4 ms·

Oracle euthanizes Solaris 12, expunging it from roadmap

33 points by amyjess 10y ago

zeveb 10y ago
Sad news — Solaris really was great once. Personally, I think it started to go to hell when it adopted XML for loads of stuff, but that's just my own bias showing.
- SwellJoe 10y ago
  I think Solaris lost because of the Linux juggernaut, not because of any particular technical decision at Sun. Free, Open Source, and massively popular, Linux has slaughtered almost everything in its path on the server (and now, on mobile, too). Solaris was far superior in the early days of Linux, everyone would agree, but Linux got better faster than Sun was able to address all of the reasons people were choosing Linux: Cost, hardware flexibility (and cheap hardware), nicer userland (every Solaris box I ever inherited the administration of already had the GNU tools on it, because they were just better), incredibly enthusiastic user base (Linux and OSS had a religious zeal that commercial vendors could never match), and it was Good Enough even early on to start making inroads into Solaris' home turf. I suspect XML had an imperceptible effect on Solaris usage. They generally had tools for most of the stuff that went XML; you didn't have to edit it directly. Sure, XML is ugly, but it aint a dealbreaker for most people, especially if there are tools for managing it.
  - zeveb 10y ago
    That's a much more mature & detailed viewpoint that I espoused in my post.
  - justinclift 10y ago
    Sun going back on promises about x86 in the earlier days didn't help. I used to do Solaris SysAdmin stuff, and was a very strong advocate for it even on x86. Sun made guarantees/assurances that it'd be available for free download for x86. Some months later (did someone look at the download bandwidth costs? ;>), they completely reneged on that and stopped allowing people to download x86. That was (literally) the last day I did any kind of advocacy for Sun or Solaris. After that it was FreeBSD and Linux, and "fuck Sun". ;)
    - SwellJoe 10y ago
      Yeah, I remember that change. I had a couple of Sun SPARC boxes at a clients office, and had an installation of Solaris x86 at my house on a spare box, just so I could stay familiar with the system and try things in a non-production environment. It was an unpleasant surprise when they pulled the free x86 version. But, I doubt it was catastrophic for Sun; I also doubt bandwidth was a major concern for them, though I'm not confident about why they pulled the x86 free version. My recollection of events and post-mortems from people within Sun about that decision and other decisions surrounding x86 Solaris was that there was a battle going on within Sun about what the future of Sun looked like (which happens when a ship is sinking as Sun was). There were those who wanted to double down on hardware, because it had always been their bread-and-butter and was still their primary profit source. Putting Solaris on x86, as with putting Mac OS on Mac clones, was cannibalizing their own hardware sales. Dell could push out high quality x86 servers for half the price of Sun, and if someone could put Solaris x86 on it, there wasn't much reason for companies to spend the extra money for a Sun box (so the thinking went for the folks who were in the "bet on hardware" camp at Sun). The other camp wanted to tackle Linux on its on turf, and pushed for x86 support, free, and eventually open source. The path to profits among that camp were a little less clear than the champions of hardware, but they did have some successes eventually in getting Solaris open sourced (but too late, as we all now know). I dunno. By the time Solaris was Open Source, Linux was a better system for most users, even pretty big users. And, then there was no entry point for Sun into any company that had left behind Solaris on SPARC for Linux on x86. They had no ties to SPARC, anymore, and no ties to Solaris. So, nothing left to leverage for service revenue or anything else. The same battle played out at several other UNIX workstation vendors. Sun was just the last one standing. SGI went whole-hog for Windows NT, trying to keep their hardware business afloat (and failed, going through a couple of acquisitions after the fact, and now they're what was formerly Rackable). I don't know what Sun could have done to survive fighting a war on two fronts, with Windows on the enterprise desktop/workstation side, and Linux on the server side. They made it longer than most of their competitors in the space, so maybe they made the right decisions but the innovator's dilemma just didn't leave them a lot of options short of shrinking the company and changing its focus and soul massively.
- jjtheblunt 10y ago
  anything that adopted XML exhibited a symptom that sensible engineers were no longer making decisions, but rather cheerleaders were.
sengork 10y ago
In some form and fashion parts of Solaris and its concepts will live on in FreeBSD and Linux, namely ZFS, zones/containers, Dtrace and SMF. I hope that SPARC itself can live on after Solaris itself tend to go hand in hand though...
- ksec 10y ago
  I hope many of those Solaris dev would join FreeBSD.
- yellowapple 10y ago
  Don't forget illumos, which forked from OpenSolaris.
  - justinclift 10y ago
    Yeah, it's commit history seems reasonable and growing (?) over time: https://github.com/illumos/illumos-gate/graphs/contributors https://github.com/illumos/illumos-gate/graphs/contributors

33 points by amyjess 10y ago

zeveb 10y ago

Sad news — Solaris really was great once. Personally, I think it started to go to hell when it adopted XML for loads of stuff, but that's just my own bias showing.

