Slashdot Mirror
The Faded Sun
jlowery writes "Robert X. Cringely seems to
think so. Forget the hardware side: what does this mean to the future of Java? Will there be enough incentive to continue to develop the language for whoever acquires Sun? Or will Java developers have to swallow hard and submit to the whims of the dark overlord? Maybe I'll switch to Mac development, after all."
The Java specs are done by the Java comminity process, if Sun goes down (I really hope not and I will be one of the first to jump on their desktop machines) someobody else (possibly IBM) will take over Suns role in the JCP. There is too much investement especially on IBMs side to let it go.
They have just renewed a commitment to the Solaris operating system, which is no longer really viable from an economic standpoint. I know, I know, Solaris users love Solaris, but they don't love Solaris prices.
This statement is wrong. The cost of Solaris is not an issue. Solaris licenses are either free or cheap depending on what kind of hardware you own and where you got it. The real problem is in the cost of Sun's hardware as well as relative performance of UltraSPARC processors compared to the 32-bit x86 processors and certain 64-bit processors. Sun executives are still living in an imaginary world thinking that Sun's future is in selling large mega-bucks systems to the data centers completely ignoring the low-profit high-volume low-end side of the market.
I don't think it is that abstract though. .NET framework might not be that bad, or even very good.
Purely technical, the
The problem is not with the technical side of Microsoft, but the contracts and legal-issues associated with licensing their software.
Microsoft has a tendency to create contracts and agreements which bind you not only by hands and feet, but which will also "dictate" a predefined Microsoft-approved-certified-blahblah direction.
Ok, his first points are very valid and I will agree. Sun is in serious trouble. They're betting the company on N1. Apple won't buy them. Java wasn't the smoking gun.
/. folk, agree - they're not doing the right thing.
But to say that a merger with Sony would be better than Apple is just plain dumb. What have the two in common? Absolutely nothing. Sony has no interest in the server market - if they had they'd be there already. Furthermore, the technology that Sun pioneers has absolutely NOTHING to do with ANY Sony market.
He article further states that N1 puts Sun in direct competition with Microsoft and IBM. Uhh, hello, where you have been dude - they've been in competition for a long time now. If he is trying to draw comparison between N1 and a MS or IBM product then he should do so. From what I've read N1 has a LOT of potential. And while IBM is certainly a contender in the distributed-computing area, MS is definitely not.
Although Cringley was clearly drunk when he wrote this, he makes good points. And I would agree that N1 is certain to fail. Not because it won't perform, or not because Sun is actually using sales people to sell it, but because the market is rather slim. N1 doesn't benefit a small or medium sized company very much. Not nearly as much as it does the enterprise.
I don't know what Sun should be doing right now. But I, and I bet a lot of you
Perhaps the problem is that the distinctions in the computer market have altered and Sun has no place for its hardware? It used to be that there were servers/mainframes, workstations and then puny PCs. PCs have grown in capability, however, essentially absorbing the workstation definition and market, leaving Sun with little room in that segment. IBM chose to make servers the core of its business, while Microsoft and Intel dominate the PC market.
.NET and the hype machine is in full force - and still Sun has failed, to my knowledge, to respond.
For quite a while I've been wondering exactly what Sun is up to. They calmly sat back while people kept repeating the mantra that Java is slow (even though it isn't; JIT-ted code and better GUI techniques improve performance markedly), allowing it to lose mindshare to competing products. Now Microsoft has shipped
Even if Cringley's article is wildly inaccurate, it does reflect the concerns and questions of a number of people, particularly those who do not use Java as part of their job. What the hell is Sun doing?
It is a shame what's happening to Sun, because it's indicative of what is happening to computing in general. Sun's old machines were solid, powerful machines that just worked. I used a sparcstation 5 up until a year ago at work, and while it was dog slow, it still worked all the time, because it was built at a time and for a market that expected that computers *worked*.
Now, thanks to the whole IBM PC/Windows thing, when a computer crashes, people say "oh, that's ok, that's what computers do" and hit reset. I'm not saying I'd rather have a blade100 on my desk than a wintel box, but I wish that my winel box had some of the engineering quality from Sun.
IBM has a lot invested in Java. It's become their common development platform for their various OS's they run from Linux on up. Native code for the heavy-duty stuff, Java for everything else. Probably saves them billions a year.
