A week ago, we reported that IBM was in acquisition talks with Sun. Sun has been in trouble for a while now, and has been shopping around the Valley for a potential buyers for the company. This report came from “people familiar with the matter”, but it seems that we now have a confirmation from none other than Intel’s CEO, Paul Otellini.
Otellini made his remarks about the IBM-Sun talks during an employee webcast earlier this week. “I can tell you that Sun was shopped around the valley and around the world in the last few months,” Otellini said, “A lot of companies got calls or visits on buying some or all the assets of the company. It looks like IBM is in the hunt now. And at a hundred and some odd percent premium, I suspect they’ll get it.”
Intel’s CEO also made a doom and gloom prediction about the future of Solaris and SPARC. “I think the stuff on Solaris and SPARC is likely to see EOLs over time through the IBM acquisition. But no strategic reason for IBM to maintain that except to attempt to convert the very large Sun SPARC Solaris base to power. I think that would be their most likely strategy as part of this.”
Of course, Otellini had a few words to say on whether Intel actually really cares about this. “Is it good or bad for us? I don’t know. I’d rather have Sun be independent, I guess.” In other words, no, Intel couldn’t care less.
Hmmmm. It might be just me but I think we need a link to what Otellini said rather than just a link back to the original story, which wasn’t really news anyway.
The brown stuff has hit the fan in the last few months with Sun. The revenue has been heading south for a number of years, costs are still far too high, they can’t lay off any more people and over a reasonable amount of time the share price doesn’t lie. Once multi billion dollar companies can’t sustain themselves any longer they tend to go quickly.
It’s a tragedy really because it just shouldn’t have happened. As happens with these things you need look no further than the top with weak leadership who didn’t understand their technology, the way things were going or the opportunites that were there. They let the lunatics run the asylum, allowed their acquisitions to be squandered and politically bullied and they tried to convince themselves that everything would return to ‘normal’ if they could just come up with one more innovation or product that would bring people to their senses.
Fixed, thanks.
Huh… revenue has been up, it was just midly flat the past year. Which given the current economy was to be expected.
Deliver us from armchair CEOs dear god!
Revenue has been hovering for a number of quarters, but the trend over the past few years has been down and it has nothing to do with the state of the current economy. That’s just pushed things over the edge.
The only way Sun have been able to break even is by having several rounds of layoffs and cheesing off their workforce even more. Silly me. Things are obviously fine.
Really?
http://www.google.com/finance?fstype=ii&q=NASDAQ:JAVA
Some people should stop assuming their “perception” is as good as actual citations of references.
SUN’s revenue has flattened out, but it had not diminished as you claimed.
I think it goes beyond just mere cost cutting; they made a string of acquisitions which were great on paper but then ran into the ground by trying to impose the bureaucratic Sun sales model instead of realising that their bureaucratic model was cutting off their ability to get new customers; try purchasing Sun hardware and you’ll see what I mean. IBM is easier to navigate when trying to purchase hardware than Sun. I want to ring up, order it and thats it. If you want to shunt me from reseller to reseller to reseller – tough, you’ve lost a customer.
They then had no direction; first it was Solaris, then it was just Solaris SPARC, then Linux but Solaris SPARC, then Solaris x86 came back but Linux kept hanging around giving the impression that Sun wasn’t confident with their own line up and direction. Then there is the middleware – where does that fit in? Then OpenOffice.org dangling off to the side like spare prick at a wedding; where does it fit into the equation? why isn’t there integration between the Sun middleware and OpenOffice.org to the same degree as Microsoft with their server and client software?
Add to that the lack of any real leadership from the top to clarify to the market where Sun is heading – and it was a matter of time before someone saw the value in Sun which has been squandered by incompetent management for the last decade or more.
Edited 2009-03-26 06:06 UTC
I hope that IBM don’t indeed have the intention of EOLing Solaris/OpenSolaris as Otellini suggests…it can not be replaced by Linux, it has its own values that are unique which can not be found in Linux or AIX. Stable ABI’s being the most glaringly obvious.
And besides, the OpenSolaris ecosystem was just beginning to warm up. Adobe have been supporting with up-to-date versions of Acrobat Reader and Fluendo providing “legal” codec support as well.
perhaps it will do what HP did when it aquired Compaq and got OpenVMS out of the deal. it mantains HP-UX (which is garbage by comparison) and OpenVMS. Though personally i would not like IBM to buy Sun, I wish that Sun could just market their products better and make it on their own.
