SoylentNews is people
SoylentNews
The January rumours were true and on Friday Oracle laid off the core talent from the Solaris and SPARC teams, in effect finally killing what they had left of Sun Microsystems. When Oracle aquired Sun, there were a lot of valuable assets, each of which, except VirtualBox, has been squandered and abandoned. Simon Phipps enumerates the main ones and what happened to them.
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Hacking's just another word for nothing left to kludge.
(Score: 2) by richtopia on Tuesday September 05 2017, @07:45PM (2 children)
I wonder if SUN had not been purchased if things would have worked out differently. For the most part I liked dealing with SUN products, but I am not a network administrator so my opinion is not well founded. Oracle's decisions lead to many forks in SUN's open source projects: OpenOffice.org, ZFS, OpenSolaris for example. However I remember SUN was in serious trouble financially, and probably couldn't have kept the lights on (they were shedding employees even during the acquisition).
Has anyone here used the T-series of SPARC processors? I remember reading about the Niagara when it was first released with so many threads and wondering what applications it would be used for. I'm still not sure if there is a niche somewhere that loves these chips or if they are mostly legacy.
(Score: 3, Interesting) by frojack on Tuesday September 05 2017, @08:21PM
The acquisition spree that Sun itself went on [wikipedia.org] was probably the seeds of its own demise.
All the stuff they gobbled up was not necessary, never went anywhere, and brought no revenue. But a few attracted attention of Oracle: MySQL. An unnecessary shot across Oracle's bow.
There were a cluster of acquisitions that made it look like Sun was going to make a serious run at Virtual computing, but I doubt Oracle was all that worried about that. Vmware or Intel might have snapped up that segment if the first IB Bubble hadn't intervened.
No, you are mistaken. I've always had this sig.
(Score: 5, Informative) by turgid on Tuesday September 05 2017, @08:24PM
It would have died years ago in that case.
Correct.
I've worked for a couple of big, famous American companies during my career, including Sun in a very junior engineering position just as the Dot Com bubble burst.
The hubris and ego in the place was enormous when I joined and quite intoxicating, but after a year or two it seemed like the company had lost its way. Sun was rightly very proud of Solaris, Java and its SPARC hardware but many of the old-timers/engineering egos refused to acknowledge the threat from Linux, they refused to see x86 performance and scalability catching up and they really didn't want to admit that not everyone wanted to code in Java for the Java platform. They just didn't get the whole FOSS thing at all. They couldn't understand why people were writing in Perl, Python, C, C++, Ruby... They didn't get GNU tools either.
They even all but killed Solaris x86 just as AMD64 came out, but eventually they saw the light and resurrected it, doing a port to AMD64 in six weeks flat, including paying to have AMD64 support in gcc and binutils fixed for Solaris.
We (young, enthusiastic, naive) tried to tell them (managers, institutionalised big egos), and they did start to listen but it was too late. They did open source Solaris, they did bring out a line of x86-64 machines, they even open sourced SPARC, they sponsored GNOME development (back when it was still good) and made it an official desktop on Solaris, they bought Star Office and open sourced it.
However, the engineering egos really didn't "get" why anyone would run Linux instead of "free" Solaris. They didn't understand that we preferred the more user-friendly and modern GNU userland. Solaris had to stick to official POSIX/SUS standards for commercial contractual reasons, but they did bolt on some GNU/FOSS stuff as an afterthought.
The PHBs and VPs liked to exert their authority and take "decisions."
The UltraSPARC development was years late. It was way behind POWER from IBM. Even though AIX was awful compared to Solaris, POWER hardware was a lot faster in some cases. Fujitsu had its own SPARC64 CPUs which were much better than UltraSPARC and Sun did eventually adopt some of them.
The fact that all the other Unix vendors had got aboard the Good Ship itanic was a golden opportunity for Sun to pull ahead with UltraSPARC but they just couldn't execute. They should have made more of AMD64 but that would have been losing face.
Now, as for the T-Series UltraSPARC (Niagara), those were very interesting. They were designed to be very simple and highly parallel, ideal for certain tasks such as web serving. They were no good for compute-intensive stuff but they had low context switch latency and high memory bandwidth.
A few years after I was laid off from Sun I found myself working for another large company that used Sun hardware and they bought some of the UltraSPARC T2 systems (with a whopping 32GB RAM) as development servers hosting ClearCase (yuck). They were much slower than the old boxes they replaced because they were the wrong CPUs for the job. ClearCase didn't rely much on multi-threading so the advantages of the T2 processors were lost, and their disadvantages amplified...
I believe that the subsequent T-processors were much more powerful. Oracle invested heavily in their development, and continued to use the Fujitsu technology for the more conventional servers.
What really upset me was when Oracle closed off Solaris development. I had collected some old but nice Sun gear (including a couple of 4-CPU Ultra 80 workstations) at home and it instantly became useless.
Being taken over by Oracle was bad, but we used to believe that being taken over by IBM would have been worse. Who knows? It's all history now.
I refuse to engage in a battle of wits with an unarmed opponent [wikipedia.org].