Remember, back in December 2016, when there were rumours Oracle was killing Solaris? And how a month later, Solaris effectively switched to maintenance mode, and then to a “continuous deliver model”?
The news from the ex-Sun community jungle drums is that theJanuary rumourswere true and Oracle laid off the core talent of the Solaris and SPARC teams on Friday. That surely has to mean a maintenance-only future for the product range, especially withSolaris 12 cancelled. A classic Oracle “silent EOL”, no matter what they claim.
With the hardware deprecated, my guess is that’s the last of the Sun assets Oracle acquired written off. Just how good were Oracle’s decisions on buying Sun?
Sun’s Solaris is dead.
Bryan Cantrill on this news (this Bryan Cantrill):
As had been rumored for a while, Oracle effectively killed Solaris on Friday. When I first saw this, I had assumed that this was merely a deep cut, but in talking to Solaris engineers still at Oracle, it is clearly much more than that. It is a cut so deep as to be fatal: the core Solaris engineering organization lost on the order of 90% of its people, including essentially all management.
[…]
Judging merely by its tombstone, the life of Solaris can be viewed as tragic: born out of wedlock between Sun and AT&T and dying at the hands of a remorseless corporate sociopath a quarter century later. And even that may be overstating its longevity: Solaris may not have been truly born until it was made open source, and – certainly to me, anyway – it died the moment it was again made proprietary. But in that shorter life, Solaris achieved the singular: immortality for its revolutionary technologies. So while we can mourn the loss of the proprietary embodiment of Solaris (and we can certainly lament the coarse way in which its technologists were treated!), we can rejoice in the eternal life of its technologies – in illumos and beyond!
I have gathered some sources about this sad day here:
https://forums.freebsd.org/threads/62320/
Sun was delt a mortal blow in 2000 by the dot-com bust.
Then finally died in 2004 when Schwartz became CEO.
After that it was just a corpse to exploit.
Thanks to Schwartz, we have open-source stuff including UltraSparc T1/2 processors.
and open source Java and open source ZFS and OpenSolaris and OpenOffice and a long list of cool projects and tons of open source friendliness.
Don’t get me wrong, Schwartz was a terrible CEO from an economic point of view, but for us, the community, he was like Jesus, he really believed in open source and contributed with the cause. Honestly, I think he was a good guy and you cannot be a good CEO and a good guy at the same time! xD
Regarding Solaris, well, it’s sad, the best enterprise OS ever created death in the hands of the most mediocre IT company in the world. But We, Sun fans and Solaris users, already knew this since 2009. Oracle is f*ing cancer no more no less, capitalism in its most pure state.
Edited 2017-09-05 00:49 UTC
sergio,
I also find it sad that a company that gave so much to the industry was not able to stay viable. The industry is consolidating so much that it’s not enough just to keep building good technology, you have to play the offensive and take others out of the game before they take you out. This process continues until there are a couple corporations running the show.
At least we have antitrust laws that kick in at 50%, but by that point most in the market have already been slaughtered. This is the future we have to look forward to
ZFS is *NOT* OpenSource. It’s not compatible with any GPL or BSD license. We had to remove it from http://www.harvey-os.org project due to issues about CDDL license when Harvey applied to sfconservancy.
Nevertheless it is open source, because you can read the sources.
It is not free software.
Not quite certain how you think the CDDL isn’t open source. It’s FSF and OSI approved, as well as DFSG compatible, so by all industry standard measures, it’s open source. Yes, it’s a limited copyleft, and it’s not GPL compatible, but that is not the same as not being open source.
You’re correct about GPL compatibility, although some Linux projects claim otherwise. However, several BSD operating systems ship ZFS (well now OpenZFS) with the OS.
Licensing is a tricky thing. Two developers reading the same license get a different feeling for what it means. We all really need lawyers for this and even then it would still be a mess.
Even Apple ships dtrace in Mac OS. I think that’s an indicator the license isn’t horrible.
What? This is quite wrong.
The CDDL is a recognised copyleft open source licence. It’s compatible with any licence, including proprietary licences. The GPL is incompatible with the CDDL. That doesn’t make it not open. The same could be said of the Apache licence…
They never made any money out of that though. The protracted open sourcing of Java came about not because of Sun’s benevolence but because they got to a point where they didn’t have the faintest idea what to do with it.
Oracle don’t have the faintest idea of what to do with Solaris either, but believe me, they will not opensource it! That’s difference, that’s the point.
I praise Sun Microsystems and J. Schwartz because they opensource’d a lot of stuff and contributed with the cause. Having a Fortune 500 company like Sun speaking and acting in favor of the opensource movement was a huge help.
I’m not sure how much is left to open source that wasn’t already in OpenSolaris when they ended the project. Maybe some container stuff, but in comparison with what OpenSolaris gave us, I don’t think it’s too bad that we don’t get what remains as open source.