I think if Sun burns up (and with numbers like $2 billion in losses, it could happen overnight, look at Enron/WorldCom... who knows what sort of tricks are being played with the books) IBM would be the first in line to grab Java.
Just my best guess.
-Russ
Me
I must agree that Sun is really in trouble. Solaris is not the blame but the hardware price and performance is.
I work in a semiconductors startup. Two years ago when the company was founded Sun hardware was the default when it came to choosing CAD servers. Sun even had a nice discount program for startups.
These days we can get a fast Pentium4 or Athlon (running Linux) to do the same work for a lot less $$$. Maintenance is also much cheaper.
All the big CAD software vendors now support i386 Linux and the platform is stable and FAST!
In fact, the only reason Sun hardware is still worth keeping around is because it supports large (>4GB) memory. When somebody finds a way around that (AMD Hammer comes to mind) Sun will loose its last asset.
It's a pitty, cause Sun is a good enigneering company. They invest heavily in research and are a major source of innovation.
They just can't keep up with the falling prices of that huge i386 market. No one can (not even Intel's own Itanic!)
You can't (well you can, but I'd rather drill a hole in my hand) script an enterprise app in PHP. Yah, PHP is great for a scripting language, but it's just that.
:)
I grow tired of everyone predicting or shouting for one thing over the other - there always has to be just one. Yah, right.
PHP is great for the non-ASP/*nix programmer. ASP (and I'm choking a little here when I say this) is great for the m$ programmer. C is good. Java is good. Jeeze, they all have their strengths and weaknesses. I'd much rather have a CHOICE when using a particular technology than not.
The Java VM exists for a reason. Just because PHP doesn't have one doesn't mean much. They're both written in C too - so what!
In the end, sure the user wants the most responsive app. But I'll say this, get a big project and try to have multiple devlopers script it and it'll probably die on the vine. You can do just as bad of a job with JSP (and believe me, I've seen it) but there are some really great frameworks out there that help fix problems like this.
Plus, with PHP and the like, they're tied to HTTP. It wouldn't be a very good idea to script a server app in PHP with multiple different types of clients accessing it. It's possible, but I can't see someone writing a Win or Linux native client that accesses a PHP server app. Java works well with the web, but is not build solely for it.
Plus there are other things, if you wanna compare (I don't know even why I'm doing this). There is no PHP message queueing, no or little 'enterprise features', no 'compile PHP to a console application', no PHP 'enterprise' transactional components, etc. Anyway, anyone who's ever had to really use both knows what I'm talkin about.
And besides that, for me, *nix and network programming are still like wide open spaces to me. There are still plenty of things to discover out on the Montana plains and I'm not gonna get all bent outta shape about a rock not being a tree and a tree not being a clear blue sky.
The segway; Sun is all about proprietary (13W3, sbus, solaris) - this is why they may make a great pairing with Sony. Sun could be to Sony, what the Xserve market is to Apple. It could be Sony's opportunity to be recognized in the corporate world. Apple and Sony share the exact same "creative market" - those don't buy Apple in the market, tend to want Sony A) brand, B )it's the same brand as their other equipment, C) Looks, D) Integration & Media nature of their product
Conversely, I have always thought Sun would be a good merge with Apple. I think Sun would be best getting away from almost TOTAL proprietary, allow Apple's genius to help with development of Java and further integrate it in to Unix/BSD, and give Apple some of the best blade technology in the industry, and possibly a stronger development partner for RISC processors.
I had even come up with a good slogan for a Steve Jobs Keynote; "Every Apple Needs a Little Sun To Grow!"
Yell & scream & rant & rave... it's no use... you need a shaaaave ~ Bugs Bunny
Your assertion is that the way to succeed is to "use inhouse". The R&D at your company must be phenomenal! Motorola thinks that way, I hope they make it!
Yell & scream & rant & rave... it's no use... you need a shaaaave ~ Bugs Bunny
I think what we've seen is that Sun had a couple of fantastic years, fueled and fanned by the dot-com boom. Remember the "we're the dot in dot-com" campaign?" That is just one example of Sun's high caliber marketing folks capitalizing on the times. If you've met with any Sun sales reps lately, you'll know that the new party line is "let us show you how we can save you some money" -- exactly what they should be saying in a down economy. Their marketing tactics are still as sharp as ever.
