SUN Solaris is effectively dead.

I wanted to wait to get a more concert announcement, but apparently it’s not going to come.  In case you missed it, on Friday Oracle robo called the people working on SUN hardware & software and effectively laid them all off.

We can now dispose of any pretense that Oracle was interested in any SUN technology, or has any plans for it in the future, rather it was a poor ruse to give the impression that there was a future that Oracle was interested in, to drum up a larger settlement amount for how Google deprived them of their licensing for the Dalvik, the Java inspired VM.

Now that Oracle has failed twice in suing Google, there is no longer any need to keep up the charade.

The mismanagement of SUN after the purchase by Oracle is nothing short of stupendous.  While busy chasing pie in the sky litigation, two very important and very lucrative markets were literally expunged from SUN’s portfolio as Oracle clearly has no vision or any competency.

SUN announces the ‘GRID’ in 2006 aka cloud computing

Seriously, look at this announcement from 11 years ago.  Oracle could have been a leader, although given their horrific support and business of treating customers like criminals, perhaps their incompetence is a good thing.

Hi. I’m Jonathan Schwartz, chief operating officer for Sun Microsystems. For nearly a quarter of a century, Sun Microsystems has been saying the network is the computer, and today I’m pleased to announce that Sun is launching the world’s first true computing utility. The Sun Grid is an Internet accessible compute utility that provides anyone in the United States easy access to a super computer for the affordable price of a buck per CPU hour.

So what does it mean for you? If you’re a biotech company, it means that you can plug into the Sun Grid to run complex modeling jobs and get results instantly. If you’re a startup, you can rely on Sun Grid now instead of spending your own capital to build out your own IT infrastructure. Sun Grid let’s you focus on your core competency and get to market faster. If you’re an online business that needs additional resources at peak times, why don’t you provision your own infrastructure for normal work — workloads and let Sun Grid handle the spikes? And if you’re an enterprise running at steady state, use the Sun Grid to test new products and innovations without jeopardizing production systems.

With Sun Grid, you’re not required to a commit to a long-term contract. There’s no hidden fees. You can dial up or dial down based on your own usage of your own compute resources. You don’t need to negotiate the price. You don’t need to negotiate the terms. And you don’t need to meet with a sales person or try to structure a funny finance contract. Simply go to, purchase hours through PayPal with a credit card and start computing.

Today we’re throwing open the doors to the public Sun Grid and inviting everyone to try it out. There’s an interesting sample application up there right now that reads text, say, from an article on the Web and converts it into an MP3 file that you can download to your iPod. Imagine everyone in New York streaming a custom version of the Wall Street Journal to their iPod every morning. With Sun Grid, they can.

We welcome all ISVs to build and deploy software services on the Sun Grid. We’ll be publishing the API specification shortly. For developers generally, Sun Grid offers a heavyweight runtime environment for low-testing your apps across multiple horizontally clustered systems.

Now, I’m sure there’s lots of uses of the Sun Grid that we haven’t even thought of, and we can’t wait to see how it evolves. One of our founders once said, “Innovation happens elsewhere.” Well, log onto and test drive today.

With Sun Grid, the network truly and finally is the computer.

Another interesting artifact of the time is that Jonathan is that he kept a blog, making SUN more accessible.  Sadly this didn’t catch on in the world of tech, where for some reason people either confine themselves to twitter, or Facebook, both platforms that they don’t own, nor can they actually control.  Instead modern CEO’s hide behind PR teams, as they have done since the advent of Public Relations (Thanks Ed!), and of course one of this many posts on the grid.

With the rise of AWS around the same time, there really could have been a far more stronger and competitive area, and as always Microsoft was totally asleep at the wheel letting Virtual Server languish with no x64 path for years to come, and a public cloud not available until 2010!

The Intel Itanium managed to kill almost all the popular RISC processors, and combined with Linux killed the majority of SYSV Unix in the world.  Just as the Itanium Linux combination upset the midrange world, AMD pushed x86_64 to the forefront upsetting the Itanium push before it even began.  In 2017 to not be on x86_64 & ARMv7 is just suicide.  Oddly enough IBM still managed to not only sell PowerPC based kit, but AIX hasn’t seen a new version since 2015 with the larger push of course going to Linux.

Open Solairs

SUN knowing that their future was imploding did the great thing of opening Solaris, Java and other great tech.  Naturally Oracle closed the door soon after buying SUN.


Naturally you’ll never see the words open over at Oracle.  To say the hacker open culture of SUN is incompatible with the litigious and highly customer toxic Oracle is an understatement.  At least this did get the OS out into the wild, with things like ZFS and Dtrace to live on, and outside of Solaris.  However even Apple was scared enough of the possibility of lawsuits from Oracle to have dumped ZFS from OS X, even though they did have it running.