SwellJoe 10y ago
I think Solaris lost because of the Linux juggernaut, not because of any particular technical decision at Sun. Free, Open Source, and massively popular, Linux has slaughtered almost everything in its path on the server (and now, on mobile, too). Solaris was far superior in the early days of Linux, everyone would agree, but Linux got better faster than Sun was able to address all of the reasons people were choosing Linux: Cost, hardware flexibility (and cheap hardware), nicer userland (every Solaris box I ever inherited the administration of already had the GNU tools on it, because they were just better), incredibly enthusiastic user base (Linux and OSS had a religious zeal that commercial vendors could never match), and it was Good Enough even early on to start making inroads into Solaris' home turf. I suspect XML had an imperceptible effect on Solaris usage. They generally had tools for most of the stuff that went XML; you didn't have to edit it directly. Sure, XML is ugly, but it aint a dealbreaker for most people, especially if there are tools for managing it.
- zeveb 10y ago
  That's a much more mature & detailed viewpoint that I espoused in my post.
- justinclift 10y ago
  Sun going back on promises about x86 in the earlier days didn't help. I used to do Solaris SysAdmin stuff, and was a very strong advocate for it even on x86. Sun made guarantees/assurances that it'd be available for free download for x86. Some months later (did someone look at the download bandwidth costs? ;>), they completely reneged on that and stopped allowing people to download x86. That was (literally) the last day I did any kind of advocacy for Sun or Solaris. After that it was FreeBSD and Linux, and "fuck Sun". ;)
  - SwellJoe 10y ago
    Yeah, I remember that change. I had a couple of Sun SPARC boxes at a clients office, and had an installation of Solaris x86 at my house on a spare box, just so I could stay familiar with the system and try things in a non-production environment. It was an unpleasant surprise when they pulled the free x86 version. But, I doubt it was catastrophic for Sun; I also doubt bandwidth was a major concern for them, though I'm not confident about why they pulled the x86 free version. My recollection of events and post-mortems from people within Sun about that decision and other decisions surrounding x86 Solaris was that there was a battle going on within Sun about what the future of Sun looked like (which happens when a ship is sinking as Sun was). There were those who wanted to double down on hardware, because it had always been their bread-and-butter and was still their primary profit source. Putting Solaris on x86, as with putting Mac OS on Mac clones, was cannibalizing their own hardware sales. Dell could push out high quality x86 servers for half the price of Sun, and if someone could put Solaris x86 on it, there wasn't much reason for companies to spend the extra money for a Sun box (so the thinking went for the folks who were in the "bet on hardware" camp at Sun). The other camp wanted to tackle Linux on its on turf, and pushed for x86 support, free, and eventually open source. The path to profits among that camp were a little less clear than the champions of hardware, but they did have some successes eventually in getting Solaris open sourced (but too late, as we all now know). I dunno. By the time Solaris was Open Source, Linux was a better system for most users, even pretty big users. And, then there was no entry point for Sun into any company that had left behind Solaris on SPARC for Linux on x86. They had no ties to SPARC, anymore, and no ties to Solaris. So, nothing left to leverage for service revenue or anything else. The same battle played out at several other UNIX workstation vendors. Sun was just the last one standing. SGI went whole-hog for Windows NT, trying to keep their hardware business afloat (and failed, going through a couple of acquisitions after the fact, and now they're what was formerly Rackable). I don't know what Sun could have done to survive fighting a war on two fronts, with Windows on the enterprise desktop/workstation side, and Linux on the server side. They made it longer than most of their competitors in the space, so maybe they made the right decisions but the innovator's dilemma just didn't leave them a lot of options short of shrinking the company and changing its focus and soul massively.
jjtheblunt 10y ago
anything that adopted XML exhibited a symptom that sensible engineers were no longer making decisions, but rather cheerleaders were.

sengork 10y ago

In some form and fashion parts of Solaris and its concepts will live on in FreeBSD and Linux, namely ZFS, zones/containers, Dtrace and SMF. I hope that SPARC itself can live on after Solaris itself tend to go hand in hand though...

ksec 10y ago
I hope many of those Solaris dev would join FreeBSD.
yellowapple 10y ago
Don't forget illumos, which forked from OpenSolaris.
- justinclift 10y ago
  Yeah, it's commit history seems reasonable and growing (?) over time: https://github.com/illumos/illumos-gate/graphs/contributors https://github.com/illumos/illumos-gate/graphs/contributors