Good luck to all those affected
Oracle is laying off staff in its hardware division, The Register has learned. Current and soon-to-be former staffers have whispered online and to El Reg that the database giant has shipped packages containing paperwork ending their employment. The workers learned of this by receiving alerts from FedEx that the parcels, which …
Isn't this a case of chicken and egg?
The salespeople are selling cloud because the customers aren't wanting to make big investments in a niche hardware platform with an uncertain future. Hardware sales across the industry are decreasing and a significant chunk is now done by the larger cloud/SaaS/PaaS providers as traditional customers either migrate or investigate how they can run workloads in that environment.
Should the sales people stick with hardware sales and disappear at the same time or should they try and sell something in the hope of retaining the customers and their jobs?
Solaris/SPARC and Sun were seeing some of it's biggest interest and growth towards acceptance when they did OpenSolaris. When Oracle took over, closed it down, closed down links to traditional high spending markets and then refused to work with external software vendors that produced market leading products that even slightly clashed with what Oracle had in-houose, it was the end.
They dropped ranges for telecommunications systems (DC power) claiming special requirements - whereas the competition had no issues putting DC power supplies into off the shelf systems.
The focus became big iron for Oracle only workloads. If you weren't buying M5/6/7 or Exa, you were a nobody.
Then they dropped the blades.
Why would I listen and care about Oracle Cloud?
When Larry bought Sun he put the hardware business up for sale immediately. He was clearly hoping to split Sun into three pieces, selling two for a net profit, and keeping the third. Larry, being a modest guy, wouldn't have mentioned his brilliance to everyone.
Note: One of those statements might not be true.
Once a buyer couldn't be found (and Oracle tried *hard* to sell it), Larry wisely claimed that keeping the hardware side was the plan all along and that it would be jewel in Oracle's crown. One that merited only the bare minimum investment to milk it until obsolescence.
It is happening all right - the conference rooms in Santa Clara are booked by HR for Friday. And just before a holiday is just when Oracle would do it.
Before I left people developing internal tools for bringup and sustaining - and CAD - were pushed off Solaris and Sparc systems onto x86 and Linux.
Oracle management - "who'd eat this? It's dog food!"
It is indeed sad to see good technologies go down agonizing this way ...
I'm no Solaris guru, having only practiced a bit, but this is a technology I'll miss. Where else do we have ldoms and Niagara CPUs ??? I remember consolidating a dozens of old systems onto a dozen of ldoms, running 4 times faster.
Zones were a vangard techno too, before containers ...
How sad to see all this go ...
One of the reasons, that the hardware didn't fly under the Oracle banner might have been the Oracle brand itself: We used SunFires for Linux workloads quite often, but the moment they became larryated we stopped immediately. Sure enough, for the next NTP on ILO debacle they wanted us to pay for a firmware upgrade. Not even HP dared so.
This post has been deleted by its author
..ask Jonathan Schwartz and Scott McNealy if they still think selling Sun was "the best decision for the company". Thousands of people laid off and demoralized, great technology stripped for parts.
But as long as Scott's kids can play in his private hockey arena there's nothing to see here, right?
Great legacy guys.
Or Jonathan Schwartz and Scott McNealy could have stripped Sun themselves.
I'm not sure Sun would have lasted another 5 years. And I'm not sure Schwartz, McNealy or any other person could have turned Sun into a company that could have survived past the global financial issues when all your products were either free or in decline.
As much as I dislike Oracle and their business practices, they've arguably provided more life to Sun than Sun could have.
True this. Before Larry arrived the potential buyer was IBM, who would have taken the customers and gotten rid of the employees. There were no engineering layoffs at the transition, and we started hiring soon after.
When Larry was in school there was a lot of interest in language and application specific architectures. I think his fantasy was to modify Sparc to run the database real fast. And some of that happened, but not enough.
If Larry fell off his yacht five years ago, the hardware division would have joined him at the bottom of the Bay soon after. I'm no fan of Larry, but I bet he finally gave in to Hurd.
If SPARC hardware wasn't hopelessly overpriced it would sell better. I appreciate the quality of the product but against a backdrop of alternatives that are "good enough" often the only reason for the architecture to persist is legacy apps deemed too difficult to move or specific cases where you want alternative architectures in the mix. For everything else, why wouldn't you dump it onto X86-64, or maybe at a push, ARM?
