back to article Oracle's systems boss bails amid deafening silence over Solaris fate

Oracle has revealed that John Fowler, a Sun veteran who stayed to serve as Oracle's executive veep for systems, has left the company. An SEC filing [PDF] dated July 27, 2017 says “John F Fowler resigned his position as Executive Vice President, Systems effective as of August 2, 2017.” No reason is offered for Fowler's …

  1. HCV

    Conventional?

    If that is the case, it will represent conventional release cadence for Solaris.

    Like a Swiss chronometer, it is, if by "like a Swiss chronometer" you mean "every November, or possibly October, unless it's April, or perhaps not at all that year."

    1. Brian Miller
      Joke

      Re: Conventional?

      So it's like a Swiss chronometer if it's like an English chronometer?

  2. Anonymous Coward
    Anonymous Coward

    It's sort of sad to hear of what sounds like the SUN going out (SUN workstations being my first introduction to unix, as a student), but on the other hand, Linux and BSD descendants power pretty much everything nowadays; the song may change, but the music plays on…

    1. This post has been deleted by its author

    2. Anonymous Coward
      Anonymous Coward

      @a/c

      "t's sort of sad to hear of what sounds like the SUN going out "

      Indeed. And while Solaris is a good OS it doesn't really have enough stand out features to make the cost worthwhile these days.

      "Linux and BSD descendants power pretty much everything nowadays;"

      The nice thing about Solaris for me wasn't the OS, it was the hardware. PCs are so damn boring and everyone knows that under the hood the hardware is pretty cheap and the architecture is a kludge. Sparc kit was a step up in quality especially the workstations.

      1. rlhamil

        Re: @a/c

        The hardware sucked vastly less than most x86 hardware. But Solaris 10 and even 11 (despite Oracle's backstabbing removal of Solaris from mostly open development) have plenty of standout features: dTrace, ZFS, Zones; with in Solaris 11 the added goodness of a proper repository based approach, eliminating the nastiness of patch-based updating; and in later increments, OpenStack integration. Looking at the Linux rough equivalents to those first three named features, and I think they'll all seem lame by comparison. Solaris 10 had insane innovation; Solaris 11 put in the maintainability and cloud-ifying aspects that were missing. Aside from drivers for everything and ports to way too many other platforms and ideology, what the heck does amateur Linux have that Solaris doesn't? :-)

        1. Anonymous Coward
          Anonymous Coward

          Re: @a/c

          Like v10, Solaris 11 did bring a fair amount of standout features. Granted, some where improvements of what v10 introduced. I particularly liked some of the networking enhancements. And yes, Oracle did Solaris a massive disservice by killing off opensolaris which was helping generate an interest in Solaris again after Sun dropped Solaris x86 prior to bringing it back w/v10. Luckily there have been some forks to OpenSolaris to help keep an amazing OS alive.

  3. Wade Burchette

    This makes me sad

    I coded all my C and C++ projects on a Solaris Unix computer in college. This makes me sad because every soul Oracle touches withers and dies inside. Just like that lady that was forced to have a mandatory binding arbitration clause in her contract by Oracle and after she won in Oracle's kangaroo court, Oracle sued their own kangaroo court.

    It is only a matter of time before MySQL and Java succumbs to soul crushing society of Oracle. If Oracle ever gets their hands on PHP or ColdFusion, then my college days will a time of great sadness because Oracle would have their hands on everything I did in college except COBOL. (And I hated COBOL because I hate accounting.)

    1. Nutria

      Re: This makes me sad

      On the plus side, they're still rolling out updates to Rdb (the OpenVMS database server which they bought from DEC back in 1994).

      1. rnturn

        Re: Rdb

        Interesting to hear that. I recall hearing that, after Oracle bought Rdb from DEC, that they had it running in their labs trying to figure out how the hell it performed so well---running rings around Oracle's software. DEC may have suffered greatly from inept marketing but their software teams wrote excellent code.

    2. Lysenko

      Re: This makes me sad

      Oracle get their hands on Zend/Personal Home Page?? PLEASE make it happen. Then we would only need to kill HHVM to render PHP as a bad memory like PowerBuilder or DataEase.

    3. Anonymous Coward
      Anonymous Coward

      Re: This makes me sad

      hardware sales were $4.15bn for the 12 months to June, but that is declining

      Well, to be fair to Oracle (I know, I know) much of that can be laid at Fowler's door. Over the last few years he's done little visible beyond keeping his chair warm. When the news of his (presumably non-voluntary) departure broke last week the most common reaction I heard from ex-Sun types was "what took them so long?".

      As to the gitgub repository being "hardly a hive of activity", most of the Solaris folks I know have spent the last 6 months working like crazy to unpick Solaris 12 from all the new stuff so that upgrades from 11.3 to 11.4 still go smoothly. I doubt if incremental updates to an opensource repo were either useful or high on anyones priority list.

