I think all the well-known ones have been thoroughly covered by this point. I can think of one or two obscure ones. For example, 3D workstation companies Stellar and Ardent were acquired by the same investors, who forced them to merge into Stardent--did they manage to achieve anything interesting before fading into oblivion? Then there were Alias and Wavefront, who also combined together. You don’t hear much about these any more (though the memory of Alias|Wavefront lives on in the “.obj” format, still popular for 3D data interchange, because of its sheer simplicity).
Sun vet here, ‘99 to 2004. The video doesn’t really get what happened. First, the “sharp elbows” were replaced around the time I got there. We were winning for many reasons, but two of them were the safe and solid OS, Solaris, and partnerships with industry Allies: Sun, Veritas, and Cisco. The performance advantages were still there, and winning deals was pretty easy. Unfortunately, hiring enough people was getting tough, and the culture was changing. Lower to middle management got taken over by a bunch of political parasites many who had learned survival tactics at DEC and Compaq. When things turned, these people completely quit working on keeping the company afloat, and started scheming to be the last thrown overboard as there really wasn’t any place left to go. Scott had seemingly lost control of the culture. Zander, who had been a decent COO totally threw Sun under the bus as he departed for Motorola, where he showed his inability to be a CEO which then surprised no one at Sun. Morale crashed, even though we still had an excellent platform. Perhaps there was no way forward, but it seemed to me the ingredients were there for a better solution than the world ended up with. Open systems are too open, and thieves reach right in.
You sound awfully bitter about the move to open software. By all accounts, the open systems that replaced you were more stable, more compatible, and drastically more user friendly.
@@jrshaul I’m bitter, and it’s not just about open systems. Open systems have lots of advantages, but none of them are stability (quite the opposite. I had customers with systems that had been forgotten about because they had simply run for YEARS without attention), and this has nothing to do with user friendliness which isn’t a factor for operating systems, and has not been for thirty years. Most users never see the OS anymore. I’m also bitter about all sorts of things mostly thrust on us by an education system that teaches amazing amounts of trash without bothering to teach the importance of prioritizing value over price.
@@Baronvonbadguy3 Sure. Name a subject. 🤣😂🤣😂 You know who else had rants? Scott McNealy. No kidding. He used to do a podcast for the employees, and he would do a rant on each one. One time, he went off on the money the company wasted buying vegetarian pizzas which were mostly thrown away. It was hilarious. Unfortunately, I’m not blessed with his talent for humor. Sorry.
@@jrshaul > You sound awfully bitter about the move to open software. Open Software killed off a lot of companies. Nobody ever thought that making a free version of UNIX was a sound business case for the companies that made proprietary HW and OSs. How they got away with copying and mimicking the UNIX code and didn't get sued is beyond me.
As a former Sun reseller and support contractor, I'm very happy for you covering my favorite company for your christmas video. Thanks! I'm still working in Solaris support today and have a large collections of systems around. So sad that Oracle has ruined what was once a great company.. but that's what happens when a software guy buys a hardware company and has no idea what they actually do or how to run them. When Ellison swooped in, the first he did was double the system prices and triple the support contract fees, and struck down any special deals that Sun had going with their largest customers for decades. This lead to most of those customers taking the jump over to Linux clusters, that were just about becoming mainstream enough at that time.
Was it that Ellison is a software guy, or simply that he's a greedy b****rd who makes decisions motivated by short-term profit (or delusions thereof) and that alienate the customer? Seemed to me it was the latter, though being a software guy I could be biased :)
I wonder if killing many of Sun technologies was Oracle's plan all along. Afraid that some of the technologies Sun had could pose threat to Oracle dominance in the enterprise markets.
But somewhere you seem to not realize that SUN was not profitable any longer when Oracle acquired it, cannot blame Oracle for that. Engineers seem to believe that there is an unlimited amount of money and that companies do not have to make profit.
I was blessed to work for sun early in my career from 1988 to 1991 as a sales engineer. We called ourselves systems engineers at that time. What a learning experience that was: The people, the technology, the momentum, and the possibilities. As an engineer, I was amazed that a unix system could have an uptime greater than a week without requiring a reboot. Like anything, the growth was too fast and the company changed in spirit in the mid to late 90s. Great summary of Sun!
To this day, I'm always going for 4-digit uptimes. As a manager for a fairly large IT department, I admonished my guys to fix problems, with my line "Restarting is for losers."
Unfortunately, OS uptime seems to be going the way of the dodo. Containers seem to be helping but all of the patch restarts today are really frustrating. I remember when I would only have to patch/reboot an OS once or twice a year if that.
funny story about a Sun Micro Mainframe. My father started working as a Junior accountant for a Large Queensland meat processing abattoir. they had a sun computer system. when costing the production of sausages by the ton, different wholesalers order them with different ingredients... and these had to be coasted down to the 4 decimal point. and this computer was worked by terminal and punch card. and a salesman would send through ingredient list and volume etc and an hour later they would get a printout handed back to them, and then they would get back to the client with a cost often a 2 hour info round trip. well dad bought a TRS80-4k Level 1 computer when they first came out, supposed to be for me. he wrote a basic program that would work out these costings almost instantly (well a few seconds) so he could talk to the sales person while on the phone. and the next month Byte magazine came out with a basic program that would turn such data into a pie chart. one day day took the computer to work and put it on his desk, and was doing the costings instantly, when the Managing director walked past, and asked what he was doing. dad put on a demo, including a pie chart, and the MD asked the Mainframe guy to make a pie chart for next weeks board meeting. and sure enough next week at the boardmeeting, the Heda of computing showed his pie chart, and dad demonstrated the little TRS-80. and the CEO asked the main frame guy aput him makingthe pie cahrtas, and he said they they had to purchase Sun's special program for pie charst that cost $15K (remember this was the late 1970's) so the MD stood up and said to the main frame guy "your fired" and had the Main frame sent to the tip and had TRs80 put on every salemans desk, and dad was promoted to head to sales... (dad made several changed to production, that had him promoted to 2IC in a few years)...
@@johancoetzer2165 ACTUALLY... Sun did make mainframes towards the end, right before being bought up by Oracle: the M-line. The design was done by Sun and manufactured by Fujitsu. The hardware is completely redundant and can be partitioned, just like IBM's zSeries mainframes, additional processors can be enabled temporarily or permanently with a license... The only difference is, this mainframe hardware is not running z/VM and z/OS, but Solaris 10 or 11. 10 in my case. M3000 was the first, code named "Teraya" (I own one privately), followed by other mainframe models like the M10. These were all made after 2007.
I had a summer job at Sun in 1996, it was an absolute dream, the epitome of cool for a computing student job. Having that on my CV opened so many doors. I miss them.
I miss Sun. I still have a couple SPARCstations sitting in the spare bedroom, a SPARC 20 and an Ultra 5. I haven't powered them up in years, though. I was given the 20 and I bought the Ultra 5 new, way back in the 1990s. They, and the IPC and IPX I used to have, made a nice little home network.
For years while maintaining Sun servers in NJ/CT/NY, I traded/bought used Sun parts from a company in New Jersey...I once spent 2 days upgrading two Sun E6800s by hand (3 quad CPU boards and four 8-card PCI "I/O boats" each). The CPU boards had 32 DIMMs per board (96 per server, 192 DIMMs over both servers)...I had bought 6 newer CPU board and new I/O cards. But I needed to transfer my RAM from the old CPU boards to the newer CPU boards before handing the old boards in as part of the trade up. So 192 DIMMs out, 192 DIMMs back in.
@@carlschumacher5510 Weren't these sun hardware incredibly expensive, and hard to maintain? Didn't they fail all the time? Those sun E series servers were designed wrong. Or that they never notified anyone how to properly mount them. The E4000 and up units had the air intake on the right side and exhaust heat to the left side instead of normal servers that do front to rear. So if you end up putting 2 E4500s next to one another then the one on the left would be constantly hot. Stupid design, and add to the fact that they never told anyone on how to properly mount them. Our company had these and they were always hot. I was a datacenter tech told to look into it. but the company had onsite sun techs to replace parts and they never said anything.
@@xcen1 From 1994 thru 2012 I worked with many models of Sun boxes (from ~$5K up to ~$800K models) but only a handful for E4000/E4500s...Every now and then I'd run into a popular Sun model that had a major flaw I had to work around. The E450 was a perhaps 10U very deep model that I don't think was meant to be racked (it had wheels iirc). The E450 was a model that could have 1 to 4 CPUs, one or two 10-drive bays I think, and various levels of RAM. It had a huge motherboard mounted vertically in iirc the middle of the chassis. Said MB was not supported well (it was easy to crack when installing parts). I wasn't the only one killing said motherboards, the local Sun techs were also...My solution?: I convinced management to let me only order fully-loaded E450s (all CPU & RAM slots filled, both drive cages installed, all the PCI cards I'd ever want). This was during the Doc Com era. In the early 2000, working with E6500s / E6800s / E6900s was a step up in Sun boxes (but not quite to the E10K). These were "I am a rack" frames that you kept for 5 - 7 years, adding CPU/RAM boards over time (for example as more Oracle databases were added). The E6800/E6900 were true monsters. They had two internal power planes, the OOB management computer was actually a cluster of two small computers (PowerPC based I think). The beast was a real friend of anyone selling electricity. Over the years, as the CPUs got a bit faster, you could add/upgrade any of the 6 CPU board slots (each board holding 4 CPUs)...The coolest thing I have EVER done to a production server, was replacing a CPU board in an E6800 while it was up and serving requests (you could tell the OS to move processes off of a given CPU board). One last comment on user expectations of Sun gear, during my E6800/E6900 period (2000 - 2007). Whenever I wanted downtime (patch, reboot, replace a failed part, etc) everyone "Why Carl? Why? I thought these things were bullet proof? Why?" (sigh!)...The Windows guys just had to say "Patch Tuesday". Zero objections...I remember almost hitting a clock bug (some issue after 520 days of uptime in the OOB server cluster (yeah, not a power of 2)) and having to get time slots to reboot the whole E6800/E6900 fleet.
@@xcen1 The 4500 was "cheap" for what it was in it's day. It was an NFS and mid-range data base server without an equal at its price point. The only real problem, as you pointed out, was the airflow design could lead to overheating and indeed fires. I worked at Sun from '96 to '99 personally saw one system catch fire and assisted another customer replacing a unit that had also caught fire. It wasn't intended to be a datacenter unit but rather a small office, stand alone design. The price performance was so good, however, that at one point that's all datacenter customers wanted to buy.
@@justinbukoski1 OMG no wonder sun went away. That's an ass backwards idea to say it was meant to be in office. Any server should be in datacenter environment for so many reasons 1 being a security issue. The E4000 machine weighs as much as a car. Me being a skinny kid in 2000 at 5'8" and 110 lbs I was able to lift up a HP DL380g1 by myself at that time, But this thing was so heavy 2 of us struggle to move it one time. After the first time i learned how to move it by removing everything inside, all the boards, and psu's then it was manageable to move. Jesus to put a server in an office, someone could walk by and spill coffee all over it. What a stupid idea..... And it has to be in a datacenter environment because it needed front and rear access. You're going to put it up against a wall in an office, then you'll move it to change and rear boards or psu. In a rack in a datacenter, you have open access to to rear and front.
When I came to study at my uni in 1994, all their services (DNS, mail etc.) ran on lone SPARCclassic workstation. Since they had no one familiar with Unix and I already knew Linux, they hired me to take care of that server and I eventually ended up working full time in their IT and this set up me for my career in networking. And it all started with that small Sun computer. Good memories.
That's how I got into the biz, so to speak. I got my hands on 8 floppies that just happened to have a Linux Slackware distro on it. I completely ditched Windows, quit college and got a job working in the field, using nothing more than a decent ability to handle the dry reading of countless manpages, coupled with knowing C. I worked as a Sun level II software support specialist for a couple of years, but lost that gig when Sun started contracting out level II support. I was an odd guy to have around, really useful, but odd. I needed help using a Windows work station, but if it ran some flavor of Unix, I was perfectly at home.
@@snorman1911All to find out you enabled the wrong driver. ;-) Good times! When the kernel went modular, that's when the overnight compile run ended.. mostly..
*_Former Boeing Everett..._* Boeing used Sun Micro Systems, Silicone Graphics, IBM Catia, Windows NT, IBM OS2, Mac, and Windows 3.xx over a very large campus. I was part of 350 member IT Desktop Support Team in Everett. There were 50,000 devices total, including Workstations, Servers, CAD, Network Printers, and more, used by 30,000 employees in Everett. It was challenging supporting all the various desktop configurations. When Boeing bought McDonald Douglas, we inherited their mess too. I transferred to Lean Manufacturing later. At that time Boeing was consolidating IT and standardizing Computing. IBM Catia was being upgraded to handle PC functions as part of Single Glass Initiative. There were different departments using 2-3 different types of workstations per desk. Single Glass solved part of that problem. Boeing had it's own versions of Windows Operating Systems and Windows Office. Boeing had the largest network running Microsoft Products in the world. Everett was just 1 of many other campuses around the globe. Our Call Center in Bellevue handled a millions calls per year. IT was 24/7 365 a year. Boeing may have slowed down over the Holidays, but it never stopped providing world class support to all it's computer users.
@@NoTraceOfSenseThanks for comment... Yes I was there during the merger. I was part of Lean Manufacturing and was sent to St Louis to iron out production kinks in the F/A-18 Super Hornets. Amazing experience.
Thank you for this trip down memory Lane. Working in the workstation marketplace in the late 1980’s you touched on many things that were formative parts of my career, and friendships I still have today. The best man at my wedding I met at HP, he left and we both ended up at a reseller. I few months in and he was poached by Apollo (for a crazy amount of money) and he ended up back at HP 😂. He left again. My oldest friend worked for ARM in its early days. I defected early to the Microsoft bandwagon as my interest was business not technical and I started working with Olivetti hardware (still with the occasional unix install at the start), so it was nice to see their name pop up. Always remember the Olivetti joke that it was going to be called Italian Business Machines but they didn’t think the initials would take off. But I’m rambling now - as us old folk tend to 😂.
I remember when engineering workstations were the most powerful stand alone desktop computers. They, at onetime were many time more powerful than any IBM PC (and its 386 clones) or Apple desktop computer. But, in time Moore's law changed everything, and consumer grade PCs eventually could run the most sophisticated CAD/EDA and simulation software, at a much lower cost. Sun and Apollo workstations had lost their edge. The last thing I remember from Sun was Java.
I worked on Sun workstations (hardware diagnostics programming) starting around 1990, and was amazed to find just a couple of years later that a 486 upgraded to a Pentium and running Linux ran compute-intensive tasks about twice as fast. Our workstations were 2-3 years old by then, but by 1995, the Linux PC was absolutely faster than all but the most expensive Sun workstations, and we stopped buying them. Sun forgot their open roots, beginning a little with SunOS, going full retard with SPARC processors, and by then they had dug too deep a hole to ever get out of. The end of the 1990s was like a sad little joke as they kept on digging that hole. I think hey could have kept going with SunOS and Solaris, wouldn't have been too hard a problem to re-open it like Linux. But Motorola botched the 68030 and 68040 advances, SPARC was a desperate gamble which paid off for a few years, but there was no going back, they had too much face invested to switch to Intel x86, and that was all she wrote. (I can understand Sun shuddering at the x86 architecture. I wrote assembler for a lot of architectures, all were fun and interesting one way or the other, but x86 always gave me the creeps, about the ugliest architecture possible, and that didn't fit their image.)
They fell to the same arrogance has Digital Equipment (DEC): they refused to acknowledge the competition and to a combination of lowering the price and raising the performance.