Believe it or not, much of IT management still subscribes to the belief that "you get what you pay for." If you work in the field and have ever suggested MySQL to an Oracle shop, PHP to an ASP or JSP shop, or Linux to a Windows/Solaris/HP-UX/AIX/SGI shop, you've heard that statement. The exceptions are software like Apache, which is nearly ubiquitous, however if you look at BEA's marketing and their broad marketshare, you might even say that they're making in-roads convincing IT management that web & app servers follow the same rule: you get what you pay for. Why is JBoss still only a developer's choice and not the enterprise's? (http://news.com.com/2100-1001-984476.html?tag=fd_ lede1_hed)
McNealy is in no danger of being replaced; he'll adapt and overcome. Sun is not resting on their laurels, and have never forgotten their base -- enterprise datacenters -- as evidenced with new product lines like N1.
I don't argue that you're very unlikely to run a Sun workstation at home on your DSL connection. Linux is just too damned good (price performance) for the Unix-oriented home user, and getting better. But for enterprises, you get what you pay for.
I think to understand what is happening to Sun one has to look at the market.
Just recently the Java vendors have noticed that Open Source has taken their share of sales. And this makes them nervous. Sun, which has fought with Open Source is falling into the same problem.
Open Source is making inroads into the market. And the problem with Open Source is that it kills, nullifies the traditional software market. By traditional I mean give money get software. Open Source opens different markets and some people are coping, eg IBM or RedHat, etc.
When I see this I truly do see the end of days of MS. That is IF MS goes this route. If MS decides to accept Open Source, then things will probably change. But I see MS going the same route as Wang. Wang a now dead company that had truly interesting technology and products. But a company that failed to adapt to changing times. But before that happens it is going to get TRULY messy with IP (in America thought).
"You can't make a race horse of a pig"
"No," said Samuel, "but you can make very fast pig"
Sun products are not any *special* improvement of Sony products and vice versa. Sony products usually do not need any servers, but when they do - other cheaper and still good servers around (read: Linux). And of course, who needs any Sony products in the server room?
Apple has already been bound to IBM and Motorlla through PowerPC. Typical Mac applications (graphics) doesn't require big database and internet servers, where Sun is still strong. In fact, typical Mac users are geeks (it was so hard to avoid typing "jerks") and not corporate users. But if otherwise would be - IBM servers are not far away.
Both Sony and Apple has no traditins of picking up failing former giants and digesting their dead meat.
But there is other company, which has very long tradition of squizing the last juice from the dying things: HP. They just bought Compaq who bought DEC. Why do they do that? Because their business model is based on support, specifically on supporting customers with legacy system, who doesn't have (almost) any other choice to get that support of their already dead platforms. But that business model requires new victims every few years.
Besides dying expensive hardware, Sun and HP has another in common: system management. Both have good ideas, both did not implement it well, at least as good as IBM did. So, by combining system management platforms from both, Sun and HP can make them a stronger competitor to IBM on that market segment.
As for Java... Sun will let HP to suck the last possible money from IBM on Java licensing. Of course untill IBM will drop Java finally and move to Python (I would love to see Eclipse for Python!). And I won't be surpised to hear that HP or Sun or merged HP-Sun, will buy Borland together with Together :)
Personally, I can bet that if in coming year we won't hear about upcoming plans of HP-Sun aquisition, then we shall hear about HP planning to acquire SGI. But any speculation about that merge would be a kind of offtopic here.
Less is more !
The real problem with this analisys is best seen here:
:))
Cheap Intel and AMD hardware running Linux is going to kill Sun unless the company does something so stop it, which they aren't
This comment is so clearly comming from a journalist, and not from someone who ever done some real work. To explain:
x86 is a 32-bit architecture. It has a 4G RAM limit which you just can't help. There are ways to put more than that - but you can't have a single process which uses more than that. There are also problems with the fact that x86 doesn't scale well above 4-8 CPU's and others - but this is the main issue. Now days there are simply too many problems which require datasets larger than 4G, both on workstation, and on servers - scientific, database, just to name a few. Now we all now that both intel and AMD are going 64, but... Would you base your buissines on a brand new platform, no one tried? Well, neither would most others.
There are other issues here - sun is known for great service and reliability of their machines, but the main issue remains within these 64-bit. PC's, and mac's (for now) just can't cope with that. There are moves forward, but no one has tried them - and that's the bottom line.