The other amazing bit of tech they did have was VirtualBox, which started out as a German virtualization product to bring the wonders of OS/2 to other operating systems.  VirtualBox coupled with the grid on Solaris x86_64 could have been an incredibly robust platform, but of course that was never to be.  VirtualBox is still around however, so there is that at least.

I suppose there could be a lot more said about the ultimate rise and decline of SUN.

Like Cisco and Oracle, SUN’s fortunes took a dramatic uptick in the late 1999 .com fueled gold rush, along with the parallel Y2K gold rush.

$63.4 to $3.28, source Bloomberg

It’s hard to keep on going when you lose the majority of your value in the .com crash.  SUN struggled to re-define itself, and ultimately rebranding itself as JAVA just wasn’t enough to save itself.  Amazon may have been a retailer, but it quickly gained a reputation as a competent data centre operator, where as SUN was unable to do so.

The death of the SPARC, and Solaris was inevitable, just as the 68000 & SunOS was doomed in the 1980s.  When I talk to enterprise architects on what platforms they are going to use in upcoming projects, the new portable .net is catching some attention, c# may be gaining some traction again, but nobody wants that ‘killer’ combination of Solaris+Java+Oracle.  I know plenty are already tied to it, but I don’t see anyone championing it.

I know from my own personal experience of having a signed contract with Oracle, and having them arrive on site, doing an audit, invalidating their own contract as being “not properly authorized by a ‘rogue’ salesman”, and having our prices doubled overnight, and being forced to downsize our infrastructure as our faster processors had too many cores (even though we disabled the additional cores in the BIOS), I had moved all future projects to Microsoft SQL, as a flat rate of $5,500 retail for SQL Server was a flat out bargain.  I didn’t care if Oracle was 10x as fast, I could buy 100x more MS Servers for the price of an Oracle 10g grid.

The other lesson is that even though Apple could have bought SUN, and had an enterprise market, the bottom line is that it doesn’t matter.  Just as Apple abandoned Enterprise Objects, the X-serve, the Mac Pro, and OS X Server as a real viable server platform, the truth is they make so much more money on iTunes royalties, that they are the ones quickly approaching a 1 Trillion dollar market cap.  The ironic thing is that although you need an enterprise, there is no money in servicing the enterprise, there is far far far more consumers out there.

I’ll keep links I find on this developing story


26 thoughts on “SUN Solaris is effectively dead.

  1. I figured as such. As it happens the fellow that I knew back when they were SUN and trying to create a largely open sourced version, has moved on. And so far I’ve seen nothing from Oracle.
    Now the question is how long does it take for software such theirs to become officially abandoned? That’s the big ticket there. I have here a SPARC box here, for while (s)he was my web server. Now the system just works as an internal file server and general purpose presence.

    Besides Oracle itself is slowly losing its own position as a database product, and we both know why, mind.

      • Now that I can believe. (About AIX.) For my part there’s an emulator out there that’s got enough stuff in it to effectively run the earlier releases.

        But here’s the catch, it will only build on a BSD based OS……….

      • Wow. Gregg, is it a full system emulator? Is it open source? How old are the early releases?
        I wonder if I wasting my time for a problem which is solved long time ago.

        • That emulator is indeed fully open source. It can successfully boot SunOS from as far back as its original release. I’m not sure how it does networking, if at all. I mean for example. its possible to bring up all of the ones for X86 inside either QEMU which is open source, and there is networking present. It is also possible to bring it up inside VMware, which of course is not.

          Last time I tried bringing up that particular emulator it kept crashing during the build process and I did not know who to write to, or how. And this was on Slackware Linux-14.1 on 64 bit. FYI for you Jason, on this laptop.

      • So, Gregg, you are talking about SunOS emulator, not AIX. Is that right? Then you probably mean tme (“The machine emulator”). It builds and works ok under Linux, *BSD is not necessary.

        • Then I must have not gotten it to properly grok my setup properly. That was Slackware64-14.1, and I now have Slackware64-14.2 available. Suppose you aim your work arrangements towards our host, and I’ll collect them.

  2. Yes, it’s really a pity that Oracle has thrown it away.
    Concerning the IBM and OpenPOWER, it may be the sign of falling down like it was with Sun and OpenSPARC.

  3. On the other hand, maybe the Oracle bought Sun not only to sue Google: they use Java in their products since Java 1.2. Maybe it’ll last a bit longer than Solaris.