Yeah, the PPro and successors ate old RISC's lunch, Windows NT/2K and Linux ate old UNIX, and when Nvidia released the og GeForce and Quadros they ate the pro 3D market. Same story in all cases, the low end market just becoming good enough to gradually eat the high end from under it. I still miss SPARC/Solaris though, and those lovely purple boxes. Not so much the E450 that almost fell out of a rack and crushed me once.
I started working on Suns around 1990 and continued using them through grad school and beyond, up until about 2010. They were a great machine to develop on since Sun controlled both the hardware and the software and created a very integrated feel. At the time, Apple's OS was a mess of spaghetti code. Now I develop on a macbook pro (Apple long ago got smart and switched to unix) and while it's gotten much better, it's still not as smooth as Solaris was in its heyday.
I thought Apple's MacOS was built on Unix, which is true but I recently learnt that it's not the full story. Apparently the older gen Mac OS was replaced with an OS that was based on a Mach kernel (ending with h), and one of the BSD's stack was built on top of this Mach kernel. This OS was named OSX, which some time later became MacOS. You probably know this being an apple user. It's strange how there is more to the story than I thought. Similar to how Android is not GNU/Linux, but Java-like Virtual Machines built on the Linux kernel.
Its a mistake to call Mac os UNIC because Mac os cant do any mission critical task or multiuser doing multitask Calling mac os unix is like calling a teenager a soldier because won a kungfu tournament
@@elumiomerk4013 System 7 to 8 was so broken that you would reboot 10x per day. Networking and dial up was broken, SCSI was bad. The things were slow and there were no graphics cards available. Shitbox by shitheads. G3 were probably the faultiest products shipped ever.
Sun workstations with Solaris were marvelous systems and we used them extensively at my place of work. Their reliability was an asset in those days. I still remember fondly using the unix base system with Motif GUI.
Sun NFS was really bulletproof. I was there for the entire ride, despite starting in 1981 on the IBM PC. The early VME bus systems were good, the 386 based design was a flop, but the Sparc based systems were just incredible. We had a Sparc 2 as our NIS master that had an uptime of 4 years when it came time to upgrade for y2k. We deployed hundreds every year in the 1990s, every software developer had one on his desktop. Too bad they couldn't catch up with the latest process nodes, it was funny to hear about dual channel memory in the PC being such a big innovation when we had been installing quad channel memory for years.
NFS was designed to be “stateless”, which caused no end of problems for Unix software that was expecting a baseline of sane behaviour from the filesystem -- ironic for a Unix-centric company, don’t you think?
@@lawrencedoliveiro9104 I obviously had a much different experience. Compared to other vendors (and especially Linux) the Sun implementation of NFS "just worked". When they enabled TCP rather than UDP transmission, it just worked. Since our primary use was an NFS based application, I had a lot of experience in that regard.
@@lawrencedoliveiro9104 NFS had some issues, but having department-wide computer transparent file system in mid 90-s, was plain fantastic. Literally, in first two years in my university department as a postdoc, I did not even know to what computer my home directory is physically attached. Compare that with Windows 95 paradigm. And I still use NFS among the three Linux machines I manage in my office now
Didn't Sun or Stanford license one of their first cards to three companies? Believe one was in Chicago, the other was Sun Micro, can't remember the third. I worked in applications my first two years out of college helping with MC68000 support for multiple customers. Supported Apollo, Apple, Sun, and many others.
The amount of Sun workstations that Nortel bought in the late 1990s and early 2000s you wouldn't believe, as many of the manufacturing rigs used them. When Nortel went into serious decline, their Sun workstation orders dried up very quickly as they had more than enough machines spare.
I had such a beast of a CRT monitor. made by Sun. My brother picked it up from a trashcan at UCSC. It was at least 1440p back when most lcd monitors came at 1080, and most tvs where at 720. It was at least 80lbs, and 27" Lasted me so many years, had to get rid of it because of its the weight after I left college in 2013.
I remember back in college in 2010 We used dated Sun Microsystems Desktop PCs in one of our classes; they were definitely unique. I remember the cases of both the monitor and tower were purple.
@@Nedski42YT It took me some time to find which one it was that they had in our labs but if you google Sun Blade 1000 that was the one; the sides are blue but to me the front always looked purple but i have been called colorblind by many exgirlfriends lol.
@@Nedski42YT $990??!! That is insane! We had a lab of about 50 of them, that is crazy, I thought it was unusual that we used them and I knew they were rare back then, It was some version of Linux that they were running but I can't remember which one.
Scott McNeely(Sun CEO): James? James Gosling: Yeah? Scott: James? Can you please quit messing around with that editor and get in here? I've got a new assignment! James: Okay boss, what is it now? Scott: "James, I'd like you to do a new version of C++. Make it compile into p-codes like the UCSD p-system. Give it a massive o-o library, why don't you borrow a bunch of stuff from Smalltakl? Oh and I almost forgot, please, drop the unsigned integers. I always hated using those back in the day when I was a C programmer." James: Okay boss, give me 10 years, 1000 programmers, and $40B and I'll be right back ...
@NRGY There is a reason why NSF completely BANNED computer language research in academia in the mid-1980's. It was folly. A bunch of people reinventing the same wheel.
Being from Linlithgow Scotland where Sun employed 800 people and had a manufacturing plant, it had a big effect on the town when they closed the manufacturing plant. There is now a smaller presence at the site with Oracle.
“Billion Dollar code” a Netflix series I saw mentioned Sun Microsystems and its CEO/Founder (I think) and some German software developers with a lawsuit over creation of google maps. Good movie
No, that was actually SGI (Silicon Graphics), which were close to Sun in some respects but also totally different. Both companies definitely were considered cool from what I was told.
Excellent video. Sun could have become what Microsoft became, and the world would be richer for it. It happened around 1988, when Sun introduced the 386i workstation. At this point, the Intel 80386 was far ahead of MS-DOS, and the PC world was desperate for a true multi-tasking, multi-process operating system. Windows 1 was a joke, and IBM's OS/2 was not ready. Power users would have several PCs at one desk, with a bank of monitors and various mazes of wiring to switch keyboards, etc. Now came the 386i, not only with a superb monitor and graphics wuth the SunOS Unix operating system, but able to run many DOS programmes in their own windows! This ability to have it all on one screen instantly made the 386i a hit with stockbrokers. Now, this was all a bit much for the 80386, and people were not used to the delayed latency of a multiprocess computer, but the 486 was already being tested in the next version. The hardware was nice, with higher end workstation features such as SCSI peripherals and the excellent monitor, but it was not PC-compatible (i.e. could not boot MS-DOS, though it ran DOS in subprocesses), even though Compaq had set the standard forever with its open ISA-based Desk Pro 386, thereby burying IBM's proprietary Microchannel bus (technically superior but royalty-laden). The 386i was years ahead of Microsoft, and had Sun gone for the general PC market, they could have taken it all. Only NT/Windows 2000 a dozen years later could compare. The NIS "Yellow Pages" administration system was being improved, and with the 486, a very attractive combination, offered at lower cost, would have captured the whole market. I bought a 386i in 1988, and two of my colleagues followed suit. It was my main computer for seven years, and served several years as a mail server. A student of mine did a massive computional thesis on one. A P2-90 PC running Linux eventually replaced my 386i. Being able to do both Unix and MS-DOS on the same machine was a godsend. Instead, Sun cancelled the 386i in 1989, after just one iteration; they also cancelled the Motorola-based workstations, to concentrate on the higher margin SPARC CPU and the professional engineering workstation market (and later servers). Admittedly, a Sparcstation motherboard was a work of minimal art compared with the always messy PC architecture. Needless to say, the PCs soon took over the workstation market, and Linux did the rest, but there was a gap of about six years between the 386i emerging and Linux becoming competitive. There was no universally available Unix for X86 PC's, only expensive, inferior products such as Xenix which had no traction in the general market. Sun could have filled that vacuum long before Linus took on the job. Sun's failure to take Microsoft and IBM head-on for the whole shebang is one of the greatest missed opportunities in history. General computing would be so much more advanced now had Sun succeeded and not Microsoft.
Sadly the group who came up with the 386i supposedly didn't really have the companies blessing, so it was starved for support and funds. Sun mgmt was both fickle, and devoted to Scott's cult of personality.
Sun engineering vet here during the boom times. 1988 - 1998. One heck of a rocket ride living through the growth and transition from workstations to enterprise cluster servers. Could see the end coming when Scott The Mouth was spending 8 out of 7 days on the golf course. Lack of leadership is what did Sun in. Great piece, thank you!
That's funny, I thought it was the PC bucket garbage undercutting the Sun Microsystems hardware and everyone's unwillingness to admit it, and design SPARC-based hardware which would undercut the PC-bucket "servers". I told them they'd go out of business in a year if they didn't, but nobody would accept the truth. Eight months later, Sun was sold to Oracle.
The original Stanford SUN CPU boards designed by Andy Bechtolsheim was the original CPU board for Cisco routers. Andy's CPU board design was also used in many time sharing UNIX mini-computers back in the mid 80's.
I grew up on Sun in the early 90s. They had really great tech, but unfortunately Intel eclipsed the SPARC in performance, and Linux made the high cost of Sun hardware hard to justify.
Thank you, I really enjoyed this video on a couple fronts. Apollo was the first start up I worked for and it was a short (1984-1986), but amazing ride that gave me the desire to hit a couple more start-ups throughout my career. It was also my first experience in wide spread layoffs and seemed so odd at the time after such fast paced growth. I worked at the Billerica campus and to walk through the manufacturing floor and see huge empty spaces and empty worktables after seeing them full of employees was heartbreaking. Those of us that remained were allowed to pick through some of the manufacturing floor tools and I still use a Fluke multi-meter and Weller soldering station to this day. Post Apollo I had the opportunity to work with Sun workstations and after using Apollo and it was an interesting contrast. Those were some great years I will never forget!
I worked in the Milpitas manufacturing building then the Sparc development group from 97 to 04. It was the best job I ever had and the 90s were a true boom time. Huge Christmas parties, quarterly outings, white water rafting, Napa wine train. And tons of Tshirts! But boy did that come to an abrupt halt. Fun while it lasted.
1:24 There was some controversy over describing the original 68000 processor as “32-bit”, since this was seen by some as marketing exaggeration. I think the best way to describe it was as a cut-down 32-bit design. When the first true 32-bit member of the family, the 68020, was released, you could see that the 32-bit extensions were mostly just a matter of filling in gaps in the original implementation. Unlike certain other vendors, whose 32-bit chips involving sticking unsightly architectural bags on the side of their older 16-bit designs.
As was briefly mentioned, Xerox didn't care and gave its permission for most of these inventions to be commercialized by others. Xerox was solely focused on the lucrative enterprise markets where they already had dominance. You could call it short sighted, but I don't think everything would have been as successful as it was if it was kept by Xerox. Xerox knew it's limitations and was risk adverse. Did Xerox leave money on the table, sure. They also didn't kill the technologies in their bureaucracy.
The original design goal of the Sun workstation was to build a 3M machine for CAD: 1 MIPS, 1 Mb memory and a 1 megapixel display. The original Sun1 design met this goal and sold for 1 mega-penny ($10,000). But that was without any peripherals like a disk drive or Ethernet networking.
I liked how Sun managed to make money in 2001 and 2002 but then went back in time to 2001 so they could lose a lot of money. If only they had been able to commercialize that time travel technology. “Stanfurd” would have been very proud.
They wrote off some several billions worth and hugely unsuccessful Linux appliance company acquisition in 2003 from the books, hence the huge one time loss. Operationally they were close to zero loss/gain in that year.
@@andreybashkin9030 who said anything about 2003 🤪 Just kidding. Thank you for explaining that. My company at the time helped them implement the Java Learning Center. Nice people to work with.
YT cut half of my response off. Let me retry: Oracle deal was McNealy's response to IBM's hostile takeover offer to save Sun's technological legacy and perhaps a better deal for the employees. Oracle was at that time hurting Sun by treating each thread on Niagra line of CPUs as a separate core in their pricing model. Sun had MySQL under its wing. Not a happy relationship to begin with. It was still in many ways a better deal than the one from IBM. Sun killed SGI. "We are the dot in dot-com". Sun succumbed to deteriorating corporate culture under poor management. It was painful to watch. Yet, to the last day it was great company to work for.
I cut my teeth programming CAD applications on a VAX 11/780 back in the early 80s. I remember the Sun and Apollo workstation wars. I finally jumped on and bought some HP Unix workstations in 1991 to run COTS CAD application (Unigraphics). PCs were simply not capable for high end CAD work back then. It is amazing how far things have come since those old days.
I remember touring data centers in the late 90's & early 2000's and everything was enterprise Sun architecture as far as you could see in any direction. Floor after floor in each data center of nothing but Sun enterprise 42U racks packed with servers, storage and network stacks. Two years later going through the same places it was all cheap 1U high density Linux Red Hat clusters. The change was pretty dramatic but once IT leaders were convinced that open source was a safe move, they got out of Sun hardware and Solaris software as fast as they could
I once met a sysadmin guy from Goldman in early 2000s who told me the Linux thing is a toy and before long people will come to their sense and come back to Solaris, I wonder where he is now😂
@@murphytalk I think the last holdouts were banking and finance corps. Their information risk people were largely against open source but eventually that changed too. Combine that with Red Hat offering enterprise "support" and it all changed very quickly
I worked for Sun for 7 years in the earlier days in the field - absolutely, the most fun company to work for in my career. I was a manager of pre-sales people out in the field sales force. Was able to give people unexpected raises at times - like, 15%! And the first low cost, color workstation (3/60), had a lead time of six months - we couldn't make them fast enough. Scott McNealy created a great culture, and I got to meet Bill Joy and drive him around to customers. He was hung over from drinking too much tequila the night before, but he was just an amazing, brilliant person, and customers were mightily impressed. And at the same time, he was a nice guy, didn't seem too full of himself. I still think that if they could have worked with HP, DEC, and IBM, to standardize Unix and a window system, we would've had a better and more secure desktop experience than Windows a long time ago.
Thank you very much ! The video took me to 1991-92 when I worked on Solaris OS based Sun workstations. Unix was the benchmark OS at that and knowledge of ‘C’ language was a must. I wrote Image processing algorithms on Sun Microsystem…
The day the workstation market died, afaict, was the day SNL ASCI Red took the 1 Teraflop crown on TOP 500. Suddenly, it didn't matter how well-designed your chips were, they would never get the price/performance of commodity chips on the latest process node. Economies of scale were here, there was no arguing with it.
I remember fondly my first workstation when I joined my first ISP as a network engineer in 1996. The Ultra1 was a great tool, albeit not the most desktop/user friendly.
I still have mine with Creator 3D graphics. The company I worked for folded - they gave away the workstations. Unfortunately it won't run Solaris 10, and without a somewhat modern browser, it's kneecapped on functionality. I have it set up in my living room, power it up every so often just to see if it works.
I remember my first demo on a Sun workstation in 1985, of a new automated publishing system called Interleaf. It offered true WYSIWYG text editing and page composition under an X Windows interface designed (reportedly) by MIT linguists for intuitiveness. It was impressive. Over the next five years Interleaf would be eclipsed by Framemaker, Adobe, and Quark.
@@BroscoWankston Retired now. I was going through all my old papers and throwing them away things. Came across Zander's note scribble on a piece of paper.
Sun's interesting history is under reported and under told. Solaris was a very innovative OS. Cheaper and more feature rich than ibm AIX. SPARC was a very interesting hardware and I thought it was open source, this video says it is not. ZFS is still a relevant storage technology. NFS was shared offered to the competition as a standard, still used today in Unixes. If Sun had sharp elbows, imagine Oracle treating the Sun open source communities as maids: Openoffice, Opensolaris, Mysql. "The network is the computer" thinking like that was very avant garde in 1980. Bill Joy, that guy in the photos folks, is the original author of Vi editor. Xerox PARC had the future in their hands. They could have been huge. But Xerox management had no clue. Any Alan Kay speech is highly recommended. The real father of the tablet computer. Some say Solaris is/was more advanced than Linux, but I'm no real expert on Unix kernels.
