Now SPARC and Solaris are really dead

Hello, ex Sun Microsystems employee here, working in Fujitsu engineering since 2010. Posting here since I think this is more relevant to Oracle given we sell SPARC basically only in Japan.

During the last year Oracle and Fujitsu worked for an highly confidential and secret agreement to have Oracle develop Solaris and Fujitsu buy the Systems LOB and develop and sell SPARC.

In the past months LE and MH killed that agreement. There have been some attempts to modify someway the agreement and go on, but unfortunately a final decision has been taken.

LE and MH refused to continue to invest in Solaris, and asked Fujitsu to buy the entire SPARC/Solaris LOB, which Fujitsu correctly refused to do since was not the original plan agreed and out of our core business (which is mainly HW, not SW).

Oracle will simply let SPARC and Solaris slowly die, i.e. Oracle will continue to sell SPARC systems until there is reasonable demand, and then will simply close the LOB and layoff all people. Estimated deadline for the closure is 2020. In the meanwhile Oracle will develop a Linux for eMAG cpus to be used in their own cloud (i.e. Oracle will also get out of the x86 business).

Fujitsu still have in the public roadmap the new SPARC64 cpu in 2020, but the plan is actually unconfirmed and given the situation there are many discussions to just freeze the SPARC64 development right now and invest the budget in a general purpose porting of Post-K. Knowing my own company, I hink this will be decided in the next weeks. My suggestion is: don't expect anything beyond S64 XII.

Unfrotunately this means the death of SPARC and of Solaris, even if the support for the existing systems is granted for many years. What a pity for such a powerful and elegant architecture.

Comment! It's anonymous!

29 replies (most recent on top)

Oracle bought Sun because of their patent portfolio. Oracle never did care about SPARCs or Solaris.

"If flies likes sh--, does that mean it's good?"

Solaris have very good file system named "ZFS" and "zones" light virtualizaction, I only mention this because I use these technologies.

Linux there is no have a good filesystem, LXC is a beginning not-refined technology.

SPARC multithreading is good for microservices.

The system is adapted to the hardware, like apple, and everything is consistent, there are not as many problems with drivers as on linux.

Just love these punks who think that Amazon and YouTube emerged from Bezo’s a--hole fully formed. Keep living in the eternal now, it will bite you in the end. Pun intended!

@U5uIwh1-1Bakt:

Spoiler: that tech built out the modern Internet, and paved the way for Amazon and a shedload of others.

Quit waxing nostalgic about how great Sun was and how Sun invented everything. Nobody cares.

If Sun was really that great and smart - smarter than everyone else on the planet - how come it crashed and burned?

Get over your Sun nostalgia and get a life.

Anyone who think Sun was “useless” doesn’t know their computing history, especially the last 20 - 30 years. Spoiler: that tech built out the modern Internet, and paved the way for Amazon and a shedload of others. The decision to let Sun “rot” was Oracle’s alone. Sun had much better customer and employee relations by any measure. So eat it, punk!

What credibility?

They're not getting any new customers, but plenty are leaving!

How could Oracle possibly parter with a top cloud provider after all the boasting MH and LE have done about the wonders of the Oracle cloud relative these competitors? Oracle would loose any credibility that they have left.

The OP's message certainly sounds realistic to me. The drastic guillotine for Solaris, as some sound like they are welcoming here, won't happen because there are many US $billions of fielded SPARC and Solaris gear out there, a lot of which need a path to the cloud and support for a few more decades. I guess they could sell-of the Sun support org to a service bureau company if they truly want the LOB gone. ARM on the server as been tried a few times, it has not lived up to the marketing hype either, you can ask Nvidia about that. Why is Wim posting Raspberry Pi info about the Oracle ARM effort? Seems silly to mention rpi if they are even halfway serious about getting Oracle RDBMS to run decently on 64-bit ARM. If you were an investor really looking for Oracle to succeed in the cloud, you would have to ask yourself why they are talking Iaas (and ARM) in the first place. Their strength is in software, they should partner with one of the top-3 cloud providers and stop pretending that they can compete at cloud-scale infrastructure with them, especially since they were a few years late to that party. Hopping to your trendy CPU and having a VP blog about a ARM Linux distro is just a very small piece of that. Good luck if you own ORCL.

very true. it seems those investing in ORCL are no more so interested in the cloud hype, and do not consider the DtE explosion in the last year a valuable indicator, but are simply looking at the fundamentals of a short term investement.

i.e. how much dividend, and the total aggregated growth of the revenues, both of which are substantiallly stable and good. so basically only short-term investors are buying ORCL, looking for short-term low profit, and do not care about mid to long-term scenarios. this means that ORCL will still be a rollercoaster in the future and will not have a stable growth.

possible and easy to make some money with it, but definitely dangerous if you want to invest for a long-term investement, possibily with double digit YoY increment.

"If Oracle was priced as a value company it would be trading closer to $35."

The entire stock market is overvalued. Even moribund firms like HPE have P/E ratios comparable to ORCL. Meanwhile, ORCL has recovered everything it lost after the last quarterly report and appears to be headed back above $50.

It's pretty sad if employees who may own company stock or options don't know if the company is a growth company or value company. I'm fairly certain the senior folks who receive a good percentage of their compensation from stock don't view Oracle as value company.

The unfortunate reality however is that Oracle may be dangerously close to transitioning from growth company to value company if earning growth doesn't accelerate. That transition is almost always a dangerous time to be owning a stock.

A: since oracle big mouth honcho Larry E said oracle was a contender for cloud solutions provider & mini-me side kick Hurd professed oracle is going to outpace amazon bc their lead won’t last.

