Ex-Sun CEO Warns Oracle of Death By Open Source 408
gearystwatcher writes "Former Sun CEO Scott McNealy talks to The Reg on where things went wrong, and acquisition by Oracle: 'We probably got a little too aggressive near the end and probably open sourced too much and tried too hard to appease the community and tried too hard to share,' McNealy said. 'You gotta take care of your shareholders or you end up very vulnerable like we got. We were a wonderful acquisition — we got stolen for a song at the bottom of the Dow.'"
Business vs Open Source (Score:4, Insightful)
Definitely, if all the valuable assets of your business is in software (Solaris, StarOffice, Java, etc) and you give away such software for free then your business does not make sense at all.
Blame open-source (Score:4, Insightful)
Right, the only mistake Sun did was open-source too much. Like all the closed shop were doing wonderfully well too.
Thanks Sun.
Re:Business vs Open Source (Score:4, Insightful)
Definitely, if all the valuable assets of your business is in software (Solaris, StarOffice, Java, etc) and you give away such software for free then your business does not make sense at all.
Those "valuable assets" of the business are now worth nothing, better free alternatives exist. The part that doesn't make sense is not successfully moving onward to a consultative / training / services based business structure.
Re:Business vs Open Source (Score:5, Insightful)
Sun just couldn't compete with Linux and Intel. Open sourcing wasn't the problem. It probably helped, just not enough.
Re:Blame open-source (Score:5, Insightful)
The mistake they made was that they forgot (or didn't know how) to monetize the open source solutions they had. OpenSolaris was great, Java was great, OpenOffice was great but there was no option to buy support or custom development for those products. The only way was to go with closed Solaris and StarOffice which were quite different products and required IT folks to migrate. Basically they pushed OpenSolaris as a development vehicle for their closed Solaris which made for a bunch of OpenSolaris installs way ahead and more feature-rich (patch-wise) than Solaris, migrating back was a pain (or impossible if you upgraded your ZFS pools), installing Sun software on it was a pain.
If anything I would say they didn't open source enough of their products for it to be a success. OpenSolaris would've been great in a well-marketed product like Nexenta did - take the closed source out of it, allow for the great amount of Linux software to run directly on it and make it easy as Ubuntu. But their stock repositories were crap and hard to find requiring signing up to get keys or stick to the handful of community repos. Their HA and Storage solutions are still the best you can find in the market but again, hard to install on OpenSolaris and not very compatible with other software and systems.
Their hardware was also overpriced which pushed them right out of the market. I can understand the higher pricing on their SPARC products but not for their generic x86 systems.
Meally mouthed (Score:5, Insightful)
Open Source mouth; insert foot (Score:4, Insightful)
I notice that your Slashdot UID is 1378985. You have clearly concluded that there are almost 1 and 1/2 million Slashdot readers, all of whom have no business experience and troll from their basement. This in spite of numerous recent articles about the fact that the mass of linux kernel and other major Open Source project developers are paid developers, many of whom work for very big name companies in the high technology industry. You have clearly forgotten who has the experience to analyze McNealy's position, and whom the guy without the experience opening his mouth out of turn is.
News Flash: There are many, many, many people in the world more qualified to analyze where Sun went wrong with their approach to Open Source than Larry Ellison and Scott McNealy, and some of them are right here on Slashdot. In fact, the decline of Sun could be viewed as specific evidence that there was a lack of understanding about Open Source on his part. HP, IBM, Intel, and many other big name hardware/software companies seem to have managed to keep on without losing their shirt in the process, for example.
Re:Business vs Open Source (Score:5, Insightful)
I'd say Java was pretty valuable. Love or hate Java, it's used all over the place.
They just never figured out a way to turn that mass user base into serious profit without losing their users.
Re:Nothing to do with it (Score:5, Insightful)
It's more that Linux became "good enough" for a lot tasks and was cheaper. Having worked around both, Solaris still has features that if it's needed are worth the money.
Re:Business vs Open Source (Score:5, Insightful)
That is the point. Java is only valuable begause it is given away for free. If Sun (or Oracle now) tried to sell it, it would be nearly worthless.
Re:Sun made a strategic mistake not tactical one. (Score:4, Insightful)
"Then Linux got ported into intel" and all this time I thought it was written originally on the 386.
Re:Business vs Open Source (Score:2, Insightful)
Where Sun went wrong was in not having open sourced these earlier. They messed around with Solaris, withdrawing the free (as in beer) x86 version at one point. Uncertainty wasn't exactly giving potential customers the "warm fuzzies" required to invest in Sparc hardware (we were evaluating Solaris on x86 at that time). Java, despite the shortcomings would have been a compelling choice for many apps had it been open source. Finally, StarOffice was never a serious threat to the monopoly until they open sourced it.
Had everything been open sourced 10 years ago, the technology landscape today would be very different. From where I've sat, it's not open sourcing their code but simple bad management that took the company down.