I don’t think he killed it at all. I think after the dotnet boom and the maturation of Linux, Sun was a walking corpse. The products were too expensive to justify use. Sparc had no price/performance advantage over intel/linux for most use cases. By open sourcing everything Schwartz was trying to get more uptake. It was probably too late at that point, but a good effort.
When they started closing things up, that was the final nail in the coffin.
Its a shame, I always wanted to use Sun, but couldn’t ever justify it. I was able to play around with some old sun workstations, which was fun but never used for any real work.
Sweeping statements like this are always dangerous as they show the author is uninformed. This basically is untrue and all you have to do to disprove a statement like this is to come up with one use case.
Use case no.1: Solaris and Sparc could be hardened to become ft-Sparc, multiple processors working in lock-step to achieve fault tolerance. Used by Telcos for mediating data types between mobile telecom switches and trackside locations for control of points and signals. Not the sort of work you would want to be done by linux or Windows. BSOD resulting in lots of RFOD (red faces of death)…
Fault tolerance was one use case – I am sure we would not be hard pressed to come up with another. Don’t even bother trying to argue that linux or Windows systems can be made fault tolerant. They simply cannot and by arguing such you would just be showing your lack of knowledge.
How many telcos are out there? How many more start ups just needed a reliablish web server to run LAMP?
Not everyone needs 100% uptime. Most probably need like 95%. You pay out the nose for each additional nine you need.
That right there explains why Sun died.
Its not that the usecases didn’t exist where Sun was the best choice, its that those usecases weren’t as popular/profitable.
Edited 2017-09-05 17:23 UTC
That’s a different argument. Let’s change our words to suit new facts.
Sorry, my first note was too brief, but that is indeed what I meant.
I have the utmost respect and admiration for the innovations that were contained within Sparc and Solaris. Technologically, they were without a peer.
I was hoping some investor group would have gobbled up both Sun and Novell back in 2000. Unfortunately you had two completely divergent cultures and too much overlap between the two. And now they are both dead.
The mistake here is to assume that Oracle purchased Sun to increase their revenue streams across [all] their product lines.
A lot of the decision was strategic. Sun’s hardware / os was largely being brought for Oracle databases. The writing was already on the wall for that business with Oracle’s move into Linux. And the move off of Sun hardware was always going to accelerate as customers sought to standardise their data centres, and reduce their TCO.
Oracle couldn’t have turned the hardware business into additional revenue – but they needed to manage the transition.
Then there is the additional software. The biggest loss was probably OpenOffice – it might have been attractive, but the community moved against them.
Java hasn’t gone anywhere, despite various grumblings. VirtualBox is also widely used.
But most importantly, don’t overlook MySQL. In buying Sun, they managed to hoover one of their largest database competitors, without it looking like a horribly aggressive move.
OpenOffice.org actually benefited from being abandoned: the LO fork made improvements at a much faster pace than if they had remained under a corporate umbrella. Just a shame that the AOO fork managed to convince some people that it was the original rather than a fork, and that it refuses to acknowledge its own death.
Mysql/Mariadb is a similar story, although Oracle still does just enough Mysql development to keep current users from looking elsewere. Still, nowadays there’s no good reason to keep using Mysql instead of postg^Wmariadb.
Virtualbox is still in good shape (I think ?) but containers, KVM, and Foo-on-Bar technologies have stolen the limelight. If Oracle completely killed VirtualBox today, most users would move to something else without too much trouble.
Java is slowly being handed over to the community, which is probably a good thing (the community is big enough, including corporate backers, to keep the technology in shape). Java will be better off without Oracle.
Disagree with respect to Mysql. Oracle has thrown a lot of man power at it, releases are more stable, more frequent, and more perform-ant.
On the other hand, the price as skyrocketed, and more non community features are being developed. I’ve heard the enterprise features are very nice, but costly.
I don’t really trust Mariadb, not because they aren’t smart people. They just don’t have the qa manpower that oracle does.
Oracle *is* working on MySQL, just not as much as MariaDB is. Oracle only does enough so that existing users don’t flee en masse.
I’d take a MariaDb support contract over an Oracle one any day : beside the price, you’ll certainly get better support (speed and quality) from the company which only sells support (not paywalled features), has to “compete” with community support, and has many people on staff who have been working on MySQL before Oracle and even Sun bought it.
While not an exact superset, it tends to have more features and more focus on performance than MySQL. It also includes (many/all of ?) MySQL’s contract-only features in MariaDb’s open-source release. And it has no internal conflict about encroaching on Oracle DB’s feature set (like simple multi-master replication).
Oracle-owned MySQL is not Free software. Development and even bugtracking happens mostly behind closed doors. The greater community is not involved. Not everybody will agree, but this is be a huge red flag against installing a database.
And lastly, a subjective opinion that is nevertheless shared by a huge chunk of the IT industry: Oracle is the last company you want to be dealing with, due to its horrible business practices.