Even Java is becoming superfluous. Java is the Dan Marino of software. Just as the former Dolphins quarterback, Java affected the world so much that history cannot be written without its mention. But nonetheless, neither Java nor Dan ever won the big one.
Blasphemy! I saw Mr. McNealy speak at JavaOne last year, and as he remarked that Java had now become the most widely used programming language, he put up a slide saying "Java Won!". It's everywhere! How can this fool say that it never "won the big one"? Since 1.4 was released, all the objections to its use have been made irrelevant: speed (thank you, HotSpot), user interface (Swing now really does look and feel the same on all platforms)... well, I can't really think of any other objections, anyway. Bottom line: be as negative as you want about Sun, but Java is not in trouble, it rules the world, from cell phones to mainframes!
So, where have you been in the last 5 or so years?! If you read Judge Mott's ruling about the Sun vs. M$ case it clearly shows that m$ abused yet another contract with another company.
.NET, and C# for that matter, might be a better development platform than VB and such, but they still have the same 'fuck the world at everyone's expense' mentality behind it. I've learned that the hard way.. Good luck!
The reason he is forcing them to bundle it is because they did cause harm to Java in more ways than one and that deserves a correction now instead of worrying about it later; as was the case of Netscape. Argh! Ya know, you need to get your shit straight and present the facts when posting stuff like this. Sure, someone reading your post would be like "yah, market forces dude! What are we now, commies forcing successful companies to carry stuff they don't want?"
Gee, and oh yah, weren't they already convicted of being a freaking monoploy?!?! Um...
I've been developing with Java for 7 years and while it's not perfect, I'll never go back to m$ development for many reasons.
- They're a highly unethical company. for a small taste see here, here, and here...
- Like someone else said, profits first, users second
- Welcome to a m$ only development world. I applaud the Mono guys and the dotGNU guys, but just wait until m$ wants to flex its patent muscle. Profits first, lock in second, whatever after
- Bugs & security. Welcome to the jungle
- Horrible, god awful, slave, er, customer service
- Service pack # 539.. and counting..
- Worms, attacks galore and shotty patch record. Just that alone would make me steer clear of that platform.
Java isn't the holy grail and I'm not looking for that. But its developer community is much better than M$s' and has its roots in the *nix world, which frankly, I'd rather have my foot in than the DOS world.
Oh, and by the way, before I bailed from ASP/VB/<fill in other m$ crap here>, I was in a constant state of perpetual screwed-ness with M$ products!
Not only too much IBM investment, there is too much IBM CUSTOMER investment. Now that the big financial institutions have implemented Java apps it will never die. Look at COBOL.
While Sun has tried to compete with MS and IBM, they haven't been successful. For one, they have no application server. Yeah, they have "iplanet", but who is doing development for this platform? Nobody. Currently, enterprises are deciding between .Net and Websphere for the most part. Many companies are also consolidating their hardware purchases, and I can tell you, they are not consolidating it into a bunch of Solaris boxes. Sure, Sun may have a stronghold here and there, but on the whole companies are getting rid of their Sun machines. Ford, for example, just decided to turn their entire hardware purchasing to IBM. That means the several thousand workstations and servers they have from Sun will all be gone in a year or two.
Java, contrary to what some of you have posted, has a solid piece of the market. This is largely due to companies like IBM, who have provided compelling development tools and deployment platforms. Many enterprises are looking at Java as a way to keep their business logic in a relatively vendor-neutral place. So, Java is not going anywhere. But, I can't see that it is going to do Sun much good. Amazingly, Sun has some of the worst Java tools available.
So, Cringley's article is not a troll or a flame, he makes some sober observations about a company that has been losing relavance for some time. He is absolutely correct when he says they need a visionary. Joy and all those other guys are not filling that role. Sun needs a Bill Gates or a Steve Jobs. Someone who will wake that company up and say, "Screw servers, we're losing our ass there. Let's do embedded devices!", or something. Anything but what they have been doing for the last several years.
I read through this thread and the thing that I kept thinking was: "These kids haven't worked on an RS/6000 recently have they?" Now I will admit that I was indoctrinated into AIX goodness over a year ago, but with Linux/Mac/Windows eating into the workstation market (which is taking market from ALL the players in that space) With the rise of linux/gnu stealing profit from the OS and developer tools business, and now IBM mounting a full frontal assault on the high-end.