    • I’m sure someone is telling Larry that line, after the lawsuits failed. But if Java was why, they would have terminated Solaris and SPARC day 1.

      This is the same Oracle that yanked support on HPUX Itanium, regardless how committed HP was.

      • Well I guess the reason to yank the Itanium support was trying to sell the SPARC hardware. They made a lot of noise with all that now-we-are-the-hardware-company-Iron-Man stuff. And as long as they had SPARC, Solaris was necessary.

        But I don’t know how it worked inside, and neither am I a lower. Maybe they were forced to keep the people for a while, and just made the best advertisement out of it.

        Closing the Solaris source was indeed a stupid move – they lost any community interest they had before.

        • I think it’s more so they don’t care about obligations, or contracts.. Much like how the MS of old would view a fine that they can pay as a licensing fee for a practice other people don’t agree with.

          Who even owns UNIX now, like what is a UNIX anymore? AIX? Is there anything left?

          • UNIX? Itself. UNIX is largely community property. AIX is actually IBM’s property. I’m not really sure about its activities. Especially since the favorite OS on Power is an OS managed by a lizard and a well dressed bird.

            Ask the gang on the Historical Society about it.

  4. That it’s two years since a new version of AIX came out isn’t due to them focusing on Linux. If you look at the release history, you’ll see that AIX 7.2 came five years after 7.1 which in turn came three years after 6.1. Between the major new versions you have the Technology Levels which introduce new features.

    I have been working with AIX since shortly after it was released and following the slow development of it is kind of like watching the Pitch Drop experiment.

    • I was more so involved in the good olde days of 3.2 back on the POWER machines, and 4.x on the early PowerPC. After that I wasn’t working at a bank to pony up the big dollars for the hardware & software so that pretty much ended that.

      XLC/XLF was sure fun to fight with… It’s amazing how things got better once we got GCC compiled from source, and later on the fine GNAT folks who provided pre-compiled binaries. A UNIX without a C compiler, really is a fish with a bicycle.

      • I thought the: “Oddly enough IBM still managed to not only sell PowerPC based kit, but AIX hasn’t seen a new version since 2015” was a rather strange statement as well. (1) IBM sold PowerPC machines up until 2009, and that really is irrelevant to AIX on POWER5/6/7/8/9. (2) Itanium never had ANY effect on POWER/PowerPC sales (IBM joined in the fun and sold a few Itanium systems to throw everyone a bone), it’s worth mentioning HPE had to pay intel to refresh another batch and sue Oracle to continue support for Itanium. (3) POWER chips will always be selling as long as z/OS, IBM i and AIX continue to rake in billions.

  5. Oddly enough Oracle does have a few mainframe customers left, but perhaps few. Most of their former customers have switched to the IBM product, who has a wider range of a hardware offering then Oracle had.

    In fact the IBM offering has a strange past.

  6. I was too young to be caught up with SUN and Solaris, but I’ve always looked at Solaris as one of the last remaining things of the 90s boom. Let’s just hope Oracle can keep VBox going for quite sometime. But, considering it’s free I doubt it. They don’t develop it for free out of the goodness of their hearts…

    • For sure, in the 90’s the premier internet serving platform was SUN Solaris. Which is why the first thing we did in that port was get it running on Solaris again, then pivot away once we had it running, as the 90’s are over, and nobody goes forward with Solaris anymore… It’s legacy support at best.

    • It’s not entirely free – as notes, the “Extension Pack” isn’t free for commercial use so you need to buy an “Enterprise License”. Various options available starting at $1220 per socket for the first year.

      My guess is that most people who want to pay for a product like this buy VMware Workstation, but maybe there is a market for it, such as if some integrator sells you some form of all-Oracle “solution”.

      I suppose the VirtualBox team is a lot smaller than the SUN team was, so it’s probably got less of a target painted on it.

      Disclaimer: This is all uninformed speculation.

  7. My university was a BIG Sun shop with Solaris running all the big iron infrastructure. They ran an entire room of Sun SPARC machines as an Andrew File System cluster at one point and there were those silly SunRay terminals all around the campus. Today everything is x86/x64, and its likely that not much of it is running Solaris.

    • Lots of places that were moving away from the VAXen behemoth’s were going down the path of SUN.. I don’t recall that many going the ways of DEC Alpha, IBM POWER or HP HPPA, but it did happen. It sure was amazing the difference from being one of 60 users online on a VAX 8600 to being farmed out onto the new and exciting Ultra’s..

      • They had one machine left running VMS while I was there, likely an Alpha. They were still teaching VAX assembler in 2002! Eventually that system was retired at the end of 2004 when vendor support ended for it.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.