The bottom line is that stable ABIs haven’t been enough of a selling point for people over the years, and alas, that’s generally used as an excuse for the glacial pace that Solaris gets developed at. Linux distributors quite happily maintain ABI stability in their own distribution lines while the rest of the community still moves things on.
I doubt they’ll get rid of Solaris, but Sparc is probably a goner. Given that there’s a fairly substantial demand for Solaris on x86/x64, they’ll probably retain it on on that platform as a commodity, possibly even continuing to license it to other OEMs, and retain AIX on Power as the premium offering. Over time, they’ll probably be merged, or at least share a substantial feature set and API’s.
As for Linux, no doubt they’ll continue to happily support it on any platform you like, but it’s still a product they make you buy from 3rd parties – you have yet to see an IBM branded Linux – and somehow I don’t see them giving up their proprietary OS’s anytime soon. They still regard them as strategic platforms, and I doubt they’re crazy enough to bet the company on a platform they don’t control. If Linux loses it’s luster or takes a direction IBM doesn’t care for, they’re going to insure they’re still the masters of their own fate.
Well, it looks like I guessed right about IBM’s Solaris strategy. Although the accommodation for Sparc customers is a bit of a surprise.
http://news.cnet.com/8301-13924_3-10206753-64.html
At this point it doesn’t really matter for Sun if IBM buys them or not, either way, “they’re dead meat.”
How so?
IBM is great! Thank you for your good information!
Here I have another good share: Tradestead, there are many kinds of good things I like it very much!
Registered today, with one post, in Engrish, and recommending a share. Not to mention, a quick google search reveals that Tradestead is a Chinese company whos employees apparently have a habit of spamming online forums.
Hmmm…
There’s no way IBM will dump Solaris. When was the last time you’ve heard of anything remotely interesting happening with AIX? If any OS is going the way of the dodo – that would be AIX.
Even if you were right, it would be surprising for IBM to dump it’s own AIX instead of Sun’s Solaris.
It’s just so rare that big companies acknoledges that their product is not the best and that they should promote another technology instead… So I don’t see IBM doing so.
All proprietary operating systems are going the way of the Dodo, apart from Windows right now. It’s extremely difficult and costly to continue to develop your own OS with all that it entails now – kernel, drivers, support for ever more diverse hardware, userspace, tools, developer tools…… The burden needs to be shared.
I don’t think IBM really cares whether AIX keeps going or not, as long as they can contiue to sell hardware and do consultancy and support off the back of it. AIX is not the reason they exist. It’s is still around for historical and support reasons but the weight of development will swing in Linux’s favour. The problem was that Sun cared rather too much about Solaris and SPARC, which they never managed to find a niche for when the x86 juggernaut set in.
Humm, I’m surprised that nobody speaks about the Java platform?
When speaking about Sun, you all seems to only think of SPARC and Solaris… Me, I instantly think of the Java language.
Doesn’t it make sense for IBM — greatest ever supporter of Eclipse and other Java technologies — to buy Sun and thus the Java trademark? What do you think?
I don’t know for SPARC and Solaris, but for the Java world I can only see this as a good thing.
That’s probably because Sun never found a way to make money off the back of Java, and for a company that wanted people to stay on Solaris and SPARC I never understood why they thought the Write Once Run Anywhere philosophy would be good for them there. As Joel Spolsky said a while ago, I don’t know why you would want to try and make hardware a commodity (and the OS as well) when you’re pumping time, effort and money into SPARC to get people to stay on it.
Most of SUN’s services stack depends on Java. A big chunk of their business comes from selling big a$$ systems to run Java apps.
Claiming that SUN never figured out how to make money off Java is a tad facetious, don’t you think? Esp. when it has been their bread and butter in a portion of their data center approaches…
Edited 2009-03-26 18:48 UTC
Running stuff on Java (and I don’t think Sun went anywhere near far enough with it) and making some money directly off the Java platform itself are two very different things.
SEGEDUNUM
When you state that SUN doesnt make money out of Java, is it something you know, wish it would be so, or just sheer phantasies? It is quite often you speak about things you have no clue, you can never back your claims up with proof. Seriously, what are you doing here on OSAlert? You are clearly not IT knowledgeable.
You state things as a supercomputer and a large server are the same thing, but they are not. They have different architecture, even wikipedia says so, but you still you insist otherwise.
You insist that ZFS requires several GB of RAM (maybe it does on FreeBSD) but it does not on Solaris – hence ZFS does not need several GB of RAM. You state that ZFS data integrity is not important (which is quite frankly, stupid to say).
You state things as there are plenty of Linux servers with 32 cpus or 64 cpus – could you show me some links on this? You didnt last time I asked. I doubt this.