To the poster upthread, Sun hasn't captured the hearts and minds of CIO's, Oracle and SAP have. When/If Oracle ever migrates to a different platform for their primary development that will be the nail in Sun's coffin. That's really what is keeping Sun in business these days. The question for the next 5 years is if Sun can transition it's business model from providing expensive big-iron as their primary money maker, to competing with IBM in the high-end and Linux in the low end. They're in somewhat of Apple's problem. Their processor has lagged against the POWER series at the high end, and the PIII/IV at the low end. The Mid-sized market is such a cut-throat environment that Sun can compete there, but it's a hard sell for everyone. As much as we hate monopoloies, Sun needs to find a niche that it can dominate (like SGI) to have a rudder to give it stability.
What if it is just turtles all the way down?
as an aside. i think everyone out there who has contact with sun somehow sees this one in the works. i have a cobalt server, and the guy who was the engineer in charge of the list got canned. i'm sure they're aware of the implications of keeping managers and canning engineers, but you can only get away with so much of it before your company goes, POOF, and becomes a Geek Story. "There used to be this company that made . . . .. " We'll miss em, hope that maybe IBM buys em out or somethin.
"You never want a serious crisis to go to waste." - Rahm Emanuel
What have they done for the company recently?
Seriously, Gosling has been involved in a lot of visionary technology before Java, but none of it got anywhere. NeWs was squished by X-Windows. Gage did net day, but what has he done for the company recently? Come to that what does Whitt Diffie do for Sun beyond consume cafe latte?
Unfortunately there is a major difference between technological firepower and technological leadership. The problem isn't with the technologists, it is with the management. They have simply failled to construct a business plan or environment that can utilise the firepower they have.
In that sense, Sun invented workstation.
My DEC Alpha was far superior to anything sun had to offer. Come to that SGI provided better firepower and a slicker integration package. Sun invented the cheap engineering workstation, mainly for the education market. Real engineers used VAXen. Now VMS didn't survive too well but it was the DEC/MIT X-Windows system that defined the workstation interface in the end.
As the author claims, Sun might be gone; on the other hand, Sun might be ruling the world by then.
I doubt it. IBM is rulling the commercial java space and OSS is rulling the freeware space. There is not much of a gap between the two.
The apple/Sun issue is key here. Apple is very well positioned to take huge bites out of Sun's core server market. They simply don't need Sun technology at this point. All they need is a hot processor - which sun notably lacks.
For Sun to survive it has to start focussing on its business, not Microsoft. Meetings with Sun engineers are painful, you get a 45 minute whinge about Microsoft. Which is pretty sad when they know you are one of Microsoft's closest allies in the industry. Even if Sun makes a billion in the lawsuit they will lose big, the suit is costing them far more than that in lost business and lost opportunities.
The first step to save Sun is to sack McNealy. Unfortunately Sun does not have a Steve Jobs figure waiting in the wings.
Looking for an Information Security student project suggestion?
Try http://dotcrimeManifesto.com/
Sparc isn't proprietary. That's why Fujitsu makes Sparc systems that run Sparc Solaris.
http://www.sparc.org
No one can tell what's going to happen in this arena. You've got "only" five years, so you are dead. That sounds too premature. Anything can happen in five years in IT industry. As the author claims, Sun might be gone; on the other hand, Sun might be ruling the world by then.
Amen. To put this in perspective, 5 years ago the dot-com boom was just getting off the ground. If you think you can predict *THIS* industry, I have some stock options I would like to sell you.
My beliefs do not require that you agree with them.
Of course there's a high end, and there's also a healthy mid-range segment.
Cray is a totally different market really, scientific supercomputing. That market still exists and is as large as ever, it's just not a growth market like PCs. However, the low end is extremely marginalized and hard to profit from.
The CPU speed of Sun hardware is only a small part of the equation in the enterprise market. That you even single it out without talking about I/O, service contracts, or other more important issues, indicates to me that your experience is not in the mid to high end enterprise market.
-Kevin
You're describing a classic retreat to the high end.
Sun may not go away, but their market share will continue to shrink and shrink and shrink as the lower end machines gain in capabilities.
Anyone care to confirm the facts mentioned?
The basic facts - and Sun's basic problem - are that Intel/AMD hardware is cheaper than SPARC hardware because of economies of scale; and that Linux is comparable with Solaris (behind in some areas, ahead in others). Do you really need confirmation of these facts?
