My thought is that Sun Microsystems was destroyed because they stuck to channel partners when everything was switching to generic PC's and Linux. If you use channels there's something like 100% margin, and who can carry that these days?
59 replies (most recent on top)
In 1999, workstation sales didn't earn the money Sun needed to pay its rent. That money came from high-end, high-margin systems.
Linux was no match in the high end. But the writing was already on the wall that IBM and SGI were going to change that.
Low end systems mattered more as a gateway drug to get people hooked on Solaris before they had any need for high end systems.
@OWQEY7M-1dul - indeed ... the SPARC mafia.
BTW, that's also who purposely killed Cobalt. They hated it.
I don't know. If you went backward in time and started selling Solaris for x86 for cheap in 1993, I think fewer people would have shown any interest in what Linux looked like back then. And we might now be living happily in a closed source world with Solaris as its king.
In 1993
And that's the problem. Solaris engineering and Sun management got stuck in 1993 and refused to see the world changing around them, until it was too late.
In case you didn't get the memo: computing in general has changed quite a bit since 1993. Punch cards are no longer in use.
Sun's failures didn't start in 1993. Sun was still a dominating force back then - 24 years ago. Asking what Sun could have done in 1993 to forestall Linux x86's success in the first half of the 2000's is a nonsensical question.
Sun's failure started around 1999, when a $2000 Pentium III laptop running RedHat Linux 5 had better performance than a $45,000 UltraSPARC workstation.
And that is something Sun could have done something about, but didn't. Because SPARC.
Windows 98 didn't exist until after Windows 95. The 95 meant 1995.
In 1993 if you wanted multitasking on a 386 based PC, it was too hard.
Windows 3.1 386 enhanced mode crashed or hung too often. OS/2 wasn't much better. DesqView had its niche, but cost $$$. Minix didn't have virtual memory unless you implemented it yourself; usually a common task for CompSci grads.
FreeBSD still wasn't free. Instead it was still called 386-BSD and it cost $500 due to AT&T licensing fees.
Solaris had superior multithreading and was rock solid. But expensive.
Some kid in Finland started sharing early prototypes of his solution to this problem. But it was still early.
If McNealy actually tinkered with technology more and played less golf and hockey, he might have noticed broad demand in the market for a Unix OS as evidenced by how quickly people started downloading and using Linus' little hobby OS.
Dual boot into what?
Windows 98 or NT or XP. Ever heard of these?
Or FreeBSD. Ever heard of that one?
Why? Because, as a developer, I don't want to be hauling 3 different laptops with me, just because Sun Microsystems wants it so. If Sun was dumb enough to force developers - you know, those weirdos who write software - to carry a separate laptop just to run Solaris x86, goodbye then. Back then, I could boot either Linux or Windows or FreeBSD from grub or LiLO.
Solaris 10 x86 was a ton better than what you described
Not really. Solaris 10 would not install on a PC with a very popular TYAN motherboard. No chipset drivers. RedHat Linux had no problems installing on the same exact PC. I had to wait until Solaris 10 update 4 - two years after the FCS release of Solaris 10 - to be able to install Solaris on that motherboard.
Not a winning strategy, my friend.
cool features like ZFS and DTrace
These cool features that you refer to did not make a dime of difference for Sun or Solaris.
Maybe they weren't really that cool or useful to begin with. Except for certain buildings in Menlo Park, CA where staring in wonderment at your own navel was standard operating procedure.
As for Solaris 10 being open source - namely OpenSolaris - that was a joke too. OpenSolaris was nothing more than a collection of mailing lists where people wasted time in idiotic bickering.
The point is that's when Sun already started failing. When there was still time to beat Linux before it took off.
When your argument about dual booting didn't even matter. Dual boot into what? DOS? Just use a floppy when you really need DOS.
Solaris 10 x86 was a ton better than what you described, and Solaris 11 was actually pretty darn good. Plus, it was open source. Plus, it was free to use. Plus, there was at least some activity with IBM, HP, and Dell. Plus, it not only had a growing effort of all the free software you wanted to run, it had a ton of cool features like ZFS and DTrace that was making some of the linux companies nervous. (see btrfs and SystemTap. Hey, SUSE's promising btrfs will finally be useful Real Soon Now!)
Oracle acquires Sun. Kills the open source. Raises licensing cost to a random high number pulled out of thin air. Kills all the partnerships (no surprise there, of course). Kills off all edu deals. Throttles any freeware porting efforts. Basically, makes it as unattractive as possible to use Solaris on low end hardware.
Hey, how come no one's using Solaris x86 any more?
Specifically in 1993.
1993 is 24 years ago. If Linux from 1993 is your benchmark, then I can't help but wonder: where have you been hiding since 1993? Have you tried a version of Linux x86 more recent than the one from 1993?
Apples to oranges, my friend. You just described Solaris vs. Linux in the late 90's. I said the early 90's. Specifically in 1993.
