> Sun had become bored with the mechanics of running a business.
I would contend that Sun was never interested in running a business. They always cared more about building amazing technology and suffered the sales to make money.
You can't deny that Sun hardware was top tier and always ahead of its time (until Oracle bought them).
Edit: I had on rose colored glasses. I was thinking of late 90s SUN. They did in fact start getting crushed in the early 2000s.
I think this sorta forgets how slow the cpus were - I was running mostly ASIC design workloads on big Sun hardware in 2001 and we eventually got some x86 boxes that were only 32 bit but could run some of our jobs - and the single-threaded perf of these commodity machines was amazing compared to the Sparc boxes. When x86-64 came around, this was the death of all the Sun hardware in that group.
Also sometime in 2000 I had interviewed at Compaq - the old Alpha group who I think later worked on Itanium at Intel. They were constantly frustrated because they couldn't even get a build from Synopsys of the design tools they needed to run on Alpha and were forced to use Sun hardware they described as "gasping and wheezing" to design their crazy fast cpu.
Sun lost on a bunch of fronts, but I think the amazing thing is they let happen to them what they did to the other big UNIX vendors before - they were eaten from the bottom by faster cheaper hardware.
Same here. I've been in the chip design industry since 1997. Everything was Sun and I was right out of college and convinced my manager to let me use a Linux/x86 machine and I used the savings to buy a large 21" monitor that would run at 2048x1536 resolution and use it as an X terminal to the Suns. Everyone else loved it and did the same.
Around 2000 we started getting some EDA software released on Linux. I remember getting some Linux kernel patches and recompiling the kernel to change the user / kernel split from 2GB / 2GB to 3.5GB user / 0.5GB kernel. But we still needed the 64-bit Suns for some jobs that needed over 4GB RAM. Itanium was really expensive and slow and then AMD released the 64-bit Opteron in 2003. We bought them immediately and never bought another Sun.
My high school in 1992 got a bunch of old Sun3 and first gen Sun4 systems. I absolutely loved them. My dream then was to have my own Sun workstation but they were way too expensive. My other dream was to go to college and then get a job at Sun. I heard about Linux in 1993 and got my Unix system at home. I did work on some Sun UltraSPARC III companion chipsets but Sun couldn't compete on the price to performance ratio with Linux/x86
We had the Sun workstations in college (some with that crazy laser mouse) and I think towards the end of my tenure, most of the labs were switching to linux boxes, but at that same time Solaris x86 actually came out - so on commodity hardware you could run Solaris which was kinda fun.
I did this a little, but the things you could do with actual Linux started to get interesting for me at that time - building routers for our shared home network, hosting Quake servers, doing actual programming with gcc etc and the hardware support was a bit better as I recall.
I also interviewed at Sun in Boston - I think the small group servers group or something - I really liked the engineers I met there and I did consider working there for a minute, but ended up in Silicon Valley which I think was better fit for me.
Yes, I had the same experience while using various Sun computers between 1995 and 2005. During all that decade the Sun computers were reliable, but excessively slow.
Around the beginning of 2006, I had a laptop with a 64-bit AMD Turion CPU and I could run on it Cadence and Mentor tools for ASIC design and the simulations were much faster than on a big Sun multi-processor server. Later that year our team finally got some servers with AMD Opteron, which accelerated dramatically the work in comparison with the traditional Sun servers.
Both SPARC and PowerPC met the same fate. A brilliant ISA done poorly was no match for a poor ISA (x86) done really well.
I had the same experience twice in a few months from those two. Moving from a Sparc server to a Xeon based system, and again when the first Intel Mac Mini's came out and was faster than the dual-G5 systems we had.
Thanks for sharing. I just read an article comparing UltraSparc III and Pentium 4. The frequency is the biggest difference, and while the article also highlights a few pros for the Sun processor, like being 64-bit and super scalar, the benchmark shows that P4 has a clear win. I don’t know the price tag then, but I guess Sparc was actually more expensive?
The article is here: https://www.silkstream.net/blog/2006/08/comparison-ultraspar...
The Sun system would have cost several thousand dollars more than a generic Pentium 4 box.
I recall a fork/exec heavy shell script at a previous job that ran about an order of magnitude (~25s vs 3s) slower on Solaris/SparcV9 vs Linux/Intel P3 circa 2003. The standard retort back in those days was to rewrite it in Perl.
I can deny it. It was OK, but not amazing. I had to make pretty frequent repair calls on their gear--much more frequently than the less-technologically-amazing-but-solid-as-a-rock Dell and Compaq gear. IBM, Digital, and HP (yes, that HP, before they lost the Way) also had very good UNIX gear, but Sun had the mindshare during the '90s.
In terms of performance, even Linux 0.x on a 33Mhz 486-class processor could outperform a uniprocessor Sun/SPARC box at a fraction of the price.