Right now, it would be impossible to replace all Sun servers by Intel hardware because Sun makes "big iron" - multiprocessors with 64 cpus. A big bank, for example, that has to process hundreds of millions of transactions at month-end needs that performance. But inevitably, Intel hardware will become available in this kind of configuration, and Linux will support it. It's not a question of "whether", it's a question of "how soon". This is clearly a threat to Sun's business model.
Who cares?
I believe that one of the causes of the dot-com implosion is that many companies discovered that their customers will actually put up with pretty crappy service. And therefore the market for co-location services and monster data centers never actually appeared and companies like Exodus were doomed. 24/7 uptime just isn't needed by that many companies.
Why pay big bucks for hardware support on a box from Sun when you can buy 5 cheapo boxes for the same price and have your own in-house monkeys do the board swapping within one hour rather than waiting for board-swapping monkeys from Sun that might not actually show up within an hour anyway?
That may be true but HP is not Sun. When I worked in a Sun shop if we had a hardware problem the techs were on-site WITH the parts they needed in under an hour. It was pretty much the same with the Cisco support contract we had.
If you pay the big bucks, you CAN have good service. And there will always be companies that pay the big bucks to have a higher measure of reliability (and lower downtime). If the downtime costs you more than the support contract, it is a smart buy.
Just because HP has crappy service doesn't mean everyone in the segment does as well.
Enigma
"some accountancy stuff that mean diddly to the companies operations" ?
Losing $2 Billion in asset or "paper" value is a huge deal for where it counts. Banks, investors, fund advisors, et. al. will never loan money to a monster company like this that loses value so rapidly. It's the equivalent of your house going down in value from $500,000 to $50,000 -- sure you still have the house and will get a lot of use from it, but you couldn't even get a Kia car loan with that. And if you think that Sun needs to make some drastic changes quick, they just lost alot of flexibility to do so.
No, $1.5 billion in cash and "cash equivalents" may not be enough to turn this once great company around. What that does is let it keep doing what its doing for the next x years as they slowly wither away (think SGI).
Unless somebody bets the farm, which is what I hope this thread will spur McNealy and crew to do. N1 isn't it and ironically IBM has taken the "king of Java" crown away.
The problem that Sun has aways had is the "reality distortion field". The Sun marketing department seriously thought that they were going to eat Microsoft's lunch early on--even though the Sun management just didn't get what GUI intensive systems(i.e. Macintosh and Windows) were about. I wrote one of the early reports for the directors to explain what stuff like Hypercard and Visual Basic were going to do--they just didn't get it. Another big problem Sun had was when Linux started to get developed. The Sun management just didn't get how Linux would impact the market.
Sun was in serious trouble years ago. The H-1b exansion at Sun was largely a means of covering up the problems that Sun management had created for themselves over time. The poor track record of Sun management had created a situation where Sun just couldn't hire the best younger American talent--Sun management rolled the dice bringing in the H-1b's and guess what, they rolled snake-eyes. The H-1b experiment at Sun has gone on 5 years--it just didn't help the company. Now what it _may_ have done is enabled the large shareholders from 5 years ago to sell their stock and leave someone else holding the bag.
I see no particular reason why folks like Scott McNealy, Vinod Kholsa and John Doehr(the CEO, co-cofounder and venture capitalist the funded Sun) should have any serious credibility with anyone. What is the real track record of these guys? What really happened to the folks that listened to these folks(investing their money and/or careers)? The real lasting legacy that I can see from this whole period is that Vinod Kholsa helped get a lot of his co-nationals green cards.
Business models based around giving stuff away for free aren't a good way to make money. Sun spends a ton of money developing Java, and then giving a lot of it away to people who haven't given them a cent.
Microsoft gives a lot of stuff away too, but the stuff they give away generally only works on platforms that you had to give them money to buy. Anything that Microsoft gives away for free has a chance of making them money, while Sun has a much much slimmer chance of the same.
"The database layer is another story entirely. They require a large amount of horsepower, and the whole operation dies if they go down. This means it *must* sit on big, reliable hardware with a support contract"
Solution: Clustered Oracle RAC Enterprise with Redhat AS 2.1 and full support contracts from Redhat and Oracle.