Excellent video. I dreamed of getting a Sparc 5 or 20 back in the day. The Sparc processor is open source. en.m.wikipedia.org/wiki/OpenSPARC#:~:text=OpenSPARC%20is%20an%20open%2Dsource,microprocessor%2C%20the%20UltraSPARC%20T1%20processor. Similar to the IBM Power 9 processor. I’d love to hear a history of the absorption of DEC, SGI, et al to better understand the changing business winds at the time. How did all of these workstation manufacturers miss it? I know that HP pulled the plug on PA RISC and drank the Itanium cool aide. But what about the others? In other words why wasn’t there a greater move towards consolidating OS, peripherals, etc? Thanks for this video- well done as always.
I got into Solaris just a *bit* too late to experience their revolutionary NeWS system that delivered UI via Postscript. Imagine my surprise when years later I realize the pattern of a horizontal web architecture that delivers HTML + JS which then communicates with remote services is almost exactly the same model. A similarity that NeWS’s inventor James Gosling would later comment on himself.
Bill Joy wrote vi, but I believe it was an incomplete implementation of 'SE' (screen editor) documented in the tome 'Software Tools' by Kernighan and Plaugher. Software Tools gives the original specification of many of the UN*X utilities still in use today, and serves as a reminder of just how prolific the pioneers were. You could actually get source in a package referred to as the 'Georgia Tech Software Tools' well before the use of any version of UN*X outside of research. I first used them in the very early 1980s on multiple operating systems, none of them unix based. It was one of the first examples of the open source movement I ever saw.
A missing piece: Academics in mathy fields used SUN in late 90s and migrated to Apple because Apple was cheaper and a 1997 PowerPC was a fraction of the cost yet ran Mathematica just as fast as the contemporaneous lab/desktop SPARCs.
In their hay day they were the best. We used to joke that the NT in Windows NT stood for Nice Try. I had always loved sun servers and I always wanted to get my hands on the bigger ones. In 2005 I was on the team that did the largest install of Sun servers in Europe at commerzbank in Frankfurt Germany. I thought I had died and gone to heaven, lol. such a Great Project
Great video as always. Just a note on the pronunciation of Bechtolsheim - the three syllables in German are separated like so: Bech-tols-heim, with “heim” meaning “home”. Therefore, the “sh” is not a single sound as in “shoe” but must be separated into “s” and “h”.
In the later half of the 2k's, a friend's company would get full rack Sun systems in lots with other equipment they actually wanted. Apparently that was a tactic used by the wholesalers to get rid of junk. Friend's company had no use for them and nobody would buy them for less than "paying them to take it away as scrap". They offered them to me, but I had no use for anything other than the racks and back then I could buy already empty used server racks for around $50+"you come get it", which is a whole lot cheaper than trying to figure out what to do with multiple dozens of very old SPARC systems.
Former Boeing Everett IT Support... We had Sun Microsystems, Silicone Graphics, and IBM Catia. The Triple 7 (777) was designed and built without prototype physical models. It greatly reduced the design cycle and 777 was an instant hit. We also had; Windows, Mac, OS2, and a few others just to keep our techs busy. Close to 50,000 devices on our network in Everett. Engineering, Factory Production, Customer Relations, and dozens of other Departments spread across a very large campus. Our team provided end user support and had over 350 people.
Sun Microsystems lost track of their customers motivations, and that was fatal. In the early 1990's they: 1) failed to recognize that their systems were being used as routers. Sun workstations with 4 port Ethernet cards in every slot. If they HAD realized this they either would have acquired Cisco (both were headquartered less than a mile apart in Menlo Park in 1993) or they would have become Cisco; 2) they could have bought Amdahl's Key Computer Labs, which had a completed mainframe ready for production in 1993). Instead, the same Amdahl team built the Cray mainframe Sun eventually acquired years later; 3) in 1999 most of Sun's sales were due to companies' back up plans to deal with the infamous Microsoft Y2K bug, with Unix OS; 4) Sun tried to undermine Microsoft by offering a fully compatible alternative to Microsoft Office called Star Office. Microsoft retaliated by funding Linux development to undermine Sun's Solaris. 5) Intel had far more processor sales than Sun, and could fund much more rapid semiconductor advances than TI and other SPARC processor foundries could keep up with.
TSMC Arizona plant. The Taiwanese engineer are paid less than their white counter part and needed to work overnight shift while the white engineers don't. This reminds me of 1863 to 1869, roughly 15,000 Chinese workers helped build the transcontinental railroad. They were paid less than White American workers. The Asian engineers are looked upon as less worthy than their white so called colleagues. History is repeating itself. Racism is alive and well in USA.
StarOffice was not originally Sun product, Sun bought it and made it into Openoffice. And that is what I use right now as we speak, when I need Microsoft Office type docs, under Libreoffice name (executable is still called soffice)
While I was at Sun Quentin one day Scott M. was talking to a tour of students who were visiting. One student asked him if he could change one thing, what would it be. Scott was (for a change) honest and admitted that he screwed up by not seeing the market for routers. But then Sun had lost interest in The Network Is the Computer (remember Network the dog?) some years before that.
Dude...... I grew up in Redmond in the '90s and Early '00s, stepped DEEP in Microsoft and Sub-California tech culture. This video from 10:45 to 13:15 explained SO MANY questions that have been festering in my mind since I was a little tot, seeing all the big-name computer brands rapidly come-and-go in school, home, work, and stores. Fascinating.... man, Sun screwed up. A tale as old a time.
At Rice University in the late 80's they were setting up Sun workstations everywhere. Big beautiful monitors and an OS that was pretty friendly. Over time some Apple machines moved in next to them, and these were junky by comparison, with very picky and nuanced methods for doing anything, it seemed to me. The Suns were really luxury items in that day.
Worked with one of the suits in this saga (and my wife was for a time at Motorola). This is the best concise account I have come across! Nicely done! 😎✌🏼
Working at Sun I got to meet Jonathan Schwartz and Scott McNeely. I got to meet the founders of Google. I got to work with some really smart people. In building 18 we had a computer closet that was filled with historical equipment. We had original Altos computers from the HP PARC facility. It was so much fun working there.
Excellent video covering all the major points accurately. I worked at Sun 2000-2011 and even in the declining years it was a great place to work. They treated employees well and and had great openness internally. A follow-up video could be a list of all the tech we now use and take for granted that started at Sun but was far too ahead of its time to be commercially viable when released.
I touched a Sun workstation once. It was a diskless workstation that booted across the network from another computer. I had never seen or even heard of that way of booting before, though subsequently I've spent a lot of time setting up and using diskless workstations on various platforms. After that first encounter, I never even saw another Sun product again. I started computing with 8 bit systems that ran proprietary OSes. The company I worked for, DHL, even made their own hardware, OS and application software for 8 bit and 16 bit hardware in the early 1980s. Then, they moved to Xenix on the Intel 80286, and NCR Unix on the NCR series of servers based on the Motorola 68020 and 68030 processors. Later they used HP-UX on the HP PA-RISC systems. Finally, when the Intel 80386 appeared, they used SCO UNIX as well. None of these systems were used with graphical workstations, just dumb Wyse50 terminals. So, I learned a LOT of command line computing. Later, I was exposed to Novell Netware, Windows 3.0/3.1/3.11/95, OS/2 2.1, Warp and eventually Linux. I've only recently started using FreeBSD on my home server. Sun certainly was a big player and very instrumental in getting the basic backbone of the internet started, but Linux, in particular Red Hat Linux completely wiped them out with their dirt cheap alternative. I also agree that IBM announcing that they had invested $1 billion in Linux, was a serious affirmation of how good Linux was. That sent shockwaves that affected both Sun and Microsoft. The availability of open source solutions has completely changed the landscape of computing forever. When Richard Stallman lauched his Free Software movement, I'll bet he never thought it would end up being used for surveillance capitalism.
Just a comment on the Cray acquisition. Sub did not acquire Cray. Cray was acquired by silicon graphics (a competitor). Cray had a division that developed sparc based systems and this division was sold to Sun by silicon graphics.
I remember inn 1991 when Eurotherm at their Reston campus got some Sun Workstations for software development. At that time we had Macs and PCs and the PCs were really only used in the lab and the Macs for documenting. This made the Macs the main information originators. The Macs were all networked and could share digital information straight forwardly. They also ran time sharing applications at the same time with real WYWYG What you see is what you get. The PCs were not the useful the Macs were and the Sun Workstation had a lot of characteristics with the Macs. They were not WYSWYG, but they have highly readable bitmapped fonts. The SPARC station was fast! Sunview Sun's window system had tool talk that was suppose allowed various parts of the windowing gadgets to talk to each other. I could cut and paste huge text logs using the windowing cut and paste at a snap while on a Mac the system might not respond for tens of seconds. Sun Workstation then went away from Sunview windowing that had been a system that was more of a advancement over the Macs with bends towards engineering and scientists to the the common desktop environment used my HP and IBM more a kin to MS Windows window system that made the Sun and other work stations seem more like a Windows PC than an engineering Macintosh. I think that was the death of the Sun Microsystems workstation. We could still use an advanced form of that Sunview windowing system today in 2022 for engineers and scientists running CAD and other scientific packages. I do remember the Daisy System workstation. It was used for electronic circuit design in 1992. We had one and I was made system administrator for the electrical engineering department which made me the one to give it attention. It had a screen that was so dim you could hardly see it and very slow to draw. I was told that was due to the high resolution screen it was using for its day. It was extremely slow and the only way to move digital information was on 5" floppy disk. If was the critical production flow of circuit design from EEs to the drafting department. We were having to schedule engineers to use the Daisy for schematic capture use. This was a bottleneck and inconvenient. I remember talking to software support for the Daisy and people at that company they were saying they thought they are going to make it and there was talk of Daisy abandoning its proprietary workstation and porting their software to a Sun Workstation. That was exciting, but Eurotherm was paying a lot for licensing the Daisy per year and only using it for schematic capture. It might have been $30,000 a year. By the time I was involved with the Daisy the discussion had changed keeping it to having me consult the various engineering teams for a replacement for schematic capture. The Daisy ran SPICE a mathematical common source library of mathematical models for electronic devices. Although SPICE was developed in the open Daisy's libraries were proprietary. Our engineering department didn't use the simulation part of the Daisy which may have been its most expensive part. The proprietary nature of a lot of the Daisy made it less appealing to Eurotheum's company uses both in Reston and in the UK. Eurotherm had huge layoff in fall 1992 which included me. Its very hard to retire old CAD systems like the Daisy and without an administrator to transition data, older designers may have never really left that old Daisy workstation. The company adopted a PC based schematic capture system which wasn't the Mac based version I was hoping for and a great deal of people preferred. I was told later the engineering team that was pushing for the PC based schematic capture system was really doing it with the hopes of getting another PC. We did everything on Macs, but they wanted another PC and decided to vote for the PC schematic capture system to get it. A couple weeks after that team got the new PC for schematic capture half of them got laid off.
Worked there a couple of times, and of the sector / type of work it was, it was certainly a unique experience, and a lot of that was embodied in the espirit d’corps, for want of a better term; not fanatical, not masochistic, but a lot of qualities one doesn’t expect or find in the field, humility, modesty, diligence, application, focus, dedication, and actually, out of all the places I have worked, one of the very few where I can honestly say I witnessed people contributing their intellect and articulating their view within the team regardless of rank, one of the few places that actually practiced the flat hierarchy, all are equal principle, which actually, many firms claim to operate according to too. It’s a very rare thing. I will say that was my experience from the people and teams I worked with, so I can’t claim this is as empirical globally within the company. I will say, a lot of people benefited significantly right across the shop floor, from how well they did, and the share options people were able to invest in.
I remember buying a copy of Solaris when they made a x86 version for around $100 or so, thinking I would be so cool to learn it. (please let me know if my memory is right!) Also recall buying a workstation with one of those optical mouse at one of those firesales. Not sure where it is now, but yeah, good memories. You can follow up with "Plan 9" OS and how that evolved to Google Cloud and what we have now days ... and where we might be going in the future.
Yes they brought out Solaris for x86 but it was rather crippled in that it lacked the catalogue of application software that the SPARC compiled version had and the performance delta was too much in favour of SPARC at the time.
@@BadatTanking the only thing it really lacked was Netscape Navigator. This was corrected later with Solaris 10. I still run Solaris 10 on i86pc in my private datacenter, and it is absolutely fabulous, I've weathered all the shit from GNU/Linux and FreeBSD thanks to sticking with Solaris. The software isn't a problem either, as I've built and packaged a vast library of the same software which runs on GNU/Linux. I'll be going to Tribblix on SPARC and SmartOS on intel soon, and with SmartOS, 15,000+ software packages will be available, in addition to having all the cool Solaris features like FMA, zones and ZFS, all running on intel hardware.
I used to work on all the home and office computers for Kleiner Perkins Dornan, Mayfield Fund, and Silverlake Partners. They were out on Sandhill Road in Menlo Park. I got to meet all the senior partners it was fun.
Thanks for the history lesson. My professional career spanned early SunOS thru Solaris servers. I moved to Linux support when linux finally became, what I called, Ready for Prime Time. I lost touch with Sun, coincidentally just as Sun was sold and I never looked back. Your review sounds spot-on. Sun Microsystems taught me a lot, and it was a good ride while it lasted. I built many good data centers based on Sun and I feel lucky to have been in the right place at the right time.
special thanks for a great 2022 season of Asianometry. Your page is one of only two we support via Patreon and your content is always topical, interesting, and educational. merry christmas!!
had worked for Sun support service before the bubble pops till 2003 ... One of the key things was high prices for cheap pc parts ... Had to replace lot of 9gb IDE drives in Ultra 5 and 10 ... Server side there were also issues with things such as the FC-AL centerplane , DC-DC converters Also at the software level we had to deal with 64 bits but had to run commercial software in 32. What was great is the ability to boot either from virtually any scsi media and the network was it from pre OBP systems to the OBP 4... My 2 cts
In the late 80's I worked at a division of a large electronics company in New York. The IT people and engineers at this electronics company were all degreed graduates from top universities. My previous computer experience was limited to PC's running DOS and Windows. I had used "inexpensive" CAD software for PCB design. At this big electronics I started doing PCB layout on Apollo, HP and Sun workstations using Mentor Graphics CAD software. I asked a few of the "smart" guys why these workstations were so much "better" than PC's with Intel CPU's. I got some odd answers. One engineer with a PHD in computer science pointed to the word "workstation" on his Sun computer and something like "It says "workstation," that's why it's better." A few of the IT experts said something like "PC's will NEVER be ready for primetime, Unix workstations will always be better." They said these sort of things all through the 90's. By the early 2000's many of the CAD software vendors had started porting their expensive tools to computers with Intel X86 CPU's running Windows NT/XP. Well, I guess they were right all along. Nobody remembers Windows or Intel based computers anymore. Or do they? ;-)
You helped me fill in an interesting blank and make the connection! I had been selling an SD-WAN product called Talari and then Oracle bought them. Oracle walked away from support of the new hardware Talari had introduced just before the purchase. Trying to force existing users to a new hardware platform. It was likely to move SUN boxes! Makes sense now!
An interesting story. In many ways reminiscent of DataPoint corporation. They had the first working network, distributed data processing, enterprise level laser printers, and many other technical innovations.Unfortunately much of the senior management kept reading their own press releases and really had their heads up their collective asses. In the late 80s they refused to consider "entrepreneurship" and instead paid talented technical people generous severance packages to start up their own companies. The USAF has a term for DPTs senior managements thought processes: up and locked.