      Don't write Solaris or SPARC off just yet. The general feeling seems to be that Fowler's departure could be a good thing for both.

      1. Doctor Huh?

        Re: This makes me sad

        "Don't write Solaris or SPARC off just yet."

        Really?!?! We are in an era of elastic pools of commodity hardware and software that can be thrown at problems as provided by every Cloud provider. You can even get GPU pools on The Cloud. If that won't do it for you, you could go the Google route and grab a cheap open source (or custom) design for hardware and scale the crap out of that. Or you can go to the low end and bulk out with Raspberry Pi.

        I loved SPARC back in the day, and Solaris was rock solid, but where today is there enough volume to justify a pricey bespoke Enterprise architecture and OS? And if there is enough volume to support that, is there really enough volume to support 3 (add in HP with PA-RISC/HP-UX and IBM with Power/AIX). The Cloud isn't the solution to everything, and X86_64/Linux/BSD may not be the solution to every problem, but is the remainder really enough to support ongoing SPARC/Solaris development?

        1. joeldillon

          Re: This makes me sad

          PA-RISC has been dead for many years now actually, modern HP-UX runs on Itanium. But yeah, if any commercial Unix vendor is going to die off it's them, chip development on the Itanium has been moribund for a while now, and HP have been making noises about making it easy to transition people to Linux.

          1. Mad Mike

            Re: This makes me sad

            @joeldillon.

            I think it's fair to say that HP-UX and Itanium is effecitvely dead. Might take a while for the corpse to stop twitching, but that's always the case. They could always create a HP-UX version for X86 hardware and that might help some people transition, but the end is nigh.

            1. Alistair
              Coat

              Re: This makes me sad

              @ MadMike:

              HPUX on x86? Why? We've managed to move enormous amounts of code from HPUX to linux on x86 with only moderate lifting. Hell, I picked up an entire MF app in java and moved it over without a hiccup, no changes needed. Mind you, the java is ancient and getting the app owners to let me advance the version to something approaching supported is somewhat painful, but progress was made.

              Cobol stuff needed some twiddling but recompiled well. License cost was stupid, at first, but when told what we were doing, the licensing became more rational. C code based stuff was a sweep and tidy on includes mostly and away we went (mind you the C stuff on hpux was originally compiled with gcc from the external depots)

              HPUX on Itanium has its place - and you wont get the speeds out of x86_64 without *almost* the same investment in $$ worth of hardware, and floor space, but it still works in certain cases.

              1. Anonymous Coward
                Anonymous Coward

                Re: This makes me sad

                " Hell, I picked up an entire MF app in java and moved it over without a hiccup"

                Err, isn't that the whole raison d'etre of Java? If it can't do that then why on earth would you use it in the first place in place instead of C++?

            2. Anonymous Coward
              Anonymous Coward

              Re: This makes me sad

              "They could always create a HP-UX version for X86 hardware and that might help some people transition, but the end is nigh."

              They could but it would hardly be worth the effort and cost. Might just as well role their own linux distro and port any tools and apps. They could port their compiler to perhaps make compiling HP-UX C/C++ programs simpler but other than that ....

              1. Anonymous Coward
                Anonymous Coward

                Re: This makes me sad

                HP/UX-x86 was there. Booting and running.

                Then came Oracle and said "We will not support Oracle RDBMS on that platform!"

                Off went 90% of the potential use case and HP/UX-x86 got killed.

                1. Anonymous Coward
                  Anonymous Coward

                  Re: This makes me sad

                  Karma is a bitch, open source DB's will do to Oracle what it did to so many others.

                  I won't miss the Nazi bastards.

              2. FrankAlphaXII

                Re: This makes me sad

                They're doing exactly that with OpenVMS y'know.

                Never say never if there's a use case for a product.

        2. Mad Mike

          Re: This makes me sad

          @Doctor Huh.

          The strange thing is that some of the modern Sparc servers are pretty reasonably priced. Not the top end stuff that still costs an arm and a leg, but certainly the lower end stuff. Price/performance is similar to x86. I guess it depends on where Oracle go with the Software in Silicon stuff. If there's enough accelerators in Sparc that aren't present anywhere else, then people may start buying it, as price/performance should improve and surpass x86.

        3. Anonymous Coward
          Anonymous Coward

          Re: This makes me sad

          Doctor Huh? quote: "...support a pricey bespoke Enterprise architecture and OS?"

          *

          Well...define pricey! Take a look at events like this:

          https://www.theregister.co.uk/2017/03/02/aws_s3_meltdown/

          *

          Then tell me how cheap a four hour outage looks to clients of AWS who have bet the farm on "cloud". And this isn't the first (or the last) time that cloud will look "cheap" only while it's up.