We are planning on this deployment simply because we can save hand over fist amounts of money. You can read all the fud about it from Oracle, but IMHO, it 'could' put a big hit on big iron.
Bye!
So what you're saying is, "Don't trust mission critical apps to dodgy Intel hardware, trust it to buggy Sun hardware"; after all, it isn't like Sun tried to cover up the data corruption and halting problems afflicting those systems, is it?
IBM. They've been doing it for years, you know. Given the cost of Sun's platinum support, there are any number of companies who would provide excellent support for you.
People used to say the same about SGI and the high end graphics market. IBM and HP both have products that slot into exactly the same area - and IBMs top end gear is a fuckload more reliable than Sun's.
A cautionary tale:
"Those whom have not studied history are bound to repeat it". Santayana
I was at SGI in 95. We read a number of articles from Red Herring and others on the need to find a new business because our old one was gone.
Execs laughed at it, pointed it out to be false. The company was doing great. TJ promised wall street 50% growth. Gary Lauer chuckled. Ed smirked.
Supporters lambasted the article, the magazine, the author. Other similar tomes appeared. More supporters screamed bloody murder. They pointed out all the problems with the articles, the physical shortcomings of the authors are "real men", and all manner of other vituperative attacks.
After all, SGI has some of the best technology around. It *couldn't* die.
Then the company death-spiraled.
They are about a year or two from final implosion. They have no physical assets, having sold those off to forestall bankruptcy. They sold their IP to others (microsoft among them) to keep afloat. Their best people left. Management remains largely clueless.
So why I am writing this about SGI when the original article was about Sun?
Simple. Substitute Sun for SGI, exec name switches, and slightly alter the facts (Sun does not have killer or even compelling technology, SGI did stuff that Sun still has wet dreams about doing).
The reality is that the Linux based machines, with IBM Global support (or HP support for that matter) behind them are far more compelling from a performance and price scenario than the Sun machines at the lower end. There goes one of McNealy's strategims. Sun used to ship absolute crap to our customer base at 1/3 the price forcing us to discount deeply. Our customers loved our machines but used Sun as a whack-a-mole 2x4 against us. I recall warning one Fortune 100 customer that if they did not perceive our value that they should consider doing without us. They said that they loved the machines, but not the price and that they would continue to whack us with the 2x4. I indicated that we could no longer afford to fund their purchases. About a year ago, this customer came back to me and indicated that they were very unhappy with their crappy Suns and that they wanted SGI back. I indicated that I was no longer with SGI, and that SGI could probably not afford them as a customer. McNealy scorched the earth, set the concept of paying as little as possible for the value, and generally destroying the workstation market. Sure NT had something to do with this, but that became the 2x4 with which to whack Sun. Now it is Linux which is even cheaper.
Sun sowed the seeds for this. It is currently reaping the rewards of what it planted in the customers minds.
I give them 3 years before stuff looks real bad, 5 years before they are bought (or 7 before they go under). Based upon similar experiences with SGI.
Sure, you can disagree with points, predictions, etc. The reality is that Sun is troubled. They have been troubled since the dot-bomb went boom. Their market is being encroached on from below by Linux. Not windows. From the sides by IBM and HP (though HP has its own problems in the Itanic).
Take a lesson from the SGI playbook and learn from their mistakes. The retreat to the high end guarantees death sooner rather than later. The market is always going downstream. You have to innovate and lead to be relevant on this high end, and that is a small market (see the IDC and others numbers). Sun is not known for innovation (Java is not a shining example of it).
One of the problems with this is the massive business case dependence upon Java. Java is, despite protestations to the contrary, a Sun product. There is input into the "community process" but Sun owns the trademarks, the IP, and so forth. This means that when (not it, but when) Sun dies Java's future is quite uncertain. The only way to fix this is to separate Java from Sun and make it a real ISO/ECMA standard. Do it like OpenGL or similar. Sure Sun will play a role. But when it goes away, others can step in and provide continuity. As long as Java remains a Sun product, this will not happen. OpenGL is going strong, and will survive SGI's imminent demise. That is because of this process.
If you are or know a CIO, they need to be quite concerned about this for Java based software. Sun has serious existential problems ahead of it (regardless of the nay-sayers or supporters positions). It is not at all clear that they will survive these challenges. Java and its future are currently tied to Sun. This is a problem, a business risk. It needs to be assessed and weighed.