On Jan 10, 10:03 am, Roy Schestowitz <newsgro...@xxxxxxxxxxxxxxx>
wrote:
> The "Year of Linux" Through Multiple Advents
> ,----[ Quote ]
> | It's like the saying, "The faster the rise, the faster the fall", and that's
> | true even in the PC world. Therefore, just because Linux doesn't suddenly
> | explode out of the stocks doesn't mean it's a failure.
To really understand the growth of Linux, you need to look at the
growth of UNIX in the server market. UNIX was an operating system
that turned a relatively basic machine into a multiuser interactive
system with powerful applications and powerful intercommunication.
In the Minicomputer market, DEC, Sperry, and Harris had proprietary
operating systems that worked with their proprietary hardware. DEC
had RT-11, RSTS, and VMS. UNIX had been ported to the VAX in 1983,
and it sold a lot of VAXen, but DEC wanted VMS on these systems
because there was more profit. Salesmen were given much larger
commissions on VAX systems sold with VMS than with UNIX.
Eventually, DEC crossed the line, sending "support staff" to apply
"support patches". If the machine was running UNIX, they would wipe
the hard drive, install VMS, and apply the patch. When you have 100
employees, each getting about $20/hour (1983 dollars) and they are
sitting there doing nothing for 2 days while UNIX gets reinstalled and
reconfigured, it doesn't win you the loyalty of customers. In the
company I worked for, we hired an armed guard, and gave him orders to
shoot the rep in the knees if he attempted to enter the building
without an escort.
Ironically, DEC's arrogance triggered interest in other UNIX platforms
including those offered by companies like Zilog, and a number of
Minicomputers based on the 68000 family chips, such as the 68020.
Eventually, UNIX became the "platform of choice" and most of the major
minicomputer manufacturers, including Sperry, Harris, and many others,
started aggressively marketing UNIX, because it opened up markets that
would have taken years to crack with proprietary systems.
In 1990, IBM offered UNIX, known as AIX, but it was often a "Bait and
Switch" product. Customers would invite IBM to tell them about UNIX
and TCP/IP, and IBM would spend 15 minutes addressing that, then
explain how AS/400, MVS, and SNA were so much better and trying to
sell their proprietary product instead of UNIX. Again, commissions,
incentives, and motivation was all oriented toward pushing proprietary
hardware instead of UNIX.
In 1991, IBM released MVS 4.0, but it needed hardware upgrades, and
all of the application software had to be upgraded. If you purchased
the computer new in 1988, it cost $4 million. The hardware and
software upgrades for MVS 4.0 in 1991 (almost 1992) was almost $5
million. Many companies had 6-8 3090 mainframes, and a $30-$40
million bill tended to raise red flags when the money was requested.
By 1991, UNIX was becoming a popular database engine, and Oracle was
improving it's customer support for UNIX users. In addition, Sun,
Pyramid, and Sequent were offering SMP systems with as many as 12
processors per server. A 6 processor Sun/6 ran about $50,000 as did
the 4 processor Sequent. The Pyramid had 12 processors per server and
cost $250,000, but this included Oracle database and consulting to
tune and optimize the database to the application. To a UNIX user,
$250,000 was a lot of money, but to a VP considering $5 million per
server for MVS upgrades, even if it took 4-6 pyramids to replace one
of those mainframs, they were still going to save the company as much
as $4 million. And the $50,000 Sun, with Oracle, was still less than
the company paid for a month of support for the MVS mainframe. In the
last quarter of 1991, IBM stock took a nose dive, and Sun and Oracle
started shooting up.
Sun also had desktop offerings, and these workstations could be
combined with servers to act like a mainframe. Their slogan was "The
mainframe is the Network". Many executives really liked the
SPArCStations, and wanted one on their desks. People working in the
financial markets also liked the Sun workstations, as well as other
power users. Compared a SparcStation, Windows looked like a primitive
toy. It was much like comparing Windows XP (Sun) to Windows 3.0. It
didn't take much to see that Sun's desktop was more powerful, more
reliable, more secure, and more functional, and more flexible.
At the CES of 1992, Sun was gathering huge crowds and customers loved
the Sun workstations. Bill sent spies to see what people liked about
the Sun workstations and summarize the hottest issues. Bill gave the
keynote speech and announced Windows "New Technology" and promised
that it would be "A Better UNIX than UNIX". Microsoft had been
working on NT, but NONE of the features promised at the keynote had
been included in current design documents. Microsoft would not
actually deliver something "almost as good" as those 1990
SparcStations until the release of Windows XP with Service pack 2 in
2003. The problem is that UNIX and Linux were not standing still.
Linux had Internet support in 1983, and by early 1994, included both
Web Browsers and Web Servers. Linux also had all of the other
software required to be an internet Server as well as being a desktop
machine. Ironically, Microsoft assumed the Internet was just a
"passing fad" and didn't persue it until very late.
In late 1994, Linux had Plug-and-Play and Windows was still a bear to
configure. Even Windows NT 3.1 was really ugly to configure. Bill
saw Linux with Plug-and-play and decided that Microsoft not only had
to have it, but they had to make sure that LInux couldn't use it.
They developed plug-n-play which was appropriately named. Microsoft
kept a very tightly controlled list of PCI vendor and device codes,
and required that vendors not disclose these values to anyone.
Linux had support for USB and FireWire in 1997, but again Microsoft
decided to hijack the USB standard, and again used tightly controlled
codes protected by nondisclosure agreements. In both the PCI and USB
cases, Linux did eventually "crack" the codes, primarily because
Microsoft had violated contracts with other hardware vendors and the
vendors decided that rather than sue Microsoft, they would give the
codes to companies like Red Hat, Caldera, and other Linux
distributors.
Linux had support for OpenGL in 1998, which enabled Linux users to
have real 3D graphics using the Mesa API set. The SGI Virtual Reality
Modelling Language (VRML) was a popular scripting language for
creating 3D "scenes" and animations. Ironically, graphics hardware
couldn't keep up with LInux, so performance wasn't good enough to make
it a "Hot Seller". Again, Microsoft decided to make 3D graphics a
strategic element of Windows XP, and developed the DirectX graphics
API standard, to keep Linux off the market.
Linux often showed up as a "server" because 3-4 people could share a
PC that was running Linux and each of them could have their own
desktop interface using X11 or VNC. The reality was that many Linux
servers were actually functioning as 3-4 desktop machines.
Linux found ways to "play nice" with Microsoft, offering such options
to OEMs as "Dual Boot", Live-CD, Live-DVD, Linux emulation on Windows
(cygwin), and desktop virtualization. Linux even found ways to let
users run Windows as a Linux application, and get better performance
than they got from Native Windows (due to improved disk buffering,
caching, and 64 bit system calls.
> | But either way, so
> | long as Linux continues to put one foot in front of the other, every day and
> | every year can be a small integral part of a perpetual "Year of Linux."
One of the things that makes Linux hard to measure is that the market
shifts. For example, Novell's Linux related revenue grew 60%, but
that's less than 5% of the total number of SUSE systems, including
OpenSUSE. And thanks to the deal with Microsoft, Novell's share of
the desktop dropped to 20 % while Ubuntu shot to almost 30%.
( http://www.desktoplinux.com/news/NS8454912761.html )
> http://www.raiden.net/?cat=2&aid=362
|
|