I still have a tiny handful of apps on UltraSPARC II. After 20 years service it just isn't cost effective to keep it on the there anymore. They have literally just a few weeks left. Even if that means replacing an x86 server every 5 years; which is probably an under-estimate of lifetime.
Hell I still have a socket 604 xeon that works quite happily; albeit it's a bit power hungry compared to a Raspberry Pi for the same kinds of mucking-about workloads that the two see!
On german IT-Portal (https://www.heise.de/forum/iX/News-Kommentare/Solaris-Sparc-Vater-John-Fowler-wirft-das-Handtuch/Quartalszahlen-Jahreszahlen/posting-30818488/show/) you can find the complete summary of the desaster:
fiscal year Q1 Q2 Q3 Q4 total
2011 1800 1100 1000 1200 5100
2012 1000 953 869 977 3799
2013 779 734 671 849 3033
2014 669 714 725 870 2978
2015 578 717 712 818 2825
2016 570 573 604 725 2472
2017 462 497 520 597 2076
comparison (loss in percentage)
comparison time Q1 Q2 Q3 Q4 total
2012 versus 2011 -44,44% -13,36% -13,10% -18,58% -25,51%
2013 versus 2012 -22,10% -22,98% -22,78% -13,10% -20,16%
2014 versus 2013 -14,12% -2,72% 8,05% 2,47% -1,81%
2015 versus 2014 -13,60% 0,42% -1,79% -5,98% -5,14%
2016 versus 2015 -1,38% -20,08% -15,17% -11,37% -12,50%
2017 versus 2016 -18,95% -13,26% -13,91% -17,66% -16,02%
comparision to 2011 (total loss in percent)
comparison time Q1 Q2 Q3 Q4 total
2012 versus 2011 -44,44% -13,36% -13,10% -18,58% -25,51%
2013 versus 2011 -56,72% -33,27% -32,90% -29,25% -40,53%
2014 versus 2011 -62,83% -35,09% -27,50% -27,50% -41,61%
2015 versus 2011 -67,89% -34,82% -28,80% -31,83% -44,61%
2015 versus 2011 -68,33% -47,91% -39,60% -39,58% -51,53%
2017 versus 2011 -74,33% -54,82% -48,00% -50,25% -59,29%
So - currently Oracle has lost nearly 60% of all hardware revenue from 2011 (merger) until today.
this is called "Success story of the decade" - comments ? questions ?
PS: Numbers pulled from official quarterly reports of Oracle - for example like the current one
search for "Hardware Products" and you got the numbers - Oracle tries to "cloak" the stuff since boosting the cloud and mergind numbers of pure hardware and service etc. together.
Comment 1: When a company wants to switch to software they tend to dump anything that's losing money into the hardware segment reporting. See, for example, IBM which famously passed around the IP licensing division and it's $2B+ no-cost income to the fair-haired child of the day.
Comment 2: It's Oracle, famous for abusing its customers. The problem is that hardware is subject to less lock-in than is software and rather than adapt to the hardware market reality, Oracle tried to keep its business-as-usual management style going. We can see how that worked out.
Oracle SPARC revenue calculations is a tough one for the simple fact that SPARC has had significant performance advances in last 6 years and so customers are buying smaller systems than previously which cost much less and why the lower revenue-just look at the unit marketshare figures and you'll see that unit volumes has actually increased in this timeframe-for example on revenue, back in 2011, a SPARC T4-4 with 4ch/32cores sold for around $80K list price while today, 6 years later, a SPARC T7-1 with the same 32 cores that’s 2x faster per core (and a whopping 7-8x faster per Chip!) sells for $40K, half the list price and hence half the revenue assuming customers needs grew by about 2x or so in timeframe (and think of the 2x performance/Database license improvement for Database customers). This had a rippling affect across the product line and why Oracle sells a majority of 1/2/4 socket systems today that easily replace the largest of M-Series from 5-6 years ago costing millions back then.
Intel CPU's have gotten faster over the period too, but they're still selling bazillions of them.
The problem is Oracle, a company I can never understand anyone getting into bed with, and fewer will going forward, oh sure they'll milk their existing hostages for a while yet but most are looking for a way out and many will find one.