What I experienced running Solaris 2.5.1 on an old sun4m at my old job is what I wanted to reproduce on my PC at home. Supposedly there was a PC version, but it cost almost $2000.
The old Linux SLS distribution with its pre-1.0 kernel was not so easy.
You had to manually set mode lines in the XFree86 config file and risked physically damaging your monitor to get X-Windows working. In many cases text mode with svgalib apps was a better option. Most vendors wouldn't publish specs to assist the driver developers either.
Sun could have dropped its price and beat Linux before RedHat even existed and made Linux so easy.
I guess I'm in dumb brain mode today, can you explain this post?
Channel partners: a sales method.
Genertic PC's and Linux: a supply chain method.
100% margin: making 100% profit on everything you make. No COGS, all profit. I'd grab me some of that!
Again, maybe it's just me, but this sounds like gibberish. Maybe a bot gone mad?
It wasn't about open source. It was about price. Pony Boy was wrong.
No, it was not about price. It was about open source and Pony Boy was right.
Between 1997 and about 2001 you could buy Solaris x86 directly from Sun for USD $50. I did. I still have the Solaris 7 and 8 for x86 boxes.
Back then, Solaris on x86 plain and simple s---ed. It was a joke. It was next to impossible to install. It wouldn't run on most available chipsets, for lack of available drivers. It wouldn't allow dual booting. Hardware drivers were a mess. It was crashy, buggy and unusable.
To make things worse, Sun was dumb enough to take core open source software such as the X server and make it closed source. Because SPARC. As a result, Sun lost the workstation market to Linux x86. ATI and NVIDIA developed hardware-accelerated drivers for Linux. They did this for free, nobody paid them to do it. You could download the accelerated drivers for free - you still can to this day.
Sun had to pay NVIDIA to write a graphics driver for Solaris x86. That's how popular and viable Solaris x86 was viewed by the industry. While Sun was still trying to sell overpriced $45,000 Solaris SPARC workstations with dual UltraSPARC-III chips, crappy graphics drivers and Slowlaris performance, I could buy a $2000 Intel laptop and get 5 times the performance running RedHat Linux x86 on an Intel Pentium III running at 700MHz.
The simple reason Linux on x86 PC's succeeded is because it was open source. Everyone qualified could contribute. Anyone could write drivers for unsupported devices. As a result, Linux on x86 had working drivers for almost every single chipset. Intel contributed to Linux. IBM contributed to Linux. Cisco contributed to Linux. Not Sun. Sun spent their time and energy undermining Linux in general and Linux on SPARC in particular. They succeeded, Linux on SPARC was abandoned. Did that help Solaris on SPARC?
Linux was easy to install. It allowed dual booting. Anyone could get the source code to Linux, to glibc and to GCC, and hack on them at will.
Not so for Solaris. It was a buggy binary blob. Closed-source. No hopes of anyone working on it to make it better. Sun certainly wasn't working on it. Sun canceled x86 for Solaris 9, again because SPARC.
By the time Sun management and engineering realized that Solaris SPARC was not going to be able to compete with Linux x86 it was already too late. Sun tried a last hail-mary pass with Solaris 10, x86_64 and OpenSolaris, but Linux had already won by then.
Sun and Solaris were destroyed by their own arrogance. Both by management and engineering, there's plenty of blame to go around. They have no-one but themselves to blame for their inept failure.
@OWQEY7M-aqp That was thing that amazed me most about the Oracle acquisition. It couldn't believe the left that management team largely in tact. I figured LE and company would see right though them.
Yet for 7.5 years oracle allowed that management team to continue to drive the systems group and ME.
There are many reasons why Sun failed, but it all stems back to one thing, terrible management.
It was great company with great leadership at one time. Many of these former execs have been tremendously successful at other companies. At some point during the mid/late 90s it was becoming obvious that it was difficult for Sun to retain top management talent. Things were never the same after that.
Sun failed before Linux succeeded.
In the early 90's before I joined Sun I wanted Unix on my PC. I couldn't afford Solaris, and Minix was weak. Then I found Linux. Solaris was by far superior, but I was too poor and Linux was good enough.
If Sun responded in the beginning by selling Solaris x86 at a low price (e.g. $50 or less), I would have used it instead of Linux. So would other people, and Linux would have died or just been someone's hobby.
It wasn't about open source. It was about price. Pony Boy was wrong.
When he open sourced Solaris and gave away the few remaining bits that still made Solaris superior, it was already too late. If Solaris had just been more affordable long ago, nobody would even care about it being closed source.
Windows is closed source and still has 90+ percent market share. But it only costs a fraction of what Sun used to charge for Solaris.
Sun failed because they did not open sourceSolaris or continue to push on the network is the computer. Part of them was visionary but the way they went about it was all wrong.
Solaris could have been what Linux is today. It would be dominant. Virtualization and everything would have been based on top of it. Java continues to be dominant. A couple of decisions a different way and they could have been the dominant computing platform.
SUN failed because of 3rd grade VP and Directors .. use less bunch...
...Never seen such incompetence..