You state that VirtualBox is suitable for production, instead of VMware.
Most of the things you can not back up. It would be nice if you proved your claims, instead of just stating claims that are easy to prove wrong. For instance your claim that SUN doesnt make money out of Java:
11 march 2009, SUNs CEO blogs
http://blogs.sun.com/jonathan/
“For those that continue to ask if we make money with Java, the answer is yes, it’s on a ramp to hit about $250m this year – one of our best businesses – and that’s just Java on consumer devices, excluding servers”
$250m USD is hardly “not making money out of Java”, eh? Why do you state that SUN is not making money out of Java? Are you lying or are you ignorant? Or simply dumb?
I seriously wonder, how on earth did you cope with school? In academia you are expected to analyse things and argue, backing up your proofs. You dont seem to know anything about how to debate academically. How to prove your claims. I dont get it, are you just simulating dumb, or are you really dumb? I mean it, it is a serious question.
Sun’s bottom line is my ultimate proof, but where are the licenses Sun is selling for Java? Where are the licenses, consultancy or support that Sun is raking in for development tools off the back of Java? I’ll tell you where they are – nowhere. Zilch, zip, none.
All Sun did with Java was to create a reasonably successful development platform for their competitors (mostly IBM) that turned the things they really wanted to sell, Solaris and SPARC, into commodities because people could easily move Java applications off those platforms. None of this inane rambling has done anything to argue my main point regarding that.
Keep deluding yourself about the reality of the situation, and we’ve also been through the rest of your inane bullshit many times. Trying to rehash arguments again when you have the memory of a goldfish and can’t remember what has been explained and argued won’t make you right.
I can see you’re obviously deeply hurt by the predicament that Sun is in and obviously cannot fathom for the life of you how all this could possibly have happened, but I’m not an Agony Uncle.
SEGEDUNUM
I have several times requested that you back up you claims. If you can not, it is FUD and you are lying. You have never backed up any claim, despite I asked you several times. Unless you want to be taken seriously, you SHOULD be prepared to defend your claims with hard proof and links. Havent you learned that in school?
So now I ask you, you state that SUN is not making money out of Java. Back that up. I have proved otherwise.
You state that ZFS requires several GB of RAM. Back that up. On Solaris ZFS requires 512MB RAM.
You state there are several Linux servers with 32 and 64 CPUs. Back that up. Show us links.
You state that a large number crunching cluster has the same purpose as big iron, which is wrong. Even wikipedia says you are wrong (they have different archictecture and purpose: “clusters are very bad at general tasks”). Back that up. I have proved via wikipedia that you are wrong. Again.
Never ONCE have you backed anything up. I have asked you many many times, but still you write this things over and over again (never backing things up). You know, as a mathematician it is important to prove things, you are taught to never claim anything unless you can prove it.
If you want me to stop asking you about these things, it is easy. Just prove you are right, and I will stop. If you can not prove them, then YOU should stop claiming these things.
It is that simple. Exactly what is that is too difficult for you to understand? Point it out, and I can explain again, in other words.
“You state that a large number crunching cluster has the same purpose as big iron, which is wrong. Even wikipedia says you are wrong (they have different archictecture and purpose: “clusters are very bad at general tasks”). Back that up. I have proved via wikipedia that you are wrong. Again.”
Oops. replace “number crunching cluster” with “supercomputer”. The moron and I was discussing supercomputers. I, wikipedia and others, tried to explain that a supercomputer is not exchangeable with big iron, despite being faster. As wikipedia says: “supercomputers are bad at general tasks”. This means that you can not exchange big iron with a large cluster and expect the same behaviour. It is simply stupid to claim so. Even wikipedia says it is wrong.
And upon request, no links are presented. No proof of the moron’s claims.
Sun recommends at least 1 GB if you’re using ZFS.
http://opensolaris.org/os/community/zfs/docs/zfsadmin.pdf
http://www.solarisinternals.com/wiki/index.php/ZFS_Best_Practices_G…
Linux handles that number of CPUs fine. In fact, Sun even submitted some TPC-H results using a 60-core SunxFire X4100 system running Red Hat Linux.
http://www.tpc.org/tpch/results/tpch_result_detail.asp?id=107102901
If you check out the TPC-C section, you’ll see some other high number of processor systems running Linux.
But even so, for most applications it’s proven far more cost effective to scale horizontally, to use smaller, less expensive systems (i.e., x86) than to use monster 32-64 CPU systems. There aren’t that many 64-processor web or app servers, because it’s far cheaper to get 32 dual-core systems and stick a load balancer in front of them.