          *

          Yes...having your systems inside the enterprise isn't risk free either....but at least you have management control over what you choose to keep in-house.

          1. Anonymous Coward
            Anonymous Coward

            Re: This makes me sad

            Public cloud has a much better uptime than anything you can engineer in your DC, the difference is when your solution has an outage it doesn't make the news. Its also more secure, despite what many would like to espouse.

            "Everything fails, all the time" should be your guiding architectural principal.

            Cloud will win over on site virtualised environments, sadly Solaris is a long way in the rear view mirror, thanks for the memories (and all the contracting money).

        4. Even Jelical

          Re: This makes me sad

          PA-RISC? did someone step in a time warp?

      2. Anonymous Coward
        Devil

        Re: This makes me sad

        > Don't write Solaris or SPARC off just yet.

        Denial is not just a river in Egypt.

        The five stages of loss and grief.

        Stage 1: Denial. It could take a while.

        After that, four more to go.

    4. Anonymous Coward
      Anonymous Coward

      Re: This makes me sad

      Oracle, following in CA's footsteps.

      I remember hearing stories--and living through a couple of them--about how CA would buy up smaller, up-coming competitors and kill off their products. The killing off usually came in the form of jacking up the pricing of the product to anyone who wanted to buy it--or the annual support to existing customers--to stratospheric levels. It got to the point that some large customers were demanding that from their software vendors that code be kept in escrow and made available to the customer should the vendor ever be purchased by CA.

      At least the base MySQL code is available online and has been forked already so users of that are relatively safe from any shenanigans by Oracle. Unless you bought into any of their proprietary add-ons that Oracle's legal team would make sure never saw the light of day in any forked code base.

  4. adamr001

    The are killing off more than just the FS1

    They also recently killed off the Oracle Virtual Networking (former Xsigo) product. It is listed as EOL on the customer portal and they scrubbed all references to the product with the exception of a small mention on the Private Cloud Appliance product page (which isn't EOL).

    They also scrubbed all references to their upcomming EDR InfiniBand fabric that was covered here. https://www.theregister.co.uk/2016/02/23/oracle_emits_carrier_enterprise_kit_at_mobile_world_congress/

    1. Anonymous Coward
      Anonymous Coward

      Re: The are killing off more than just the FS1

      Looks like the SPARC/Linux version of Exadata (SL6) is also gone, after a Scaramucci-esque tenure on the price list?

  5. ptribble

    Not the repo you're looking for

    You can't read anything into the activity of the solaris-userland repo - that's just the holder for the external open-source projects they ship as part of the product. Activity there is largely driven by CVEs and releases of those external projects, and it may not be what they're using as the primary internal repo in any case.

    But yes, Oracle's stewardship of a flagship product has hardly been encouraging. You want that stuff, there's always the illumos project.

    1. Hans 1

      Re: Not the repo you're looking for

      there's always the illumos project.

      I raise you to OpenInidiana, a humoungous amount of pacthes have come in as of late, if you like Solaris, check it out again, I tell ya!

    2. RantyDave

      Re: Not the repo you're looking for

      I've recently been using OmniOS (http://www.omniosce.org) and, thus far, has been awesome. A significantly less 'opinionated' SmartOS.

      1. Fazal Majid

        Re: Not the repo you're looking for

        Unfortunately OmniTI disengaged itself from OmniOS. Whether the project is sustainable as a community project is anyone's guess:

        https://lists.omniti.com/pipermail/omnios-discuss/2017-April/008699.html

  6. Mad Mike

    Future of Sparc

    Oracle is currently making a very big thing of 'Software in Silicon'. This mostly relates to the Sparc platform, where they are busy implementing hardware assistance for numerous function previous performed solely in software. Now, this implies Sparc has a future as a hardware platform, but what operating system will it run? Of course, Linux is an option, but there doesn't seem a lot of movement in that area, especially from oracle. Maybe they're just waiting and lining everything up for a big announcement; or maybe not. Who knows.

    1. Anonymous Coward
      Anonymous Coward

      Re: Future of Sparc

      I would not put it past Oracle to make a play into network switching. They have people who know silicon (SPARC, etc). Solaris has some history as a strong networking OS. They can do appliances. Just a thought.

      1. Roo
        Windows

        Re: Future of Sparc

        "I would not put it past Oracle to make a play into network switching. They have people who know silicon (SPARC, etc)"

        They'd need to sink a lot of money into licensing & cross licensing to pursue something beyond SPARC, chipmaking has been dominated by lawyers and accountants for a very long time now.

    2. Even Jelical

      Re: Future of Sparc

      Not as Good as Software on Chip..... as in M10/M12

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like