I also question the sanity of anyone who thinks Sun's choice of 13W3 for a video connector was a good one. :-)
I don't know, the clustering on Solaris was pretty amazing, if we are just judging "ahead of time"
DEC's VAX/VMS had decent clustering in the early 80's.
Any Unix "clustering" system I've used (Veritas on Sun, HA/CMP on AIX) always seemed bolted on.
I'd say the basic idea of 13W3 connector was fine, what was not was not trying to agree with others on pinout, and evolving the connector in a way that made invisible incompatible difference whether you had DDC-capable or not
I loved Sun hardware but I'll deny that they were ahead of their time after 2003 or so. I had a boss who loved his pizza box but Intel/Linux crushed them in the 2000s on a price/performance basis.
One of the things that Larry Ellison found utterly baffling when he acquired Sun was the way that sales commission was calculated. It was all on the value of the contract, with no requirement for the contract to be profitable.
At least in the later stages of life Sun sales folks were literally selling the company into bankruptcy. They'd find out what competitors were offering, undercut it to the degree they'd make a loss, just to get the contract and their sweet commission.
To not realise you've got this going on really speaks to how little they were interested in the actual running of the business.
> You can't deny that Sun hardware was top tier and always ahead of its time (until Oracle bought them).
From Brian Cantrill's original HN comment in 2011(https://news.ycombinator.com/item?id=2287033): "I never hesitated to tell anyone that was listening that our x86 boxes were starting to smoke the hell out of UltraSPARC". SPARC stopped being ahead of its time way before being bought by Oracle.
>> You can't deny that Sun hardware was top tier and always ahead of its time (until Oracle bought them).
> SPARC stopped being ahead of its time way before being bought by Oracle.
You seem to be focusing strictly on performance, but there were other factors: e.g., remote management is still a bit of a mess with x86. With SPARC everything can be done on the CLI (serials consoles or SSH), with x86 quite often you're still clicking on "open KVM window" for so many things. And of course many/most vendors make you purchase an 'advanced license' to use the remote KVM functionality of the LOM system.
Why is that a property of the ISA?
It's path dependence. If you built a Sun-syle x86 server with a serial console and inserted a random Linux installer CD, the installer would probably only output VGA and thus it wouldn't work.
That's part that was changed with UEFI when the old IBM PC "contract" was finally severed.
I would wager that almost no Linux installs are done by CD or displaying VGA anymore (maybe 0.01%).
> I would wager that almost no Linux installs are done by CD or displaying VGA anymore (maybe 0.01%).
When one of your Proxmox (or XCP-ng, or VMware/Nutanix/OpenStack) servers won't boot and you need go into single-user/rescue mode, then talk to me about the convenience or not of a remote KVM (versus having to drive to the DC and hook up the crash cart).
It's all very well if all you manage is VMs/instances, but some of us deal with the hardware that those VMs run on, and things break.
That 0.01% of servers is probably the layer that the other 99.99% run on (and yes, a lot of that 0.01% can also be automated, but if PXE isn't working (or you're getting strange CPU3VOLT or SLOT17ECC or BUS1RESET errors) then neither is re-imaging).
The alternative isn't driving to the DC; it's remote serial console through a console server or the BMC.
Path dependence again. x86 servers in the 1990s had to have VGA so it was never removed. (Except Oxide.)
Oxide happened long after (U)EFI brought proper serial support to boot process
Yet every UEFI server has VGA and remote KVM.
Debian installer runs fine over serial, at least.
Yeah, that was amazing. And so, I learned how to provide LILO + agetty on COM1 on every server I was admining :) So, unless there was HW fault, I could just connect to server via serial to check things out.
> You seem to be focusing strictly on performance,
Just like the vast majority of buyers do. SPARC was abandoned by customers due to poor performance.
> but there were other factors: e.g., remote management is still a bit of a mess with x86
That's a property of the systems integrator, not the ISA. Dell's iDRAC works quite well.
> Dell's iDRAC works quite well.
I currently admin several racks of Dells (and Lenovo, and SuperMicro, and previously HP). Just make sure you buy the license for KVM in case you ever need to go into single-user/rescue mode (otherwise better not mind having to drive to the DC and hooking up a crash cart).
Ask @bcantrill (the author of this article/story) about how well integrators integrate (as a hint to the answer, he co-founded Oxide Computers because he ran a cloud service with Dells/HPs/etc).
You can ignore KVM license if you have good netboot infrastructure and only install OS in UEFI mode - SOL works fine.
The big issue is usually "remote media" support, thus my recommendation for setting up netbooting infrastructure
Integrators integrate very well at the level that they're operating, i.e. the single server. Rack-level integrated systems are a whole different thing, very specialised and unrelated to the old thing.
Sun's various divisions all cared very much about their respective vendor lock-in, so they didn't innovate fast enough. The divisions that didn't have that innovated. On the way downward Sun cannibalized itself (think Sun PS getting closed).
That's the whole gist.