The narrative on Cray and Sun misses the point that starting in the mid 1980's Sun's 50K work stations had taken the wind out of Cary's multi- million dollar super computers' sales (pun intended). Few engineers wanted to stand inline to share a Cray when their own personal Sun work station would give them the same answer... albeit a little slower. Departments that had supported Cray usage quickly sprinkling Suns around like hanging lampshades and a financially important component of Cary's sales disappeared literally overnight.
What people miss about the RISC revolution is that in the 1980s with Intel's 8086 and similar products, the increasingly complex CPUs of the day were using a technique called "microcoding", or a lower level instruction set inside the CPU to run instruction decoding, etc. It was assumed that the technique, inherited from mini and mainframe computers, would be the standard going forward, since companies like intel were increasing the number of instructions at a clip. RISC introduced the idea that if the instruction set were simplified, CPU designers could return to pure hardware designs, no microcode, and use that to retire most or all instructions in a single clock cycle. In short, what happened is the titanic turned on a dime: Intel dropped microcode like a hot rock and created pure hardware CPUs to show that any problem could be solved by throwing enough engineers at it. They did it by translating the CISC x86 instructions to an internal RISC form and deeply parallelizing the instruction flow, the so called "superscalar" revolution. In so doing they gave x86 new life for decades. I worked for SUN for a short time in the CPU division when they were going all in on multicore. The company was already almost in freefall. The Sparc design was flawed and the designers knew it. CEO Johnathan faced questioning at company meetings when he showed charts with Java "sales" presented as if it were a profit center (instead of given away for free). I came back to SUN again on contract after the Oracle merger. They had the same offices and the little Java mascots on their desks. It was probably telling that after my manager invited me to apply for a permanent position, I couldn't get it though their online hiring system, which was incredibly buggy, and then they went into a hiring freeze so it was irrelevant. I should also mention that not all companies did chip design in that era with SUN workstations. At Zilog we used racks full of MicroVaxes and Xwindow graphics terminals. I still have fond memories of laying out CPUs and chainsmoking in the late 1980s until midnight.
RISC vs CISC is about the bandwidth between the CPU and the memory. CISC (Originating with the DEC PDP11) was about fetching one instruction for as much data as you could. RISC Originating with the DEC PDP8) is about fetching smaller, but more of them, into a simpler, cheaper engine which can go faster, and could be developed at the pace of Moore's law. At first (in the days of magnetic core memory) CISC pulled ahead for obvious reasons. Later, semiconductor memory made caches possible - you could fetch a bunch of instructions form core memory into semi-conductor cache - possibly 10 times faster - better performance - if you paid for it. Motorola (CISC) could not reinvent their processors at the speed of RISC. Intel got ahead by superior advertising, not superior product. Purchasing in large organisations is not done by engineers and programmers. My grandmother was famous for saying "the adverts speak very highly of it" meaning that no one would speak highly of it who was not paid to do so. No one who had a choice ever switched from Unix to Windows.
The loss of Sun is a tragedy we haven't fully recovered from yet. The retreat from the workstation market left them without an on-ramp for SPARC/Solaris and confined it to niches and legacy system, most of which now run on x86 processors running Linux.
Precisely. It was giving up the workstation/desktop market that was really the beginning of the end for Sun. It just took a long time for Sun to bleed out.
As a Linux user for years, this is an interesting video about the history, it also explains what I saw during my childhood as I was using MS-DOS then Windows 3.1 then 95 to 7 before switching to Linux. It is interesting to see how scalable the CPUs are nowadays, nowadays we have more cores but only 2 socketed CPU per mainboard. That being said, I think you would refer to Threadrippers/Xeons as High end workstations than Mac Studio ^^"
Sun was a juggernaut and I was a proud evangelist in the burgeoning Ecommerce space in the mid-to-late 1990s. Where I felt they missed the boat was they failed to defend their workstation space and I emphatically felt they had an opportunity to acquire Apple to address that in the late 90s to early 2000s. Instead, they entrenched and pushed OpenOffice and the market wasn't buying. Cheap intel hardware and open source linux eventually killed their server space and Microsoft & Apple cleaned up the workstation space. The final insult to injury was Apple adopting a Unix standard without Sun and pivoting to consumer electronics to outlast them. I feel the video missed a key point in "the network is the machine". Sun was absolutely revolutionizing in this mantra and had they not lost sight of that, they might well have survived.
I programmed and supported Unix/Linux systems professionally from 1980 until 2021...Of the dozens of Unix/Linux hardware/software varieties that I worked with, I liked Sun the best (extremely stable)...I like to say that "Solaris put my daughters through college" and in my personal Linux-based home lab, I use ZFS to manage my storage...The largest server type I ever configured myself, was a set of Sun E6900s (perhaps $800K in ~2005 dollars , 1,200 lbs per server).
That whole "Unix era" was a cautionary tale in tech arrogance. Both the Unixae and Apple were too haughty to come down to the office secretary's level, while Microsoft went after the business world without hesitation, offering a full spectrum of software apps and OS for the business world. MS was for so long just pathetic quality-wise, but good enough to limp along. Literally, the Unix snobbery and Apple elitism made MS's poor quality all there was left. IBM tried with their desk OS, but too late and not hard enough. Linux has never taken off on the desktop because MS simply had too great a lead -- and eventually got Windows fairly respectable, reliable.
@@awuma Sometimes you don't have to be the best, just the most consistent. For all intents and purposes, MS is a stable, reliable product now, but not back in the 1990s. MS did strictly "show" bets back then, i.e., third-place was good enough. Meanwhile, the Unixae and Apple were chasing this or that idealistic or purist soap bubbles and totally forgot or mehhed the business world.
Back when Windows NT first came out, the US Government had a requirement that all its divisions buy “open systems”. This specifically included POSIX compatibility. So Windows NT initially shipped with a POSIX “personality layer”, just to satisfy the purchasing requirements. In fact, this was never really used in production, since all the customers were quite happy to run software specifically developed for NT’s own proprietary APIs. The internal modularity in Windows that allowed the creation of these alternative personality layers has long since bit-rotted away. Which is why their more recent attempt to emulate Linux APIs in WSL 1.0 didn’t work very well.
15:40 the chart labels the year 2001 twice. It should have 2003 after 2002, but it goes back to 2001 when showing the -3,429 (M) net income loss. The chart is also labeled 1999 to 2001 on the top, but also shows 2002 and what I assume should be 2003. Otherwise great video!
I was working at the Southern Alberta Institute of Technology in the late 1990s and we had put in a request for a lab full Sun Servers for a IT infrastructure course. Windows 98 was out and Windows NT just appeared. When we had our meeting with management, they had just come out of a presentation with Microsoft on Windows NT. None of the managers were IT, they were educators making decisions on which operating system we should use. They liked Windows NT because it looked like Windows 98 - they didn't have a clue about Sun. There was extreme pressure to go with NT. All of us in the IT field were rolling our eyes in disbelieve. Eventually the whole school migrated to Windows NT. We had Novell Netware running our email on one little Pentium 90 server when NT replaced it, it took 4 physical servers to run the email and it was slow.
Well done! Sun was my last Valley job and it was a terrific place to work. Anyone remember OpenStep? My first Valley job was for System Development Corporation. They did the code for the early rocket systems - straight assembly language, then JOVIAL. 50s nerds still working when I hit the scene. Legendary bunch to my mind. Ford Aerospace and IBM Fed Sys Div were all around the "Blue Cube" aka Onizuka AFB, were all the action was. SDC was bought by Burroughs later Unisys. The Blue Cube had computers that came out of submarines and were like 6bit or 10bit - some strange number. Next I worked for Metaware Inc. Frank DeRemer and Tom Pennello made the first 32 bit C compiler to hit the scene. Anyone remember when compilers and tools were actually sold? That was where I made my technical bones. Crazy place to work, so I moved over to Borland Int'l. MS buried Borland by giving away the MS spreadsheet and db, in true monopolistic fashion. Borland was also a crazy place, and going to Sun was sweet relief. Any of these outfits would be cool to hear about. SDC - truly historic. I'd add Xerox Parc to the top of the list - that is where many things began.
I hired on to a research university to run several server rooms in 1999. They had just purchased a bunch of Sun server hardware on the used market, and I had literally never logged into one (Novell, Windows). I soon discovered how good server hardware could actually be. At one point the servers running files, auth, web, and remote access had uptimes of over 4 years. When the Oracle purchase happened, we replaced all the then 12 year old Sun hardware with Redhat on x86 and VMWare, then got rid of VMWare for Redhat's virtualization when they drastically increased licensing costs. We went from 100% Sun to zero percent Sun in the space of a couple months - seems we were not alone.
I was a Unix SA starting in the mid-80s and worked with Sun systems from the late 80s. In my humble opinion there were ruined by ATT forcing them to move to SYS V which was known as Solaris. It was a huge step backwards and many companies avoided it for as long as possible. This opened up the way to competition that didn't have to be there. I jumped on Linux when it came out and have never looked back.
@@patrikfloding7985 Early versions were very buggy and performed terrible compared to SunOS. I worked supporting a trade floor and the attempts to migrate was unnecessarily painful. It didn't help that they completely changed so much so that SAs were often confused and frustrated. Of course over time it was improved greatly. But the damage that I saw as a customer, was done.
BSD vs SysV preferences totally depend on your history. Either path makes the other look bad. They both lost out by not merging support for the syntax differences!
@@marshallwilensky7932 One of the things I really liked about Linux or really Gnu is that it supported both variants for common sys admin tools like the ps command.
According to my memory files, Sun workstation cost more than twice what PCs cost in the mid 90s. Sure, Suns were more powerful... But with networking you could scale those cheap PCs... and regardless, the 'cheap' PCs caught up in power very fast. Thank-you Intel. Thank you Nvidia.. Than-you AAAAMMMMDDDDD!
I have fond memories of using a Sun workstation for CAD in Engineering Design class back in 1986. We used DOGS (Drawing Office Graphics System) with the optional BOXER solid modelling software. Wish they could make a comeback.
We still run old Sun servers on legacy machines at my job. There are no plans to replace them since they still do the job and we have plenty of spares. I get to use one every once in a while. The desktop (CDE I think) looks like it's from the 80s.
McNealy was an early advocate of the networked environment; his company's motto was "The Network is the Computer". At times, he has been known to be skeptical of products that do not integrate well with networked environments. One example McNealy has given involved the Apple iPod. As quoted in The Register, McNealy said, "There’s a pendulum thing where stuff is on the client side and then goes back into the network where it belongs. The answering machine put voicemail by the desk, and then it went back into the network. Your iPod is like your home answering machine. I guarantee you it will be hard to sell an iPod five or seven years from now when every cell phone can access your entire music library wherever you are."
Having grown up in the tech industry working initially for HP, this is a fascinating overview of Sun and the high tech industry! I had the privilege to hear Scott McNeally speak at an industry conference. Interesting enough Scott took a cheap shot at MicroSoft then followed up with a cheap shot at Intel. Sad but I get he felted slighted by the competition. I still have huge respect for what he accomplished in his he tech industry. Scott was ahead of his time.
I remember Scott McNealy, former CEO of Sun Microsystems, being in the cover of the Fortune magazine in 1994. That was the feature about their new programming language, Java ☕. Now, they are part of Oracle Systems (the same company that attempted to buy TikTok).
Apollo wasn’t shut down by HP, but it’s product lines adopted the IP through out the HP UX workstation class. You could say they lasted longer than SUN in that sense.
HP Apollo Domain was a migration path for organizations that had invested in Apollo boxes, to give them time to get onto the HP platform. I wasn't involved directly where I worked; but from what I recall hearing, converting a lot of the mission critical code base to be POSIX compliant was where a fair amount of effort was placed during the migration.
@@jonp4846 The V-class servers and Superdomes where a direct result of the Apollo acquisition and "evolution" of the acquired IP. Also the Apollo techies I worked with where legends.
@@xmaniac99 My first encounter with Apollos was at college in 1981. Initially it was pretty much just a room with a few terminals and teletypes, the punch card room, IBM mainframes and Pr1me minis. There were 2 Tektronix 4010s and a large pen plotter, but that was it as far as graphics devices go. The, in sophomore year, they put up a prefab building next to the computer center, bought 4 DN300s for the new lab, and off we went. Nothing but respect for the engineers that came before me and provided amazing machines like these.
Who remembers the massive amount of thefts of sun hardware? It was almost impossible to insure SPARC servers at one time. We had three racks of sun servers and the sun racks had to be secured inside a massive cage, hardened steel, special locks etc etc. Each cage weighed in at just under 1,000KG, had to be assembled onside by a team of four people. Some of the steel panels were about an inch thick. Needless to say the sun racks are long gone but the cages remain in our datacentre as nobody wants to take on the disassembly of them.
Any old Silicon Valley companies I should consider taking a look at?
I think all the well-known ones have been thoroughly covered by this point.
I can think of one or two obscure ones. For example, 3D workstation companies Stellar and Ardent were acquired by the same investors, who forced them to merge into Stardent--did they manage to achieve anything interesting before fading into oblivion? Then there were Alias and Wavefront, who also combined together. You don’t hear much about these any more (though the memory of Alias|Wavefront lives on in the “.obj” format, still popular for 3D data interchange, because of its sheer simplicity).
i would suggest SGI (silicon graphics international)
SCO?
Motorola Semiconductors, the original Hewlett-Packard, and ISC Systems (out of Spokane, Washington).
I'd suggest Silicon Graphics
Sun vet here, ‘99 to 2004. The video doesn’t really get what happened. First, the “sharp elbows” were replaced around the time I got there. We were winning for many reasons, but two of them were the safe and solid OS, Solaris, and partnerships with industry Allies: Sun, Veritas, and Cisco. The performance advantages were still there, and winning deals was pretty easy. Unfortunately, hiring enough people was getting tough, and the culture was changing. Lower to middle management got taken over by a bunch of political parasites many who had learned survival tactics at DEC and Compaq. When things turned, these people completely quit working on keeping the company afloat, and started scheming to be the last thrown overboard as there really wasn’t any place left to go. Scott had seemingly lost control of the culture. Zander, who had been a decent COO totally threw Sun under the bus as he departed for Motorola, where he showed his inability to be a CEO which then surprised no one at Sun. Morale crashed, even though we still had an excellent platform. Perhaps there was no way forward, but it seemed to me the ingredients were there for a better solution than the world ended up with. Open systems are too open, and thieves reach right in.
You sound awfully bitter about the move to open software. By all accounts, the open systems that replaced you were more stable, more compatible, and drastically more user friendly.
@@jrshaul I’m bitter, and it’s not just about open systems. Open systems have lots of advantages, but none of them are stability (quite the opposite. I had customers with systems that had been forgotten about because they had simply run for YEARS without attention), and this has nothing to do with user friendliness which isn’t a factor for operating systems, and has not been for thirty years. Most users never see the OS anymore.
I’m also bitter about all sorts of things mostly thrust on us by an education system that teaches amazing amounts of trash without bothering to teach the importance of prioritizing value over price.
@@nunyabidness3075 Do you have any long form rants? I'm down for some "elder" perspective.
@@Baronvonbadguy3 Sure. Name a subject. 🤣😂🤣😂
You know who else had rants? Scott McNealy. No kidding. He used to do a podcast for the employees, and he would do a rant on each one. One time, he went off on the money the company wasted buying vegetarian pizzas which were mostly thrown away. It was hilarious.
Unfortunately, I’m not blessed with his talent for humor. Sorry.
@@jrshaul > You sound awfully bitter about the move to open software.
Open Software killed off a lot of companies. Nobody ever thought that making a free version of UNIX was a sound business case for the companies that made proprietary HW and OSs. How they got away with copying and mimicking the UNIX code and didn't get sued is beyond me.