This is where Linux really killed Sun around 2001. Sun’s idea of a web server was a $30,000 E250. You could get a Linux or Windows server from HP or Dell for about $3,000 with roughly the same web serving power. Sun was very anti-x86 at the time, so people dropped SPARC hard.
The only systems that tend to use that many CPUs are database systems anymore. But with Oracle and other DB clustering as well as SAN storage, many are moving to x86-based DB clustering solutions.
So while yes, Sun has a good presence in the high number of processor area, those systems aren’t experiencing much market growth unfortunately.
Clusters, by nature, are typically geared towards a single task. Web cluster, app cluster, database cluster, high-performance computing cluster. 64-processor systems are usually similarly narrow-scope dedicated, and typically databases. Large, expensive systems typically aren’t considered “general use”.
Also, clusters give you horizontal scaling (need more power? add another server into the cluster) as well as redundancy. It really depends on the situation whether scaling vertically (adding more CPUs) or horizontally (adding more servers) is the better route. Horizontally is usually more cost effective, however.
$250 million? I’m assuming that’s revenue, and 250 million out of about $13.8 billion in revenue (FY’08) is a little less than 2%. While not chump change, it’s not a substantial portion of the bottom line.
Someone already answered that before me… :p
I am not trying to take side, but, AFAIK, Sun doesn’t make much money out of Java.
Err… From what I have read, Sun revenue per quarter is about a few billions. So, for a whole year, it should be around 10 billions.
Hence, $250m a year, is really very small contribution to the overall revenue. If, $250m is indeed one of their “best” businesses, then I think, it is a worrying sign for Sun.
The question is not about that. Some moron stated that SUN is not making money out of Java, without knowing anything about that matter. That is clearly wrong.
As for “best business”, you dont have all the facts, do you? It could be that the lion share of the income SUN has to fight very hard for. And it could be that the Java business is easy.
If I could sell air for a minor profit, that would be extremely good business. And if I could handle nuclear waste for a large amount of money, that could potentially be very risky and could punish me severely, with large potential liabilities later. Taking risk is not always a good thing, it could be bad business.
Well, I agree that Java and Solars/SPARC don’t have much to do at first sight.
However I’d guess that actually, what “Sun’s plan” were to make Java popular so that everybody would want to set up a Java application server, and then people would buy these servers preferably from Sun, since Sun would be the absolute reference for Java stuff.
Introducing a web technology makes sense for a company selling servers. Remember all the hype that Sun tried to make around Java applets at first? It looks like that was the original goal of Sun.
So, from some point of view, it may have been some valid strategy for Sun at that time.
Even if now we see that it failed: from what I see Java applications are quite common but still it’s very rare to find embedded Java applets in web pages…
Part of the CDDL – that Solaris is licensed under – states that Sun can relicense without permission from any developers that have contributed. I predict IBM will exercise this option and work on integrating the interesting technology into Linux itself.
That is one obstacle that Linus has with the GPL that many overlooked when it was a big deal that he said he likely wouldn’t switch to GPLv3… he didn’t really have any say because the copyrights stayed with the original author, he just owns the trademarks on the brand his pet project has developed into.
Anyway, hopefully I’m not rushing the gun by already welcoming the Solaris guys into the Linux community proper…
Edited 2009-03-26 15:24 UTC
No doubt the Solaris guys are just as enthusiastic as you are…
Sorry, but that is so incredibly wrong it’s not even funny. The CDDL does not give Sun any special rights to re-license code placed under the CDDL.
In fact, the CDDL is a generic, reusable license that is basically an improved version of the MPL (Mozilla Public License).
The Sun Contributor Agreement, that gives Sun join copyright on contributions, is what gives them that right.
Edited 2009-03-28 08:39 UTC
This means ZFS will be GPL ?
There’s a great article on Sun from 1999 in Fortune:
http://www.forbes.com/global/1999/1129/0224102a.html
“IBM’s problem isn’t Y2K, it’s S-U-N.”
Talk about role reversal.
I wonder what will become of niagara.
Things don’t look good, but I think it’s a little too early for all the doom and gloom. Sun has an infamous track record for bad management and marketing. What they need is some competent people in management who can handle an IT sector increasingly moving away from traditional big iron. They’re still stuck in this one mindset. They should be pushing Java in other markets where it could easily dominate and targeting low end clusters and servers. In any case I hope they stick around for a good long while. SPARC is really interesting and if the some of the talk on the tubes is correct I can’t wait to see Niagra 3.