Honestly, thats more concerning than reassuring. Mysql 5.1 for instance… pleh. Quality really did improve when Sun/Oracle took over. Everything else I agree with.
I think they wanted to offer an integral solution: software+hardware like others did in the past and to develop “key in hand” solutions.
There are some truth & some lie in all that posts.
One of the facts is that there are few hundreds developers in Solaris team right now.
SPARC team was deeply hit by that lay-off not Solaris.
And be sure, it is not simple lay-off, it is a part of internal race for power between different management groups & every thing can be radically changed in the future.
I’m not all that sad to see Solaris die. I’m not happy either, just indifferent. I know a lot of people loved the platform, but I’ve not ever done much with it myself (and my limited experiences have been mixed, managing some of the really old versions at work has been a serious pain in the arse for example).
SPARC on the other hand, I will be sad to see go. It’s a good ISA, and I’d take a SPARC CPU over an equivalent vintage x86 for a server or embedded system any day.
But then again, I’d take an equivalent vintage MIPS or Alpha (or even PA-RISC) over SPARC any day. SPARC really was the runt of the RISC litter.
Love their workstations though. I have a particular fondness for the IPX (first UNIX workstation I ever used) and the other lunchbox SPARCstations.
Man I wish we could get SGI to open up IRIX. It’s so sad what’s happened to UNIX. While I love Linux it feels sad seeing the big brothers falling one by one
Sun/Oracun was only one sparc vendor. A lot of their systems were made by Fujitsu who traditionally makes SPARC64 rather than the T-Series. They also have an O/S license from the Sun era and still produce hardware called “PrimePower” .. they’ve not been allowed to sell around the world so it’s hardly known but great hardware. Perhaps this will free them up?
Who knows.
Another prediction I’ve had for the past decade or so comes true. Solaris was always on borrowed time as soon as Sun’s decline became irreversible. Sun’s hardware has really been pointless since the late nineties, but peaked with the dot com boom.
Quite what Oracle was going to do with it was anyone’s guess.
What about OpenSolaris et al. future?
What do you think guys on their future? Will they we viable and used in production somewhere? Or will they become as hobby OSes?
More talent to -> ?
https://www.openindiana.org/
I think it was load balancers that killed Sun (or rather, gave it a wound that it never sought appropriate action for). Back in the late 90s/Early 2000s, web sites were what pushed the server buying for the most part. You could do it with x86 or you could do it with Unix/RISC, the later being far more expensive.
Sun’s idea of a web server, even after the dot-com crash, was a $25,000 server (E280R, for example). Dell would sell you a pizza-box for around $2K. The E280R was faster, but not nearly 10x faster. It was more reliable, but behind a load balancer it didn’t much matter. You could scale out as you needed.
Sun didn’t get that it couldn’t sell a $25K web server anymore until it was far too late.
There was also its dumb-ass battle with Linux.
Edited 2017-09-05 21:26 UTC
They got MySQL, which is probably all they ever wanted.
Yes, I know, the community will fork and fork, but they cannot get back the paid developers from Sun that MySQL had.
You must know that database software is the ultimate lock-in, because no serious business will risk all the micro-optimizations they have done to their queries (some of them accomplished with proprietary extensions to standard SQL), and you must also know that MySQL had reached a level it was seriously threatening Oracle’s database software. Taking the paid devs from MySQL was a way to make MySQL stop evolving into a bigger threat to Oracle’s database software.
Everything else was a distraction. Nobody really believed that some SPARC CPUs with an inefficient ISA (duplication of registers due to register windows, sorry RISC fans,RISC doesn’t automatically mean good ISA) and stuck on an old 40nm node process could ever be a threat to AMD, or that Solaris could beat RHEL in anything.
Edited 2017-09-06 09:40 UTC
MariaDB is binary-compatible with MySQL, and while no DB migration should be taken lightly, the MySQL -> MariaDB one is likely to be completely transparent.
MariaDB attracted most of the original talent (paid and voluntary) behind MySQL. This is a trend that started before Oracle bought Sun, and dramatically accelerated once Oracle owned MySQL. The company behind MariaDB seems to be growing healthily, as are the contributions from the volunteer community.
They bought SUN for Java. Period.
Eh, Mysql wasn’t really an Oracle Competitor. It sorta was kinda trying for a while, but discovered replicating Oracle is really hard ( see the 5.0 features foreign keys, triggers, stored procedures, views, etc). While at the same time Mysql was really good at things Oracle was not ( speeeed!) .
No sane person would run something like peoplesoft/SAP on mysql. And your hipster social media/tech companies wouldn’t touch Oracle with a fifty foot pole.
HOWEVER:
What many people forget is that postgres development was being supported by SUN. Postgres was much more of an oracle alternative, especially the commercial offspin Enterprise Db. So I think you have the right motivation, but the wrong victim.
Edited 2017-09-06 17:11 UTC