As a former Sun reseller and support contractor, I'm very happy for you covering my favorite company for your christmas video. Thanks!
I'm still working in Solaris support today and have a large collections of systems around. So sad that Oracle has ruined what was once a great company.. but that's what happens when a software guy buys a hardware company and has no idea what they actually do or how to run them. When Ellison swooped in, the first he did was double the system prices and triple the support contract fees, and struck down any special deals that Sun had going with their largest customers for decades. This lead to most of those customers taking the jump over to Linux clusters, that were just about becoming mainstream enough at that time.
Was it that Ellison is a software guy, or simply that he's a greedy b****rd who makes decisions motivated by short-term profit (or delusions thereof) and that alienate the customer? Seemed to me it was the latter, though being a software guy I could be biased :)
Oracle is the Electronic Arts of the Open Source world, buy up and ruin stuff basically.
I wonder if killing many of Sun technologies was Oracle's plan all along. Afraid that some of the technologies Sun had could pose threat to Oracle dominance in the enterprise markets.
oracle ruins everything
But somewhere you seem to not realize that SUN was not profitable any longer when Oracle acquired it, cannot blame Oracle for that. Engineers seem to believe that there is an unlimited amount of money and that companies do not have to make profit.
I was blessed to work for sun early in my career from 1988 to 1991 as a sales engineer. We called ourselves systems engineers at that time. What a learning experience that was: The people, the technology, the momentum, and the possibilities. As an engineer, I was amazed that a unix system could have an uptime greater than a week without requiring a reboot. Like anything, the growth was too fast and the company changed in spirit in the mid to late 90s. Great summary of Sun!
My home Unix (Linux) machine had once uptime of 6 years, from 2001 to 2007
To this day, I'm always going for 4-digit uptimes.
As a manager for a fairly large IT department, I admonished my guys to fix problems, with my line "Restarting is for losers."
Unfortunately, OS uptime seems to be going the way of the dodo. Containers seem to be helping but all of the patch restarts today are really frustrating. I remember when I would only have to patch/reboot an OS once or twice a year if that.
@dražen CRO The Network Is The Computer
Miss Solaris and SPARC. I had the pleasure of using the systems for about 8 years.
@dražen ZAGREB "The Network is the Computer" - "mreža je računalo". Servus!
funny story about a Sun Micro Mainframe. My father started working as a Junior accountant for a Large Queensland meat processing abattoir. they had a sun computer system. when costing the production of sausages by the ton, different wholesalers order them with different ingredients... and these had to be coasted down to the 4 decimal point. and this computer was worked by terminal and punch card. and a salesman would send through ingredient list and volume etc and an hour later they would get a printout handed back to them, and then they would get back to the client with a cost often a 2 hour info round trip. well dad bought a TRS80-4k Level 1 computer when they first came out, supposed to be for me. he wrote a basic program that would work out these costings almost instantly (well a few seconds) so he could talk to the sales person while on the phone. and the next month Byte magazine came out with a basic program that would turn such data into a pie chart. one day day took the computer to work and put it on his desk, and was doing the costings instantly, when the Managing director walked past, and asked what he was doing. dad put on a demo, including a pie chart, and the MD asked the Mainframe guy to make a pie chart for next weeks board meeting. and sure enough next week at the boardmeeting, the Heda of computing showed his pie chart, and dad demonstrated the little TRS-80. and the CEO asked the main frame guy aput him makingthe pie cahrtas, and he said they they had to purchase Sun's special program for pie charst that cost $15K (remember this was the late 1970's) so the MD stood up and said to the main frame guy "your fired" and had the Main frame sent to the tip and had TRs80 put on every salemans desk, and dad was promoted to head to sales... (dad made several changed to production, that had him promoted to 2IC in a few years)...
Sun was founded in early 1980's. Your story happened in 1970's. It must be another company's computer.
A great general often introduce new weapons to the battle field. 👏
Kudos for your dad! But felt sorry for the mainframe guy!
Sun never made mainframes. They made minicomputers.
@@johancoetzer2165 ACTUALLY... Sun did make mainframes towards the end, right before being bought up by Oracle: the M-line. The design was done by Sun and manufactured by Fujitsu. The hardware is completely redundant and can be partitioned, just like IBM's zSeries mainframes, additional processors can be enabled temporarily or permanently with a license... The only difference is, this mainframe hardware is not running z/VM and z/OS, but Solaris 10 or 11. 10 in my case. M3000 was the first, code named "Teraya" (I own one privately), followed by other mainframe models like the M10. These were all made after 2007.
I had a summer job at Sun in 1996, it was an absolute dream, the epitome of cool for a computing student job. Having that on my CV opened so many doors. I miss them.
Steve Jobs once called Xerox management "copierheads", because they had the Holy Grail but weren't interested in anything beyond copiers.
I miss Sun. I still have a couple SPARCstations sitting in the spare bedroom, a SPARC 20 and an Ultra 5. I haven't powered them up in years, though. I was given the 20 and I bought the Ultra 5 new, way back in the 1990s. They, and the IPC and IPX I used to have, made a nice little home network.
For years while maintaining Sun servers in NJ/CT/NY, I traded/bought used Sun parts from a company in New Jersey...I once spent 2 days upgrading two Sun E6800s by hand (3 quad CPU boards and four 8-card PCI "I/O boats" each). The CPU boards had 32 DIMMs per board (96 per server, 192 DIMMs over both servers)...I had bought 6 newer CPU board and new I/O cards. But I needed to transfer my RAM from the old CPU boards to the newer CPU boards before handing the old boards in as part of the trade up. So 192 DIMMs out, 192 DIMMs back in.
@@carlschumacher5510 Weren't these sun hardware incredibly expensive, and hard to maintain? Didn't they fail all the time? Those sun E series servers were designed wrong. Or that they never notified anyone how to properly mount them. The E4000 and up units had the air intake on the right side and exhaust heat to the left side instead of normal servers that do front to rear. So if you end up putting 2 E4500s next to one another then the one on the left would be constantly hot. Stupid design, and add to the fact that they never told anyone on how to properly mount them. Our company had these and they were always hot. I was a datacenter tech told to look into it. but the company had onsite sun techs to replace parts and they never said anything.
@@xcen1 From 1994 thru 2012 I worked with many models of Sun boxes (from ~$5K up to ~$800K models) but only a handful for E4000/E4500s...Every now and then I'd run into a popular Sun model that had a major flaw I had to work around. The E450 was a perhaps 10U very deep model that I don't think was meant to be racked (it had wheels iirc). The E450 was a model that could have 1 to 4 CPUs, one or two 10-drive bays I think, and various levels of RAM. It had a huge motherboard mounted vertically in iirc the middle of the chassis. Said MB was not supported well (it was easy to crack when installing parts). I wasn't the only one killing said motherboards, the local Sun techs were also...My solution?: I convinced management to let me only order fully-loaded E450s (all CPU & RAM slots filled, both drive cages installed, all the PCI cards I'd ever want). This was during the Doc Com era.
In the early 2000, working with E6500s / E6800s / E6900s was a step up in Sun boxes (but not quite to the E10K). These were "I am a rack" frames that you kept for 5 - 7 years, adding CPU/RAM boards over time (for example as more Oracle databases were added). The E6800/E6900 were true monsters. They had two internal power planes, the OOB management computer was actually a cluster of two small computers (PowerPC based I think). The beast was a real friend of anyone selling electricity. Over the years, as the CPUs got a bit faster, you could add/upgrade any of the 6 CPU board slots (each board holding 4 CPUs)...The coolest thing I have EVER done to a production server, was replacing a CPU board in an E6800 while it was up and serving requests (you could tell the OS to move processes off of a given CPU board).
One last comment on user expectations of Sun gear, during my E6800/E6900 period (2000 - 2007). Whenever I wanted downtime (patch, reboot, replace a failed part, etc) everyone "Why Carl? Why? I thought these things were bullet proof? Why?" (sigh!)...The Windows guys just had to say "Patch Tuesday". Zero objections...I remember almost hitting a clock bug (some issue after 520 days of uptime in the OOB server cluster (yeah, not a power of 2)) and having to get time slots to reboot the whole E6800/E6900 fleet.
@@xcen1 The 4500 was "cheap" for what it was in it's day. It was an NFS and mid-range data base server without an equal at its price point. The only real problem, as you pointed out, was the airflow design could lead to overheating and indeed fires. I worked at Sun from '96 to '99 personally saw one system catch fire and assisted another customer replacing a unit that had also caught fire. It wasn't intended to be a datacenter unit but rather a small office, stand alone design. The price performance was so good, however, that at one point that's all datacenter customers wanted to buy.
@@justinbukoski1 OMG no wonder sun went away. That's an ass backwards idea to say it was meant to be in office. Any server should be in datacenter environment for so many reasons 1 being a security issue. The E4000 machine weighs as much as a car. Me being a skinny kid in 2000 at 5'8" and 110 lbs I was able to lift up a HP DL380g1 by myself at that time, But this thing was so heavy 2 of us struggle to move it one time. After the first time i learned how to move it by removing everything inside, all the boards, and psu's then it was manageable to move. Jesus to put a server in an office, someone could walk by and spill coffee all over it. What a stupid idea..... And it has to be in a datacenter environment because it needed front and rear access. You're going to put it up against a wall in an office, then you'll move it to change and rear boards or psu. In a rack in a datacenter, you have open access to to rear and front.
When I came to study at my uni in 1994, all their services (DNS, mail etc.) ran on lone SPARCclassic workstation. Since they had no one familiar with Unix and I already knew Linux, they hired me to take care of that server and I eventually ended up working full time in their IT and this set up me for my career in networking. And it all started with that small Sun computer. Good memories.
i had a similar "career" 🙂
You were using Linux in '94?
That's like buying bitcoin when it was two bucks.
@@jrshaul back when you had to compile the kernel overnight 😂
That's how I got into the biz, so to speak. I got my hands on 8 floppies that just happened to have a Linux Slackware distro on it. I completely ditched Windows, quit college and got a job working in the field, using nothing more than a decent ability to handle the dry reading of countless manpages, coupled with knowing C. I worked as a Sun level II software support specialist for a couple of years, but lost that gig when Sun started contracting out level II support. I was an odd guy to have around, really useful, but odd. I needed help using a Windows work station, but if it ran some flavor of Unix, I was perfectly at home.
@@snorman1911All to find out you enabled the wrong driver. ;-) Good times! When the kernel went modular, that's when the overnight compile run ended.. mostly..
*_Former Boeing Everett..._*
Boeing used Sun Micro Systems, Silicone Graphics, IBM Catia, Windows NT, IBM OS2, Mac, and Windows 3.xx over a very large campus. I was part of 350 member IT Desktop Support Team in Everett. There were 50,000 devices total, including Workstations, Servers, CAD, Network Printers, and more, used by 30,000 employees in Everett. It was challenging supporting all the various desktop configurations. When Boeing bought McDonald Douglas, we inherited their mess too.
I transferred to Lean Manufacturing later. At that time Boeing was consolidating IT and standardizing Computing. IBM Catia was being upgraded to handle PC functions as part of Single Glass Initiative. There were different departments using 2-3 different types of workstations per desk. Single Glass solved part of that problem.
Boeing had it's own versions of Windows Operating Systems and Windows Office. Boeing had the largest network running Microsoft Products in the world. Everett was just 1 of many other campuses around the globe. Our Call Center in Bellevue handled a millions calls per year. IT was 24/7 365 a year. Boeing may have slowed down over the Holidays, but it never stopped providing world class support to all it's computer users.
its
your take on the McDonnell-Doug merger?
@@NoTraceOfSenseThanks for comment... Yes I was there during the merger. I was part of Lean Manufacturing and was sent to St Louis to iron out production kinks in the F/A-18 Super Hornets. Amazing experience.
Thank you for this trip down memory Lane. Working in the workstation marketplace in the late 1980’s you touched on many things that were formative parts of my career, and friendships I still have today. The best man at my wedding I met at HP, he left and we both ended up at a reseller. I few months in and he was poached by Apollo (for a crazy amount of money) and he ended up back at HP 😂. He left again. My oldest friend worked for ARM in its early days. I defected early to the Microsoft bandwagon as my interest was business not technical and I started working with Olivetti hardware (still with the occasional unix install at the start), so it was nice to see their name pop up. Always remember the Olivetti joke that it was going to be called Italian Business Machines but they didn’t think the initials would take off. But I’m rambling now - as us old folk tend to 😂.
I remember when engineering workstations were the most powerful stand alone desktop computers. They, at onetime were many time more powerful than any IBM PC (and its 386 clones) or Apple desktop computer. But, in time Moore's law changed everything, and consumer grade PCs eventually could run the most sophisticated CAD/EDA and simulation software, at a much lower cost. Sun and Apollo workstations had lost their edge. The last thing I remember from Sun was Java.
Yes. By the time the PII was released, the workstation was done. Started ProE on a PII.
Started ‘70 on 360, ‘77 U1100, ‘79 Σ7/Networked 360/PC/Lisa&II, ‘83 anything networked (local “PC Guru”), ‘84-present Cysec.
I worked on Sun workstations (hardware diagnostics programming) starting around 1990, and was amazed to find just a couple of years later that a 486 upgraded to a Pentium and running Linux ran compute-intensive tasks about twice as fast. Our workstations were 2-3 years old by then, but by 1995, the Linux PC was absolutely faster than all but the most expensive Sun workstations, and we stopped buying them.
Sun forgot their open roots, beginning a little with SunOS, going full retard with SPARC processors, and by then they had dug too deep a hole to ever get out of. The end of the 1990s was like a sad little joke as they kept on digging that hole.
I think hey could have kept going with SunOS and Solaris, wouldn't have been too hard a problem to re-open it like Linux. But Motorola botched the 68030 and 68040 advances, SPARC was a desperate gamble which paid off for a few years, but there was no going back, they had too much face invested to switch to Intel x86, and that was all she wrote.
(I can understand Sun shuddering at the x86 architecture. I wrote assembler for a lot of architectures, all were fun and interesting one way or the other, but x86 always gave me the creeps, about the ugliest architecture possible, and that didn't fit their image.)
They fell to the same arrogance has Digital Equipment (DEC): they refused to acknowledge the competition and to a combination of lowering the price and raising the performance.
Yeah, the PPro and successors ate old RISC's lunch, Windows NT/2K and Linux ate old UNIX, and when Nvidia released the og GeForce and Quadros they ate the pro 3D market. Same story in all cases, the low end market just becoming good enough to gradually eat the high end from under it.
I still miss SPARC/Solaris though, and those lovely purple boxes. Not so much the E450 that almost fell out of a rack and crushed me once.
I started working on Suns around 1990 and continued using them through grad school and beyond, up until about 2010. They were a great machine to develop on since Sun controlled both the hardware and the software and created a very integrated feel. At the time, Apple's OS was a mess of spaghetti code. Now I develop on a macbook pro (Apple long ago got smart and switched to unix) and while it's gotten much better, it's still not as smooth as Solaris was in its heyday.
I thought Apple's MacOS was built on Unix, which is true but I recently learnt that it's not the full story. Apparently the older gen Mac OS was replaced with an OS that was based on a Mach kernel (ending with h), and one of the BSD's stack was built on top of this Mach kernel. This OS was named OSX, which some time later became MacOS. You probably know this being an apple user. It's strange how there is more to the story than I thought. Similar to how Android is not GNU/Linux, but Java-like Virtual Machines built on the Linux kernel.
Its a mistake to call Mac os UNIC because Mac os cant do any mission critical task or multiuser doing multitask
Calling mac os unix is like calling a teenager a soldier because won a kungfu tournament
@@Teluric2 MacOS is Unix both in lineage and in Open group certification. What more could you want?
@@elumiomerk4013 System 7 to 8 was so broken that you would reboot 10x per day. Networking and dial up was broken, SCSI was bad. The things were slow and there were no graphics cards available. Shitbox by shitheads. G3 were probably the faultiest products shipped ever.