Meanwhile Catz has muddied up the financial performance reports like river after a storm. No way to track any kind of cloud sales from legacy licensing and support ... it’s miraculously ALL CLOUD now — licensing and support “as cloud @ customer” 🤞😉😉🤞 and “public cloud” solutions all mushed into one reported sales figure . . Good luck ever getting to the truth in that mess.

Growth companies don't trade at a P/E of 19! The market is pricing Oracle as company that should be delivering earning growth of at least 10% or more. If Oracle was priced as a value company it would be trading closer to $35.

"Since when are we supposed to be a growth company?"

True. Sales have been steady/stagnant for several years, and the company pays a dividend. That is known as a "value" company, not a "growth" company,

It makes no sense for a company that is suppose to be a growth company

Since when are we supposed to be a growth company? You don't get substantial growth with the way we do things, period. It would take all new management and a total culture change to become a growth company. The growth expectation is just a line of BS we feed investors to try and keep the stock price from plummeting. Our executive leadership team just doesn't have their hearts in it to take on Amazon anymore. They are out of juice and looking to retire.

Hardware sales just being steady is not good. In fact hardware sales have been steadily declining. Neither SPARC or x86 sales are growing. It makes no sense for a company that is suppose to be a growth company to stay in this business. It's not a question of comparing SPARC/Solaris to IBM/AIX. Both are a losing proposition. The only possible justification for staying in the hardware business is case for building custom systems to deploy in the cloud. Even in that case most of that work could be outsourced.

"I agree that SPARC and Solaris are dying, only old legacy customers not able to switch to cloud or x86 are still buying this technology, and even in this very specific situation Sparc is also loosing more and more marketshare to IBM"

So you're saying that people are switching from Solaris to AIX? Given that HW sales were fairly steady last quarter, that doesn't seem likely. Plus AIX has been on life support for a long time.

Oracle is a software company. Mistake was buying Sun and loosing KB and bringing in MH.

There were rumors on another thread that Oracle we eliminate x86 development in 2020 and outsource x86 through an agreement with Dell.

AWS has been offering ARM services for several years. As with everything else in cloud, Oracle has a lot of catching up to do here.

As with everything else Oracle does in the Linux area, I suspect there very little value Oracle adds in the area of ARM support in general.

if I'm reading it right, the OP wrote that we will develop a Linux for EMAG, not the cpu itself. and again if I'm reading right, the 2020 deadline is for sparc not x86.

could be that we will develop an Arm strategy, but I don't think we will never completely abandon the x86 business, EXADATA is one of the best selling HW product we have.

I agree that SPARC and Solaris are dying, only old legacy customers not able to switch to cloud or x86 are still buying this technology, and even in this very specific situation Sparc is also loosing more and more marketshare to IBM. but I honestly don't think the death will be so close, the IB is still quite large, Systems team could go on for years just refreshing and consolidating.

So SPARC is dead and the rumors are that x86 development will be eliminated and outsourced in 2020. Looks like the only hardware development will be ARM eMAG? I would imagine that could easily be outsourced too.

Hmmm ... so a Fujitsu employee has insider info on Oracle that Oracle employees don't have.

Sounds legit to me /sarc

Losers can and should self identify to self terminate now from Oracle so you can move on and destroy the next company with your loser attitude. Later, don't show us your azz on the way out.

It rings true. Oracle shat on Solaris from the time they obtained it. Who cares how many point releases of 11 are released if there is never a version 12?

And you have the contracts of this "secret agreement" you can post to us to prove it really exists? Do you hear rumors in the halls and think they are all real? And all is dead.. so dead that Solaris 11.4 will be released very soon. We are all still playing with the beta with new developments all the time. Of course that is just for marketing as well you will say...on and on

Reading this, I'm glad I left voluntarily after the big Sep RIF.

M8+??? you joking, right? the so called M8+ is just an overclocked M8, nothng new, no improvement at all, and only going to be used in the largest M8-8 server because of the outrageous power consumption and heat generation.

do you think 10% more freq, consuming tons of kW and only in the most expensive server, is going to revitalize the demand?

moreover, you know of course that as of today nor LE nor MH nor SC have definitively approved the M8+ project, this is just a possibility that the SPARC team has arranged for almost 0$ investment, and that ES has presented with some shame to the board.

nope, the M8+ has been included in the roadmap only for marketing reasons, to be able to tell customers that "yes, we have a roadmap" and still sell what we have left in the stock, and most probably will never see the light.

PS I'm not the OP, but from everything I see here in O I think he/she is damned spot on.

I think your article is based on some outdated info. Please check the latest roadmap for Sparc M8+ chipset. https://www.oracle.com/assets/sparc-roadmap-slide-2076743.pdf

Oracle Linux for ARM is not a secret at all....

https://blogs.oracle.com/linux/announcing-the-general-availability-of-oracle-linux-7-for-arm

as it is not a secret at all that Oracle invested heavily in Ampere Computing, the developer of the ARM eMAG platform

https://www.youtube.com/watch?v=-FyqYp-TzKg

Why Oracle should continue with a rotted old technology like Sparc or Solaris? it is not accepted by the market, there are no real compelling advantages in using it, everything around Sparc and Solaris is just old-fashined and useless marketing hype just like the so called software in silicon that noone is using.

gimme a break, Oracle should have wiped out this uselss technology long ago.

better said, Oracle should have never bought that useless company called Sun Microsystems.

Thread regarding Oracle Corp. layoffs

Now SPARC and Solaris are really dead

29 replies (most recent on top)

Post a reply