Re:Business vs Open Source (Score:5, Insightful)
Exactly, and part of the reason the free alternatives exist is because Sun made them free (e.g. Openoffice, or Java open-source friendliness).
What Sun failed to do when open-sourcing their "valuable software assets" was to establish a business plan to go with it. RedHat has a business plan related to go with their open-source Linux distributions; IBM has a business plan to go with their Eclipse open-source software... Sun? even though I like them a lot ... it is true that they were not business sound from a long time.
They had the complete vertical stack (hardware [Sparc], middleware [Java] and software [Solaris] and services [cloud services]) but never really came up with a business plan.
Again, it has been really good for us (the open source community, free software advocates) but it was terrible for the economic viability of the Sun corporation (thus resulting in its end).
Re:Ayn Rand? (Score:3, Insightful)
And which makes you feel better about yourself if you are not rich and powerful?
American Idol.
Re:Business vs Open Source (Score:5, Insightful)
Sun bought Cobalt, a successful Linux server business, and instead of capitalizing on it, buried it. That alone is worth the corporate death penalty. If Sun had swallowed their pride and put their weight fully behind their Linux server business they would now own a serious share of the world's server rooms, not to mention the personal server business. Consider: Red Hat's market cap is now over 9 billion, and that without any hardware offering. How on earth did Sun miss the party?
The weirdest thing is, Larry Ellison fully intends to continue the idiocy of shoving Solaris and Sparc down the throats of customers who don't want it. The inevitable result couldn't happen to a nicer guy.
Re:IBM did well with Java (and other F/OSS softwar (Score:5, Insightful)
IBM has seen open source as a means of creating "solutions" for customers not a money maker in of itself. Ibm views linux as a sturdy and inexpensive tool that it can spend money to become very expert at. Sun sells expensive tools, IBM sells solutions to customers needs using inexpensive tools. That is why IBM is very very rich.
Re:IBM did well with Java (and other F/OSS softwar (Score:5, Insightful)
Interestingly, IBM and Oracle got more value out of using Java than Sun did by writing it. That appears to have been one of the decision points for Oracle.
One of the extra advantages of Linux for IBM was that it offered a new OS for the 390s, and of a very popular flavor. Sun already had a Unix OS for SPARC, so they didn't get the added value.
Sun was, IMHO, always a "BSD vs Bell" shop: they understood the struggle to free BSD, and learned how to deal with Bell and the commercial world, but that's where they stopped.
--dave
Re:Business vs Open Source (Score:4, Insightful)
What Oracle should have done is start hitting the R&D, and start offering SPARC and x86 hardware with enterprise friendly features. Open-sourcing Solaris would get more people onto that environment, be it college students, or others who want to test stuff out.
If Oracle hardware supported even a fraction of some of these, they would still be head to head with IBM for the enterprise market, and not being squeezed in a vise with IBM hardware (zSeries and pSeries) on the high end, and commodity x86 on the low end:
1: ZFS. ZFS could have sold Oracle hardware once it started being able to handle the enterprise slings and arrows. Sun could have added hooks for hardware, so things like rebuilding a failed HDD could be done on a lower level and not bother the CPU with I/O.
2: VM capabilities. Zones and LDoms should be an integral part of the hardware a long time ago, as it is on the IBM POWER7s. Add hooks for moving VMs between physical machines while the VM is still running (vMotion essentially), and high availability, and this will bring the enterprise dollars.
3: Get college friendly, like the Sun of old. The students who imprint on the Oracle hardware with day to day work will be the ones speccing out the big machines later on in life.
4: Start making backend systems with applications where there is a need. For example, a way to get some type of solution that is 100% compatible with Exchange. This way, E-mail and messaging can run on SPARC hardware, and that would get it into enterprises where only x86 machines go now. Another example is document management, like Adobe's LiveCycle. Hardware will not sell unless it has applications on it. Databases are just one facet of enterprise computing.
5: Differentiate from x86 hardware. IBM does this by having reliability as one of their selling points. It isn't uncommon to see 99.999% uptime on POWER hardware, and mainframes pretty much guarantee this.
6: Start working on more R&D with Internet protocols. Sun pioneered the landscape with NIS, NIS+, NFS, and many other protocols. Most are antiquated now, but they were better than nothing.
7: Start doing security innovations. For example, consider having NIC cards that have independent packet filters in them. This way, an attacker would have to compromise the NIC card (with a hardened hardware attack surface) before they could get access to the machine. DoS attacks could be handled by the NICs leaving the machine unscathed. More points if an IDS/IPS is built in. Solaris has come a long way with regards to security, but it doesn't hurt to keep advancing.
8: Work on new hardware projects. Take IBM's ZTIC. This is a simple device, but greatly ups the ante on bank fraud and ID theft. Oracle needs to work on projects like that.