@@808bigisland what are you trying to say?
I really enjoyed this. I'd love to see a similar piece on Silicon Graphics.
Sun workstations with Solaris were marvelous systems and we used them extensively at my place of work. Their reliability was an asset in those days. I still remember fondly using the unix base system with Motif GUI.
One of Sun's most iconic lines to me was when they coined the phrase, "The Network Is The Computer." That still resonates today & tomorrow!
Sadly they lost sight of the importance of that in the late 90's.
Sun NFS was really bulletproof. I was there for the entire ride, despite starting in 1981 on the IBM PC. The early VME bus systems were good, the 386 based design was a flop, but the Sparc based systems were just incredible. We had a Sparc 2 as our NIS master that had an uptime of 4 years when it came time to upgrade for y2k. We deployed hundreds every year in the 1990s, every software developer had one on his desktop. Too bad they couldn't catch up with the latest process nodes, it was funny to hear about dual channel memory in the PC being such a big innovation when we had been installing quad channel memory for years.
The sysadmins at the company I worked for, said that NFS stood for No F’ing Security
NFS was designed to be “stateless”, which caused no end of problems for Unix software that was expecting a baseline of sane behaviour from the filesystem -- ironic for a Unix-centric company, don’t you think?
@@lawrencedoliveiro9104 I obviously had a much different experience. Compared to other vendors (and especially Linux) the Sun implementation of NFS "just worked". When they enabled TCP rather than UDP transmission, it just worked. Since our primary use was an NFS based application, I had a lot of experience in that regard.
@@lawrencedoliveiro9104 NFS had some issues, but having department-wide computer transparent file system in mid 90-s, was plain fantastic.
Literally, in first two years in my university department as a postdoc, I did not even know to what computer my home directory is physically attached. Compare that with Windows 95 paradigm. And I still use NFS among the three Linux machines I manage in my office now
Didn't Sun or Stanford license one of their first cards to three companies? Believe one was in Chicago, the other was Sun Micro, can't remember the third. I worked in applications my first two years out of college helping with MC68000 support for multiple customers. Supported Apollo, Apple, Sun, and many others.
The amount of Sun workstations that Nortel bought in the late 1990s and early 2000s you wouldn't believe, as many of the manufacturing rigs used them. When Nortel went into serious decline, their Sun workstation orders dried up very quickly as they had more than enough machines spare.
I had such a beast of a CRT monitor. made by Sun. My brother picked it up from a trashcan at UCSC. It was at least 1440p back when most lcd monitors came at 1080, and most tvs where at 720. It was at least 80lbs, and 27"
Lasted me so many years, had to get rid of it because of its the weight after I left college in 2013.
Sun was the largest buyer of Sony Trinitron tunes for a number of years, and got only the highest QC spec units
BORING DUDE
you should give it to me
I remember back in college in 2010 We used dated Sun Microsystems Desktop PCs in one of our classes; they were definitely unique. I remember the cases of both the monitor and tower were purple.
Purple? Didn't Sun computers have blue colored accents?
SGI computers definitely lots of purple!
@@Nedski42YT It took me some time to find which one it was that they had in our labs but if you google Sun Blade 1000 that was the one; the sides are blue but to me the front always looked purple but i have been called colorblind by many exgirlfriends lol.
@@belltolls1984 Yeah, I found an image and it looks purplish to me too. There is one for sale for $990 on eBay.
@@Nedski42YT Indigo!
@@Nedski42YT $990??!! That is insane! We had a lab of about 50 of them, that is crazy, I thought it was unusual that we used them and I knew they were rare back then, It was some version of Linux that they were running but I can't remember which one.
How about a video on Java? This is such an important part of Sun's legacy!
We had to program in J++ for our undergrad thesis in engineering.
@NRGY Good one! And not very well, either.
Scott McNeely(Sun CEO): James?
James Gosling: Yeah?
Scott: James? Can you please quit messing around with that editor and get in here? I've got a new assignment!
James: Okay boss, what is it now?
Scott: "James, I'd like you to do a new version of C++. Make it compile into p-codes like the UCSD p-system. Give it a massive o-o library, why don't you borrow a bunch of stuff from Smalltakl? Oh and I almost forgot, please, drop the unsigned integers. I always hated using those back in the day when I was a C programmer."
James: Okay boss, give me 10 years, 1000 programmers, and $40B and I'll be right back ...
@NRGY There is a reason why NSF completely BANNED computer language research in academia in the mid-1980's. It was folly. A bunch of people reinventing the same wheel.
Andy is a decent human being. Treats regular people well.
Being from Linlithgow Scotland where Sun employed 800 people and had a manufacturing plant, it had a big effect on the town when they closed the manufacturing plant. There is now a smaller presence at the site with Oracle.
“Billion Dollar code” a Netflix series I saw mentioned Sun Microsystems and its CEO/Founder (I think) and some German software developers with a lawsuit over creation of google maps. Good movie
No, that was actually SGI (Silicon Graphics), which were close to Sun in some respects but also totally different. Both companies definitely were considered cool from what I was told.
Excellent video. Sun could have become what Microsoft became, and the world would be richer for it. It happened around 1988, when Sun introduced the 386i workstation. At this point, the Intel 80386 was far ahead of MS-DOS, and the PC world was desperate for a true multi-tasking, multi-process operating system. Windows 1 was a joke, and IBM's OS/2 was not ready. Power users would have several PCs at one desk, with a bank of monitors and various mazes of wiring to switch keyboards, etc.
Now came the 386i, not only with a superb monitor and graphics wuth the SunOS Unix operating system, but able to run many DOS programmes in their own windows! This ability to have it all on one screen instantly made the 386i a hit with stockbrokers. Now, this was all a bit much for the 80386, and people were not used to the delayed latency of a multiprocess computer, but the 486 was already being tested in the next version. The hardware was nice, with higher end workstation features such as SCSI peripherals and the excellent monitor, but it was not PC-compatible (i.e. could not boot MS-DOS, though it ran DOS in subprocesses), even though Compaq had set the standard forever with its open ISA-based Desk Pro 386, thereby burying IBM's proprietary Microchannel bus (technically superior but royalty-laden).
The 386i was years ahead of Microsoft, and had Sun gone for the general PC market, they could have taken it all. Only NT/Windows 2000 a dozen years later could compare. The NIS "Yellow Pages" administration system was being improved, and with the 486, a very attractive combination, offered at lower cost, would have captured the whole market. I bought a 386i in 1988, and two of my colleagues followed suit. It was my main computer for seven years, and served several years as a mail server. A student of mine did a massive computional thesis on one. A P2-90 PC running Linux eventually replaced my 386i. Being able to do both Unix and MS-DOS on the same machine was a godsend.
Instead, Sun cancelled the 386i in 1989, after just one iteration; they also cancelled the Motorola-based workstations, to concentrate on the higher margin SPARC CPU and the professional engineering workstation market (and later servers). Admittedly, a Sparcstation motherboard was a work of minimal art compared with the always messy PC architecture. Needless to say, the PCs soon took over the workstation market, and Linux did the rest, but there was a gap of about six years between the 386i emerging and Linux becoming competitive. There was no universally available Unix for X86 PC's, only expensive, inferior products such as Xenix which had no traction in the general market. Sun could have filled that vacuum long before Linus took on the job. Sun's failure to take Microsoft and IBM head-on for the whole shebang is one of the greatest missed opportunities in history. General computing would be so much more advanced now had Sun succeeded and not Microsoft.
Sadly the group who came up with the 386i supposedly didn't really have the companies blessing, so it was starved for support and funds.
Sun mgmt was both fickle, and devoted to Scott's cult of personality.
Was IBM planning to give PC DOS to Sun?
Sun engineering vet here during the boom times. 1988 - 1998. One heck of a rocket ride living through the growth and transition from workstations to enterprise cluster servers. Could see the end coming when Scott The Mouth was spending 8 out of 7 days on the golf course. Lack of leadership is what did Sun in. Great piece, thank you!
That's funny, I thought it was the PC bucket garbage undercutting the Sun Microsystems hardware and everyone's unwillingness to admit it, and design SPARC-based hardware which would undercut the PC-bucket "servers". I told them they'd go out of business in a year if they didn't, but nobody would accept the truth. Eight months later, Sun was sold to Oracle.
The original Stanford SUN CPU boards designed by Andy Bechtolsheim was the original CPU board for Cisco routers. Andy's CPU board design was also used in many time sharing UNIX mini-computers back in the mid 80's.
I grew up on Sun in the early 90s. They had really great tech, but unfortunately Intel eclipsed the SPARC in performance, and Linux made the high cost of Sun hardware hard to justify.
Thank you, I really enjoyed this video on a couple fronts. Apollo was the first start up I worked for and it was a short (1984-1986), but amazing ride that gave me the desire to hit a couple more start-ups throughout my career. It was also my first experience in wide spread layoffs and seemed so odd at the time after such fast paced growth. I worked at the Billerica campus and to walk through the manufacturing floor and see huge empty spaces and empty worktables after seeing them full of employees was heartbreaking. Those of us that remained were allowed to pick through some of the manufacturing floor tools and I still use a Fluke multi-meter and Weller soldering station to this day. Post Apollo I had the opportunity to work with Sun workstations and after using Apollo and it was an interesting contrast. Those were some great years I will never forget!
I worked in the Milpitas manufacturing building then the Sparc development group from 97 to 04. It was the best job I ever had and the 90s were a true boom time. Huge Christmas parties, quarterly outings, white water rafting, Napa wine train. And tons of Tshirts! But boy did that come to an abrupt halt. Fun while it lasted.
1:24 There was some controversy over describing the original 68000 processor as “32-bit”, since this was seen by some as marketing exaggeration.
I think the best way to describe it was as a cut-down 32-bit design. When the first true 32-bit member of the family, the 68020, was released, you could see that the 32-bit extensions were mostly just a matter of filling in gaps in the original implementation.
Unlike certain other vendors, whose 32-bit chips involving sticking unsightly architectural bags on the side of their older 16-bit designs.
Another lost opportunity by Xerox Parc that some other person/company successfully commercialized
Couldn't help but think of that😂
As was briefly mentioned, Xerox didn't care and gave its permission for most of these inventions to be commercialized by others. Xerox was solely focused on the lucrative enterprise markets where they already had dominance. You could call it short sighted, but I don't think everything would have been as successful as it was if it was kept by Xerox. Xerox knew it's limitations and was risk adverse. Did Xerox leave money on the table, sure. They also didn't kill the technologies in their bureaucracy.
I think Xerox got Apple stock for letting Jobs use some of their technology. If they had held on to that until now, imagine what that’d be worth.
The original design goal of the Sun workstation was to build a 3M machine for CAD: 1 MIPS, 1 Mb memory and a 1 megapixel display. The original Sun1 design met this goal and sold for 1 mega-penny ($10,000). But that was without any peripherals like a disk drive or Ethernet networking.
A whole 1MB? That's just crazy! Who'd ever need such a thing?
I liked how Sun managed to make money in 2001 and 2002 but then went back in time to 2001 so they could lose a lot of money.
If only they had been able to commercialize that time travel technology. “Stanfurd” would have been very proud.
They wrote off some several billions worth and hugely unsuccessful Linux appliance company acquisition in 2003 from the books, hence the huge one time loss. Operationally they were close to zero loss/gain in that year.
@@andreybashkin9030 who said anything about 2003 🤪
Just kidding. Thank you for explaining that.
My company at the time helped them implement the Java Learning Center. Nice people to work with.
YT cut half of my response off. Let me retry:
Oracle deal was McNealy's response to IBM's hostile takeover offer to save Sun's technological legacy and perhaps a better deal for the employees.
Oracle was at that time hurting Sun by treating each thread on Niagra line of CPUs as a separate core in their pricing model. Sun had MySQL under its wing. Not a happy relationship to begin with. It was still in many ways a better deal than the one from IBM.
Sun killed SGI.
"We are the dot in dot-com".
Sun succumbed to deteriorating corporate culture under poor management. It was painful to watch. Yet, to the last day it was great company to work for.
I cut my teeth programming CAD applications on a VAX 11/780 back in the early 80s. I remember the Sun and Apollo workstation wars. I finally jumped on and bought some HP Unix workstations in 1991 to run COTS CAD application (Unigraphics). PCs were simply not capable for high end CAD work back then. It is amazing how far things have come since those old days.
I remember touring data centers in the late 90's & early 2000's and everything was enterprise Sun architecture as far as you could see in any direction. Floor after floor in each data center of nothing but Sun enterprise 42U racks packed with servers, storage and network stacks. Two years later going through the same places it was all cheap 1U high density Linux Red Hat clusters. The change was pretty dramatic but once IT leaders were convinced that open source was a safe move, they got out of Sun hardware and Solaris software as fast as they could
I once met a sysadmin guy from Goldman in early 2000s who told me the Linux thing is a toy and before long people will come to their sense and come back to Solaris, I wonder where he is now😂
@@murphytalk I think the last holdouts were banking and finance corps. Their information risk people were largely against open source but eventually that changed too. Combine that with Red Hat offering enterprise "support" and it all changed very quickly
I worked for Sun for 7 years in the earlier days in the field - absolutely, the most fun company to work for in my career. I was a manager of pre-sales people out in the field sales force. Was able to give people unexpected raises at times - like, 15%! And the first low cost, color workstation (3/60), had a lead time of six months - we couldn't make them fast enough. Scott McNealy created a great culture, and I got to meet Bill Joy and drive him around to customers. He was hung over from drinking too much tequila the night before, but he was just an amazing, brilliant person, and customers were mightily impressed. And at the same time, he was a nice guy, didn't seem too full of himself. I still think that if they could have worked with HP, DEC, and IBM, to standardize Unix and a window system, we would've had a better and more secure desktop experience than Windows a long time ago.
All of them did standardize on a windowing system - it was called "CDE", the Common Desktop Environment. Sooo... OY! Mate! What are you on about?
Thank you very much ! The video took me to 1991-92 when I worked on Solaris OS based Sun workstations. Unix was the benchmark OS at that and knowledge of ‘C’ language was a must. I wrote Image processing algorithms on Sun Microsystem…
Man, in the 90s, if you were a Unix wizard, you could write your own ticket. Nevermind Novell, Microsoft and Apple talk, Unix gurus were rock stars.
The day the workstation market died, afaict, was the day SNL ASCI Red took the 1 Teraflop crown on TOP 500. Suddenly, it didn't matter how well-designed your chips were, they would never get the price/performance of commodity chips on the latest process node. Economies of scale were here, there was no arguing with it.
No matter price performance when you have a chip that can run an OS(AIX) that run circles around competition OSes.
Retired it tech in 2022. Fun to watch this hi tech vids once a while.
I remember fondly my first workstation when I joined my first ISP as a network engineer in 1996. The Ultra1 was a great tool, albeit not the most desktop/user friendly.
I still have mine with Creator 3D graphics. The company I worked for folded - they gave away the workstations. Unfortunately it won't run Solaris 10, and without a somewhat modern browser, it's kneecapped on functionality. I have it set up in my living room, power it up every so often just to see if it works.
I remember my first demo on a Sun workstation in 1985, of a new automated publishing system called Interleaf. It offered true WYSIWYG text editing and page composition under an X Windows interface designed (reportedly) by MIT linguists for intuitiveness. It was impressive. Over the next five years Interleaf would be eclipsed by Framemaker, Adobe, and Quark.
I was the guy who did the report to Sun's Office of the President to buy Cray Research. Ed Zander sent me a personal Thank You note.
I thought you were the guy in the commercials for the hamburger restaurant
@@BroscoWankston Retired now. I was going through all my old papers and throwing them away things. Came across Zander's note scribble on a piece of paper.