Every former Sun customer I talk to is trying to get away from Solaris/SPARC as fast as they can. In large part this is because of the way Oracle treats them and not necessarily because of the product. And all the Exagarbage, Exajunk, Exaexpensive boxes they have had shoved down their throats were the icing on the cake.
And just FedEx-ing layoff packages to employees is a class act that only Larry could conceive. I am just surprised he wasted money on FedEx when USPS would have been cheaper.
I set up an account on each fedex, dhl, and ups to see if I was getting notice. I even though Larry would use that ontrac amazon service to cut a few pennies. This is just so sad to see. Sparc is gone. The world is now left with intel/amd and arm. I see a monopoly brewing and box prices increasing.
This won't affect server prices because:
1) Sparc was basically irrelevant now, probably less than 1% of the CPU's sold
2) OEM hardware of any kind is set for rapid decline, most companies will simply go Cloud Iaas then PaaS with SaaS whenever that fits (why would you build an in-house CRM for example?). The cloud companies make their own server so Intel will be fine, Dell, CISCO & HP less so.
To be blunt, this layoff is only a logical follow-up to the layoff of the networking division in January which triggered a WARN notice in Santa Clara and made the San Jose Mercury news. The "pivot" to cloud is trying to catch up to the competition as fast as possible after the failed prediction 5 yrs ago by C-level bosses that Ethernet-based cloud would never take off and that they should stick with their traditional enterprise model of Exa- racks and InfiniBand networking. But, as usual, even if the C-level boss gets the axe too, he'll get a golden parachute, while the rest of us get/got peanuts and a "good luck".
I'm sorry for all those affected.
Why would anyone use Oracle Cloud anyway, AWS, MS and Google are set to be the only winners in this game, going cloud is the perfect time to ditch Oracle and their awful customer service, as a company (no the good employees who work there) their future looks bleak and karma is a bitch.
This post has been deleted by its author
Every place I've personally seen or heard evidence of that wasn't absolutely married to Solaris and SPARC has been trying to get away from it forever. I can't imagine it's easy for Oracle to get new customers for anything other than their database applications, given their reputation.
You did see a lot of Sun hardware in the academic space back in the day, because their model was similar to Apple's -- give it away to students so that those students will buy it at a much higher margin when they're executives. But, the only places I've seen Solaris lately are those with 30 years of code that's totally married to Solaris, or workloads that have characteristics making SPARC the best or only choice.
Hopefully Larry isn't being too stingy with the severance packages. I feel bad for anyone who's still around keeping SPARC alive.
"I can't imagine it's easy for Oracle to get new customers for anything other than their database applications, given their reputation."
Surely you mean especially for their database applications?! MS SQL Server is usually a no brainer alternative at a far lower cost and with far fewer security vulnerabilities. (Unless you need RAC which until certain patents expire there is no alternative for...)
This is old news, people! Many hundreds have been laid off in hardware-related jobs for well over a year now. See thelayoff.com/Oracle for the many threads detailing the various stealth massacres. I guess they fooled most of the public, including you!
Oracle never had strategies to sell hardware - like some leaders we know in national politics, somehow they thought just saying it would happen would make it happen. But of course that is difficult to do when you screw over salespeople and customers alike. So those of us who worked on these products, some of them at the top of the technology, suffered along with the slow death.
Oracle treats employees like pure commodities - if they can buy them cheaper elsewhere, they will. But in the case of hardware, they were just total failures. They could not adapt to that side of the business, away from their pure software roots, where you don't have to deal with the complexities of material costs, development cycles, manufacturing gyrations, and more.
Recent SPARC hardware is really good, honest! New servers like the SPARC S7-2 are also comparatively priced with their enterprise x86 competition. Linux is still catching up with a number of unique functions of Solaris and 11 now has a really decent packaging system and boot environment management. If you still run the Oracle DB the the SPARC/Solaris platform runs it faster, more stable and cheaper to license!
So why doesn't it sell? IMHO it is marketing, the core Oracle sales guys never push SPARC or Solaris, just cloud or Engineered Systems.
...and, no I don't and never have worked for Oracle or Sun!
Biting the hand that feeds IT © 1998–2022