desktop, Server and Enterprise (support).	No free versions, no downloads available. Entry level price is $599US.	In my opinion - Poor. SCO abandoned their previous Linux releases (eDeskTop and eServer) in favor of the United Linux. And, their market share is poor (single digits). No longer shipping and will never ship again. Look elsewhere.
Turbo Linux	Mid Level. Minimize risk while adding features.	Server and Enterprise (clustering)	Commercial on Server.	No opinion. Regional distributor that focuses on Asia. Limited presence in North America & Europe. UnitedLinux partner.
Conectiva Linux	Mid Level. Minimize risk while adding features.	Desktop, Server and Enterprise (clustering)	Commercial on all platforms.	No opinion. Regional distributor that focuses on South America. Limited presence in North America & Europe. UnitedLinux partner.
Mandrake Linux	Leading Edge. Always the newest features	Desktop, Server, Enterprise and appliances.	Both commercial and open source. Always makes releases available via ftp.	Strong player in Europe and North America. Great software with lots of polish. However, they stumbled in non-Linux related areas that have cost them financial stability. Not sure if they are solvent so I can't predict long term viability.

The Winner Is?

My focus is Linux business solutions. As a result, I am looking for a well regarded Linux distributor who offers everything from free (so I can play) to commercial solutions that fit the needs of business (support, upgrades, updates, clustering, etc.). I also look at market share and the long term prospects of the vendor.

It is not clear that any vendor is truly safe. Most distributors are subject to the whims of the financial markets and are profit seeking companies. Their fortunes may wax and wane as third party vendors choose to to give or withdraw support. Even totally free free Linux's (Debian) may be subject to financial pressures (as the BSD folks recently found when government funding for one of the BSD projects was pulled).

For me, the choice is down to one of two distrubutors; RedHat if you live in North America or SuSE if you live in Europe. If you don't like my selection criteria, live anywhere else, don't speak English, want leading edge or have other selection criteria, you should examine your local market and pick accordingly.