Sun's interesting history is under reported and under told.
Solaris was a very innovative OS. Cheaper and more feature rich than ibm AIX.
SPARC was a very interesting hardware and I thought it was open source, this video says it is not.
ZFS is still a relevant storage technology.
NFS was shared offered to the competition as a standard, still used today in Unixes.
If Sun had sharp elbows, imagine Oracle treating the Sun open source communities as maids:
Openoffice, Opensolaris, Mysql.
"The network is the computer" thinking like that was very avant garde in 1980.
Bill Joy, that guy in the photos folks, is the original author of Vi editor.
Xerox PARC had the future in their hands. They could have been huge.
But Xerox management had no clue.
Any Alan Kay speech is highly recommended. The real father of the tablet computer.
Some say Solaris is/was more advanced than Linux, but I'm no real expert on Unix kernels.
Excellent video. I dreamed of getting a Sparc 5 or 20 back in the day. The Sparc processor is open source. en.m.wikipedia.org/wiki/OpenSPARC#:~:text=OpenSPARC%20is%20an%20open%2Dsource,microprocessor%2C%20the%20UltraSPARC%20T1%20processor. Similar to the IBM Power 9 processor.
I’d love to hear a history of the absorption of DEC, SGI, et al to better understand the changing business winds at the time. How did all of these workstation manufacturers miss it? I know that HP pulled the plug on PA RISC and drank the Itanium cool aide. But what about the others? In other words why wasn’t there a greater move towards consolidating OS, peripherals, etc? Thanks for this video- well done as always.
I knew Bill Joy sounded familiar. Vi/Vim has been installed on every one of my PC’s since ‘95.
I got into Solaris just a *bit* too late to experience their revolutionary NeWS system that delivered UI via Postscript. Imagine my surprise when years later I realize the pattern of a horizontal web architecture that delivers HTML + JS which then communicates with remote services is almost exactly the same model. A similarity that NeWS’s inventor James Gosling would later comment on himself.
I am sure Solaris was more advanced than Linux but that was a very long time ago.
Bill Joy wrote vi, but I believe it was an incomplete implementation of 'SE' (screen editor) documented in the tome 'Software Tools' by Kernighan and Plaugher. Software Tools gives the original specification of many of the UN*X utilities still in use today, and serves as a reminder of just how prolific the pioneers were. You could actually get source in a package referred to as the 'Georgia Tech Software Tools' well before the use of any version of UN*X outside of research. I first used them in the very early 1980s on multiple operating systems, none of them unix based. It was one of the first examples of the open source movement I ever saw.
"Stanfurd." Oof, shots fired.
A missing piece: Academics in mathy fields used SUN in late 90s and migrated to Apple because Apple was cheaper and a 1997 PowerPC was a fraction of the cost yet ran Mathematica just as fast as the contemporaneous lab/desktop SPARCs.
Apple has a tiny fraction of sun reliability. There s no way to replace sun with macs unless your workload is small
This is an excellent documentary. Sun couldn't have been described better so perfectly and informatively in ten minutes. Congratulations.
In their hay day they were the best. We used to joke that the NT in Windows NT stood for Nice Try. I had always loved sun servers and I always wanted to get my hands on the bigger ones. In 2005 I was on the team that did the largest install of Sun servers in Europe at commerzbank in Frankfurt Germany. I thought I had died and gone to heaven, lol. such a Great Project
Great video as always. Just a note on the pronunciation of Bechtolsheim - the three syllables in German are separated like so: Bech-tols-heim, with “heim” meaning “home”. Therefore, the “sh” is not a single sound as in “shoe” but must be separated into “s” and “h”.
That might be the case in Germany, but everyone at Sun said Bechtol-sheim.
@@wjcroft72 Fair enough, just thought I’d point it out.
Imagine if xerox had seen the potential of what they had.
Who ever made that decision should be remembered as one of the biggest missed opportunities.
Thanks for the memories. I worked for sun from 1988 to 1998, but got out before the big fall. I really loved working there at the time.
In the later half of the 2k's, a friend's company would get full rack Sun systems in lots with other equipment they actually wanted. Apparently that was a tactic used by the wholesalers to get rid of junk. Friend's company had no use for them and nobody would buy them for less than "paying them to take it away as scrap". They offered them to me, but I had no use for anything other than the racks and back then I could buy already empty used server racks for around $50+"you come get it", which is a whole lot cheaper than trying to figure out what to do with multiple dozens of very old SPARC systems.
Former Boeing Everett IT Support... We had Sun Microsystems, Silicone Graphics, and IBM Catia. The Triple 7 (777) was designed and built without prototype physical models. It greatly reduced the design cycle and 777 was an instant hit. We also had; Windows, Mac, OS2, and a few others just to keep our techs busy. Close to 50,000 devices on our network in Everett. Engineering, Factory Production, Customer Relations, and dozens of other Departments spread across a very large campus. Our team provided end user support and had over 350 people.
Sun Microsystems lost track of their customers motivations, and that was fatal. In the early 1990's they:
1) failed to recognize that their systems were being used as routers. Sun workstations with 4 port Ethernet cards in every slot. If they HAD realized this they either would have acquired Cisco (both were headquartered less than a mile apart in Menlo Park in 1993) or they would have become Cisco;
2) they could have bought Amdahl's Key Computer Labs, which had a completed mainframe ready for production in 1993). Instead, the same Amdahl team built the Cray mainframe Sun eventually acquired years later;
3) in 1999 most of Sun's sales were due to companies' back up plans to deal with the infamous Microsoft Y2K bug, with Unix OS;
4) Sun tried to undermine Microsoft by offering a fully compatible alternative to Microsoft Office called Star Office. Microsoft retaliated by funding Linux development to undermine Sun's Solaris.
5) Intel had far more processor sales than Sun, and could fund much more rapid semiconductor advances than TI and other SPARC processor foundries could keep up with.
TSMC Arizona plant. The Taiwanese engineer are paid less than their white counter part and needed to work overnight shift while the white engineers don't.
This reminds me of 1863 to 1869, roughly 15,000 Chinese workers helped build the transcontinental railroad. They were paid less than White American workers.
The Asian engineers are looked upon as less worthy than their white so called colleagues.
History is repeating itself.
Racism is alive and well in USA.
StarOffice was not originally Sun product, Sun bought it and made it into Openoffice. And that is what I use right now as we speak, when I need Microsoft Office type docs, under Libreoffice name (executable is still called soffice)
While I was at Sun Quentin one day Scott M. was talking to a tour of students who were visiting. One student asked him if he could change one thing, what would it be. Scott was (for a change) honest and admitted that he screwed up by not seeing the market for routers. But then Sun had lost interest in The Network Is the Computer (remember Network the dog?) some years before that.
Dude...... I grew up in Redmond in the '90s and Early '00s, stepped DEEP in Microsoft and Sub-California tech culture.
This video from 10:45 to 13:15 explained SO MANY questions that have been festering in my mind since I was a little tot, seeing all the big-name computer brands rapidly come-and-go in school, home, work, and stores.
Fascinating.... man, Sun screwed up. A tale as old a time.
You probably mean “steeped”. No idea what “Sub-California” is; Baja doesn’t have much tech.
At Rice University in the late 80's they were setting up Sun workstations everywhere. Big beautiful monitors and an OS that was pretty friendly. Over time some Apple machines moved in next to them, and these were junky by comparison, with very picky and nuanced methods for doing anything, it seemed to me. The Suns were really luxury items in that day.
Worked with one of the suits in this saga (and my wife was for a time at Motorola). This is the best concise account I have come across! Nicely done! 😎✌🏼
Working at Sun I got to meet Jonathan Schwartz and Scott McNeely. I got to meet the founders of Google. I got to work with some really smart people. In building 18 we had a computer closet that was filled with historical equipment. We had original Altos computers from the HP PARC facility. It was so much fun working there.
Excellent video covering all the major points accurately. I worked at Sun 2000-2011 and even in the declining years it was a great place to work. They treated employees well and and had great openness internally. A follow-up video could be a list of all the tech we now use and take for granted that started at Sun but was far too ahead of its time to be commercially viable when released.
I touched a Sun workstation once.
It was a diskless workstation that booted across the network from another computer.
I had never seen or even heard of that way of booting before, though subsequently I've spent a lot of time setting up and using diskless workstations on various platforms.
After that first encounter, I never even saw another Sun product again.
I started computing with 8 bit systems that ran proprietary OSes. The company I worked for, DHL, even made their own hardware, OS and application software for 8 bit and 16 bit hardware in the early 1980s.
Then, they moved to Xenix on the Intel 80286, and NCR Unix on the NCR series of servers based on the Motorola 68020 and 68030 processors. Later they used HP-UX on the HP PA-RISC systems. Finally, when the Intel 80386 appeared, they used SCO UNIX as well.
None of these systems were used with graphical workstations, just dumb Wyse50 terminals.
So, I learned a LOT of command line computing.
Later, I was exposed to Novell Netware, Windows 3.0/3.1/3.11/95, OS/2 2.1, Warp and eventually Linux.
I've only recently started using FreeBSD on my home server.
Sun certainly was a big player and very instrumental in getting the basic backbone of the internet started, but Linux, in particular Red Hat Linux completely wiped them out with their dirt cheap alternative.
I also agree that IBM announcing that they had invested $1 billion in Linux, was a serious affirmation of how good Linux was. That sent shockwaves that affected both Sun and Microsoft.
The availability of open source solutions has completely changed the landscape of computing forever.
When Richard Stallman lauched his Free Software movement, I'll bet he never thought it would end up being used for surveillance capitalism.
I worked with Solaris for many years, then it was gone. Thanks for filling in the gaps in my knowledge of this company.
Just a comment on the Cray acquisition. Sub did not acquire Cray. Cray was acquired by silicon graphics (a competitor). Cray had a division that developed sparc based systems and this division was sold to Sun by silicon graphics.
I remember inn 1991 when Eurotherm at their Reston campus got some Sun Workstations for software development. At that time we had Macs and PCs and the PCs were really only used in the lab and the Macs for documenting. This made the Macs the main information originators. The Macs were all networked and could share digital information straight forwardly. They also ran time sharing applications at the same time with real WYWYG What you see is what you get. The PCs were not the useful the Macs were and the Sun Workstation had a lot of characteristics with the Macs. They were not WYSWYG, but they have highly readable bitmapped fonts. The SPARC station was fast! Sunview Sun's window system had tool talk that was suppose allowed various parts of the windowing gadgets to talk to each other. I could cut and paste huge text logs using the windowing cut and paste at a snap while on a Mac the system might not respond for tens of seconds.
Sun Workstation then went away from Sunview windowing that had been a system that was more of a advancement over the Macs with bends towards engineering and scientists to the the common desktop environment used my HP and IBM more a kin to MS Windows window system that made the Sun and other work stations seem more like a Windows PC than an engineering Macintosh. I think that was the death of the Sun Microsystems workstation. We could still use an advanced form of that Sunview windowing system today in 2022 for engineers and scientists running CAD and other scientific packages.
I do remember the Daisy System workstation. It was used for electronic circuit design in 1992. We had one and I was made system administrator for the electrical engineering department which made me the one to give it attention. It had a screen that was so dim you could hardly see it and very slow to draw. I was told that was due to the high resolution screen it was using for its day. It was extremely slow and the only way to move digital information was on 5" floppy disk. If was the critical production flow of circuit design from EEs to the drafting department. We were having to schedule engineers to use the Daisy for schematic capture use. This was a bottleneck and inconvenient.
I remember talking to software support for the Daisy and people at that company they were saying they thought they are going to make it and there was talk of Daisy abandoning its proprietary workstation and porting their software to a Sun Workstation. That was exciting, but Eurotherm was paying a lot for licensing the Daisy per year and only using it for schematic capture. It might have been $30,000 a year. By the time I was involved with the Daisy the discussion had changed keeping it to having me consult the various engineering teams for a replacement for schematic capture. The Daisy ran SPICE a mathematical common source library of mathematical models for electronic devices.
Although SPICE was developed in the open Daisy's libraries were proprietary. Our engineering department didn't use the simulation part of the Daisy which may have been its most expensive part. The proprietary nature of a lot of the Daisy made it less appealing to Eurotheum's company uses both in Reston and in the UK. Eurotherm had huge layoff in fall 1992 which included me. Its very hard to retire old CAD systems like the Daisy and without an administrator to transition data, older designers may have never really left that old Daisy workstation. The company adopted a PC based schematic capture system which wasn't the Mac based version I was hoping for and a great deal of people preferred. I was told later the engineering team that was pushing for the PC based schematic capture system was really doing it with the hopes of getting another PC. We did everything on Macs, but they wanted another PC and decided to vote for the PC schematic capture system to get it. A couple weeks after that team got the new PC for schematic capture half of them got laid off.
Worked there a couple of times, and of the sector / type of work it was, it was certainly a unique experience, and a lot of that was embodied in the espirit d’corps, for want of a better term; not fanatical, not masochistic, but a lot of qualities one doesn’t expect or find in the field, humility, modesty, diligence, application, focus, dedication, and actually, out of all the places I have worked, one of the very few where I can honestly say I witnessed people contributing their intellect and articulating their view within the team regardless of rank, one of the few places that actually practiced the flat hierarchy, all are equal principle, which actually, many firms claim to operate according to too. It’s a very rare thing. I will say that was my experience from the people and teams I worked with, so I can’t claim this is as empirical globally within the company. I will say, a lot of people benefited significantly right across the shop floor, from how well they did, and the share options people were able to invest in.
I remember buying a copy of Solaris when they made a x86 version for around $100 or so, thinking I would be so cool to learn it. (please let me know if my memory is right!) Also recall buying a workstation with one of those optical mouse at one of those firesales. Not sure where it is now, but yeah, good memories. You can follow up with "Plan 9" OS and how that evolved to Google Cloud and what we have now days ... and where we might be going in the future.
Yes they brought out Solaris for x86 but it was rather crippled in that it lacked the catalogue of application software that the SPARC compiled version had and the performance delta was too much in favour of SPARC at the time.
@@BadatTanking the only thing it really lacked was Netscape Navigator. This was corrected later with Solaris 10. I still run Solaris 10 on i86pc in my private datacenter, and it is absolutely fabulous, I've weathered all the shit from GNU/Linux and FreeBSD thanks to sticking with Solaris. The software isn't a problem either, as I've built and packaged a vast library of the same software which runs on GNU/Linux. I'll be going to Tribblix on SPARC and SmartOS on intel soon, and with SmartOS, 15,000+ software packages will be available, in addition to having all the cool Solaris features like FMA, zones and ZFS, all running on intel hardware.
Solaris 7 for intel cost $50 USD.
I used to work on all the home and office computers for Kleiner Perkins Dornan, Mayfield Fund, and Silverlake Partners. They were out on Sandhill Road in Menlo Park. I got to meet all the senior partners it was fun.
Thanks for the history lesson. My professional career spanned early SunOS thru Solaris servers. I moved to Linux support when linux finally became, what I called, Ready for Prime Time. I lost touch with Sun, coincidentally just as Sun was sold and I never looked back. Your review sounds spot-on. Sun Microsystems taught me a lot, and it was a good ride while it lasted. I built many good data centers based on Sun and I feel lucky to have been in the right place at the right time.
special thanks for a great 2022 season of Asianometry. Your page is one of only two we support via Patreon and your content is always topical, interesting, and educational. merry christmas!!
had worked for Sun support service before the bubble pops till 2003 ...
One of the key things was high prices for cheap pc parts ...
Had to replace lot of 9gb IDE drives in Ultra 5 and 10 ...
Server side there were also issues with things such as the FC-AL centerplane , DC-DC converters
Also at the software level we had to deal with 64 bits but had to run commercial software in 32.
What was great is the ability to boot either from virtually any scsi media and the network
was it from pre OBP systems to the OBP 4...
My 2 cts
In the late 80's I worked at a division of a large electronics company in New York. The IT people and engineers at this electronics company were all degreed graduates from top universities.
My previous computer experience was limited to PC's running DOS and Windows. I had used "inexpensive" CAD software for PCB design.
At this big electronics I started doing PCB layout on Apollo, HP and Sun workstations using Mentor Graphics CAD software. I asked a few of the "smart" guys why these workstations were so much "better" than PC's with Intel CPU's. I got some odd answers. One engineer with a PHD in computer science pointed to the word "workstation" on his Sun computer and something like "It says "workstation," that's why it's better."
A few of the IT experts said something like "PC's will NEVER be ready for primetime, Unix workstations will always be better."
They said these sort of things all through the 90's.
By the early 2000's many of the CAD software vendors had started porting their expensive tools to computers with Intel X86 CPU's running Windows NT/XP.
Well, I guess they were right all along. Nobody remembers Windows or Intel based computers anymore.
Or do they? ;-)
You helped me fill in an interesting blank and make the connection! I had been selling an SD-WAN product called Talari and then Oracle bought them. Oracle walked away from support of the new hardware Talari had introduced just before the purchase. Trying to force existing users to a new hardware platform. It was likely to move SUN boxes! Makes sense now!
An interesting story. In many ways reminiscent of DataPoint corporation. They had the first working network, distributed data processing, enterprise level laser printers, and many other technical innovations.Unfortunately much of the senior management kept reading their own press releases and really had their heads up their collective asses. In the late 80s they refused to consider "entrepreneurship" and instead paid talented technical people generous severance packages to start up their own companies. The USAF has a term for DPTs senior managements thought processes: up and locked.
The narrative on Cray and Sun misses the point that starting in the mid 1980's Sun's 50K work stations had taken the wind out of Cary's multi- million dollar super computers' sales (pun intended). Few engineers wanted to stand inline to share a Cray when their own personal Sun work station would give them the same answer... albeit a little slower. Departments that had supported Cray usage quickly sprinkling Suns around like hanging lampshades and a financially important component of Cary's sales disappeared literally overnight.
I went through the entire SUN life cycle from start to finish and it was a hell of a good ride for many years.
What people miss about the RISC revolution is that in the 1980s with Intel's 8086 and similar products, the increasingly complex CPUs of the day were using a technique called "microcoding", or a lower level instruction set inside the CPU to run instruction decoding, etc. It was assumed that the technique, inherited from mini and mainframe computers, would be the standard going forward, since companies like intel were increasing the number of instructions at a clip. RISC introduced the idea that if the instruction set were simplified, CPU designers could return to pure hardware designs, no microcode, and use that to retire most or all instructions in a single clock cycle. In short, what happened is the titanic turned on a dime: Intel dropped microcode like a hot rock and created pure hardware CPUs to show that any problem could be solved by throwing enough engineers at it. They did it by translating the CISC x86 instructions to an internal RISC form and deeply parallelizing the instruction flow, the so called "superscalar" revolution. In so doing they gave x86 new life for decades.
I worked for SUN for a short time in the CPU division when they were going all in on multicore. The company was already almost in freefall. The Sparc design was flawed and the designers knew it. CEO Johnathan faced questioning at company meetings when he showed charts with Java "sales" presented as if it were a profit center (instead of given away for free). I came back to SUN again on contract after the Oracle merger. They had the same offices and the little Java mascots on their desks. It was probably telling that after my manager invited me to apply for a permanent position, I couldn't get it though their online hiring system, which was incredibly buggy, and then they went into a hiring freeze so it was irrelevant.
I should also mention that not all companies did chip design in that era with SUN workstations. At Zilog we used racks full of MicroVaxes and Xwindow graphics terminals. I still have fond memories of laying out CPUs and chainsmoking in the late 1980s until midnight.
The chainsmoking in the late 1980s until midnight is very familiar to me too.
RISC vs CISC is about the bandwidth between the CPU and the memory. CISC (Originating with the DEC PDP11) was about fetching one instruction for as much data as you could. RISC Originating with the DEC PDP8) is about fetching smaller, but more of them, into a simpler, cheaper engine which can go faster, and could be developed at the pace of Moore's law. At first (in the days of magnetic core memory) CISC pulled ahead for obvious reasons. Later, semiconductor memory made caches possible - you could fetch a bunch of instructions form core memory into semi-conductor cache - possibly 10 times faster - better performance - if you paid for it. Motorola (CISC) could not reinvent their processors at the speed of RISC.
Intel got ahead by superior advertising, not superior product. Purchasing in large organisations is not done by engineers and programmers. My grandmother was famous for saying "the adverts speak very highly of it" meaning that no one would speak highly of it who was not paid to do so. No one who had a choice ever switched from Unix to Windows.
The loss of Sun is a tragedy we haven't fully recovered from yet. The retreat from the workstation market left them without an on-ramp for SPARC/Solaris and confined it to niches and legacy system, most of which now run on x86 processors running Linux.
Precisely. It was giving up the workstation/desktop market that was really the beginning of the end for Sun. It just took a long time for Sun to bleed out.
As a Linux user for years, this is an interesting video about the history, it also explains what I saw during my childhood as I was using MS-DOS then Windows 3.1 then 95 to 7 before switching to Linux. It is interesting to see how scalable the CPUs are nowadays, nowadays we have more cores but only 2 socketed CPU per mainboard. That being said, I think you would refer to Threadrippers/Xeons as High end workstations than Mac Studio ^^"
Sun was a juggernaut and I was a proud evangelist in the burgeoning Ecommerce space in the mid-to-late 1990s.
Where I felt they missed the boat was they failed to defend their workstation space and I emphatically felt they had an opportunity to acquire Apple to address that in the late 90s to early 2000s. Instead, they entrenched and pushed OpenOffice and the market wasn't buying. Cheap intel hardware and open source linux eventually killed their server space and Microsoft & Apple cleaned up the workstation space.
The final insult to injury was Apple adopting a Unix standard without Sun and pivoting to consumer electronics to outlast them.
I feel the video missed a key point in "the network is the machine". Sun was absolutely revolutionizing in this mantra and had they not lost sight of that, they might well have survived.
I programmed and supported Unix/Linux systems professionally from 1980 until 2021...Of the dozens of Unix/Linux hardware/software varieties that I worked with, I liked Sun the best (extremely stable)...I like to say that "Solaris put my daughters through college" and in my personal Linux-based home lab, I use ZFS to manage my storage...The largest server type I ever configured myself, was a set of Sun E6900s (perhaps $800K in ~2005 dollars , 1,200 lbs per server).
Those E6900s could likely be replaced by a single socket 96-core EPYC processor these days.
OMG this video brought back so much memory and showing how much the IT industry has evolved
That whole "Unix era" was a cautionary tale in tech arrogance. Both the Unixae and Apple were too haughty to come down to the office secretary's level, while Microsoft went after the business world without hesitation, offering a full spectrum of software apps and OS for the business world. MS was for so long just pathetic quality-wise, but good enough to limp along. Literally, the Unix snobbery and Apple elitism made MS's poor quality all there was left. IBM tried with their desk OS, but too late and not hard enough. Linux has never taken off on the desktop because MS simply had too great a lead -- and eventually got Windows fairly respectable, reliable.
Very interesting take on how Microsoft captured the desktop with inferior product.
@@awuma Sometimes you don't have to be the best, just the most consistent. For all intents and purposes, MS is a stable, reliable product now, but not back in the 1990s. MS did strictly "show" bets back then, i.e., third-place was good enough. Meanwhile, the Unixae and Apple were chasing this or that idealistic or purist soap bubbles and totally forgot or mehhed the business world.
Back when Windows NT first came out, the US Government had a requirement that all its divisions buy “open systems”. This specifically included POSIX compatibility. So Windows NT initially shipped with a POSIX “personality layer”, just to satisfy the purchasing requirements. In fact, this was never really used in production, since all the customers were quite happy to run software specifically developed for NT’s own proprietary APIs.
The internal modularity in Windows that allowed the creation of these alternative personality layers has long since bit-rotted away. Which is why their more recent attempt to emulate Linux APIs in WSL 1.0 didn’t work very well.
This is more informative than the other videos covering Sun's history, great job.
15:40 the chart labels the year 2001 twice. It should have 2003 after 2002, but it goes back to 2001 when showing the -3,429 (M) net income loss. The chart is also labeled 1999 to 2001 on the top, but also shows 2002 and what I assume should be 2003. Otherwise great video!
Was the Stanford label a joke?
I was working at the Southern Alberta Institute of Technology in the late 1990s and we had put in a request for a lab full Sun Servers for a IT infrastructure course. Windows 98 was out and Windows NT just appeared. When we had our meeting with management, they had just come out of a presentation with Microsoft on Windows NT. None of the managers were IT, they were educators making decisions on which operating system we should use. They liked Windows NT because it looked like Windows 98 - they didn't have a clue about Sun. There was extreme pressure to go with NT. All of us in the IT field were rolling our eyes in disbelieve. Eventually the whole school migrated to Windows NT. We had Novell Netware running our email on one little Pentium 90 server when NT replaced it, it took 4 physical servers to run the email and it was slow.
Well done! Sun was my last Valley job and it was a terrific place to work. Anyone remember OpenStep? My first Valley job was for System Development Corporation. They did the code for the early rocket systems - straight assembly language, then JOVIAL. 50s nerds still working when I hit the scene. Legendary bunch to my mind. Ford Aerospace and IBM Fed Sys Div were all around the "Blue Cube" aka Onizuka AFB, were all the action was. SDC was bought by Burroughs later Unisys. The Blue Cube had computers that came out of submarines and were like 6bit or 10bit - some strange number. Next I worked for Metaware Inc. Frank DeRemer and Tom Pennello made the first 32 bit C compiler to hit the scene. Anyone remember when compilers and tools were actually sold? That was where I made my technical bones. Crazy place to work, so I moved over to Borland Int'l. MS buried Borland by giving away the MS spreadsheet and db, in true monopolistic fashion. Borland was also a crazy place, and going to Sun was sweet relief. Any of these outfits would be cool to hear about. SDC - truly historic. I'd add Xerox Parc to the top of the list - that is where many things began.
I hired on to a research university to run several server rooms in 1999. They had just purchased a bunch of Sun server hardware on the used market, and I had literally never logged into one (Novell, Windows). I soon discovered how good server hardware could actually be. At one point the servers running files, auth, web, and remote access had uptimes of over 4 years. When the Oracle purchase happened, we replaced all the then 12 year old Sun hardware with Redhat on x86 and VMWare, then got rid of VMWare for Redhat's virtualization when they drastically increased licensing costs. We went from 100% Sun to zero percent Sun in the space of a couple months - seems we were not alone.
I was a Unix SA starting in the mid-80s and worked with Sun systems from the late 80s. In my humble opinion there were ruined by ATT forcing them to move to SYS V which was known as Solaris. It was a huge step backwards and many companies avoided it for as long as possible. This opened up the way to competition that didn't have to be there. I jumped on Linux when it came out and have never looked back.
AT&T. What was wrong with Sys V? I liked it.
@@patrikfloding7985 Early versions were very buggy and performed terrible compared to SunOS. I worked supporting a trade floor and the attempts to migrate was unnecessarily painful. It didn't help that they completely changed so much so that SAs were often confused and frustrated.
Of course over time it was improved greatly. But the damage that I saw as a customer, was done.
BSD vs SysV preferences totally depend on your history. Either path makes the other look bad. They both lost out by not merging support for the syntax differences!
@@marshallwilensky7932 One of the things I really liked about Linux or really Gnu is that it supported both variants for common sys admin tools like the ps command.
According to my memory files, Sun workstation cost more than twice what PCs cost in the mid 90s.
Sure, Suns were more powerful... But with networking you could scale those cheap PCs... and regardless, the 'cheap' PCs caught up in power very fast.
Thank-you Intel. Thank you Nvidia.. Than-you AAAAMMMMDDDDD!
I have fond memories of using a Sun workstation for CAD in Engineering Design class back in 1986. We used DOGS (Drawing Office Graphics System) with the optional BOXER solid modelling software. Wish they could make a comeback.
We still run old Sun servers on legacy machines at my job. There are no plans to replace them since they still do the job and we have plenty of spares. I get to use one every once in a while. The desktop (CDE I think) looks like it's from the 80s.
"The network IS the computer" my brain when I see old Sun servers
I always thought that was a silly tag line.
At our university there was a poster saying "Do it in the Sun" (and a big Sun logo )
McNealy was an early advocate of the networked environment; his company's motto was "The Network is the Computer". At times, he has been known to be skeptical of products that do not integrate well with networked environments. One example McNealy has given involved the Apple iPod. As quoted in The Register, McNealy said, "There’s a pendulum thing where stuff is on the client side and then goes back into the network where it belongs. The answering machine put voicemail by the desk, and then it went back into the network. Your iPod is like your home answering machine. I guarantee you it will be hard to sell an iPod five or seven years from now when every cell phone can access your entire music library wherever you are."
@@derbagger22 it is a good slogan
Having grown up in the tech industry working initially for HP, this is a fascinating overview of Sun and the high tech industry! I had the privilege to hear Scott McNeally speak at an industry conference. Interesting enough Scott took a cheap shot at MicroSoft then followed up with a cheap shot at Intel. Sad but I get he felted slighted by the competition. I still have huge respect for what he accomplished in his he tech industry. Scott was ahead of his time.
I remember Scott McNealy, former CEO of Sun Microsystems, being in the cover of the Fortune magazine in 1994. That was the feature about their new programming language, Java ☕.
Now, they are part of Oracle Systems (the same company that attempted to buy TikTok).
I always get the shivers when installing Oracle with Java, telling me „3 bln systems are running Java“...which means 3bln vulnerable computers...😂
Great video. As a former DEC employee it's good to know we were not the only flash in the pan.
such a sad comment in a way... digital a name that should be remembered just want 'poof' ... including all the signs on buildings near my then home.
I wouldn’t call DEC a “flash in the pan.” They had a good, long run in the minicomputer era.
Excellent video as always, and I think the parallel with MS at the end is very appropriate indeed!
“The network is the computer.”
Xerox Parc was making magic in their lab. It's incredible that not a single managerial type identified their work for the gold mine that it was.
Apollo wasn’t shut down by HP, but it’s product lines adopted the IP through out the HP UX workstation class. You could say they lasted longer than SUN in that sense.
HP Apollo Domain was a migration path for organizations that had invested in Apollo boxes, to give them time to get onto the HP platform. I wasn't involved directly where I worked; but from what I recall hearing, converting a lot of the mission critical code base to be POSIX compliant was where a fair amount of effort was placed during the migration.
@@jonp4846 The V-class servers and Superdomes where a direct result of the Apollo acquisition and "evolution" of the acquired IP. Also the Apollo techies I worked with where legends.
@@xmaniac99 My first encounter with Apollos was at college in 1981. Initially it was pretty much just a room with a few terminals and teletypes, the punch card room, IBM mainframes and Pr1me minis. There were 2 Tektronix 4010s and a large pen plotter, but that was it as far as graphics devices go. The, in sophomore year, they put up a prefab building next to the computer center, bought 4 DN300s for the new lab, and off we went.
Nothing but respect for the engineers that came before me and provided amazing machines like these.
Who remembers the massive amount of thefts of sun hardware? It was almost impossible to insure SPARC servers at one time. We had three racks of sun servers and the sun racks had to be secured inside a massive cage, hardened steel, special locks etc etc. Each cage weighed in at just under 1,000KG, had to be assembled onside by a team of four people. Some of the steel panels were about an inch thick. Needless to say the sun racks are long gone but the cages remain in our datacentre as nobody wants to take on the disassembly of them.