Slashdot Mirror
Sun's Last Stand
non writes "Wired has an article by Gary Revlin in the July edition about the current state of affairs at Sun. He attributes half of Sun's problems to failure to recognize the emergence of Linux, and the other half to their failure to make up with Microsoft, and finishes up with a server price comparison. An interesting read."
Sun's current "low-end" tactic of trying to replace Linux with Sun on x86 is going to win a lot of converts. There are a lot of applications out there and companies that are used to Solaris and that installed base isn't going to just go away peacefully.
The biggest argument for converting servers to smaller x86 boxes has been scalability and cost. Linux is a popular way to do that, but many companies have been using various BSD variants as well because they are more comfortable with server vs. desktop oriented software. Sun will do very well in those areas with their new emphasis.
For a company that wants to keep their big hardware on Solaris for some stuff, it makes a lot of sense to standardize on Solaris for their cheap x86 servers as well.
The party of stupid and the party of evil get together and do something both stupid and evil, then call it bipartisan.
as the leading Unix server seller in the last few months?
I am the Alpha and the Omega-3
well, obviously it's easier for someone to cut their teeth on linux, since anyone can download a distro and install it. not so easy with sun/solaris.
but, you're still not going to convince a large, lumbering manufacturing company running decidely unsexy applications to switch to linux. i've worked with a number of clients that couldn't switch from solaris even if they wanted to, because their apps just don't exist for linux (think erp/mrp systems).
plus, if you need lots of processing power, you can certainly set up a cluster of cheap lintel boxes, but why spend the time/money on that when you could just throw an enterprise server in there, and just have it work?
Whereof we cannot speak, thereof we must be silent. --Ludwig Wittgenstein
Focusing on beating Microsoft in any way possible might actually *not* be as effective as innovating and creating products people want to use.
Sun's anti-MS strategy was quite interesting (e.g. it was quite a bit more innovative than just reimplementing the GUI part of windows on top of a big teetering stack of different projects
Whence? Hence. Whither? Thither.
It's servers are too expensive for most tasks. They might be better, but the value for your $$$ is on x86. On the OS side Windows and Linux are kicking Sun's ass. On the hardware side Intel and Dell have created an efficient business model that is increasingly moving up higher and higher in the enterprise.
Sun needs to figure out a business model that will work in the new economic reality. They will either need to be a software company or a hardware one. But like a lot of companies they will probably die off because they couldn't adapt. They were successful once because they filled a market need, but when the market changed they couldn't adapt fast enough.
Ironically, at the same time my knewsticker told me about this article on /., I caught this article about how Sun just scored a big deal with HBO.
This should be modded off the front page.
I thought the summary already said what you're saying... and the other half to their failure to make up with Microsoft
What crack are you smoking? If you compare a 4-way V480 with 8Gbytes of RAM, it's a bit cheaper than that IBM (and comes with an enterprise class OS too): $34,995.00
...and you'd probably be running Solaris 9 on it nowadays, not Solaris 8.
Stick Men
Actually, once IBM spent US$1 billion to port Linux over to their AS/400 and mainframe hardware, the writing was on the wall for Sun: IBM will not take Sun's threat sitting down.
I personally believe much of Linux's rapid acceptance for large-scale computing needs is due to the very fact IBM big iron hardware can run Linux easily today.
What galls Sun quite a lot is the fact IBM's own development tools for Java are vastly preferred over Sun's own development tools. I think IBM should just buy out Java from Sun and save developers a lot of grief. =)
We have got Suns and Linux machines in the datacenter.
Where is really matters (big database engines, 4+ CPUs, a lot of external storage) Linux/Intel is just not capable of the task. Sun plainly does not have much competition there. At least not from Linux. HP-UX, AIX -- may be (though not here).
What are these whacky analysts talking about? What Linux? 8-CPU, 64bit, fibre storage attached and Linux? Have you ever tried it? I have, I know what a pain it is. It DOES NOT FLY. Period.
What REALLY hurts Sun is Windows on the low end. Not the hardware, not the price, but all these litty-bitty apps, that do not work anywhere but on Windows. Espetially Web apps. All these moronic developers with only Windows experience and mantra "does not work -- reboot it!", "open MS-DOS command prompt window and type c:".
There is Sun's biggest problem. They are lacking in the software, not the hardware.
Their problem maybe slashdot.org
They are innovative (Java), competitive (low cost Linux), open minded (open source, Linux). They have a large share of developers. They created Java. They support Linux. They make profit during the down turn of the economy.
...because it is what I know.
/etc might be moved around,
You might be doing yourself a disservice, and maybe selling yourself short.
File structures may be different, configuring stuff in
but the diff between Solaris (or IRIX, or OS X, or AIX) and Linux isn't any wider
than variants in Linux distributions. Just running an unfamiliar shell on a Linux
box could make it seem like a foreign machine if that's not what you're used to.
What's easiest for you also might be blinding you to choosing the best box for
the task at hand. I think Solaris tends to have more "torque" under load than
Linux, OS X is better at interoperability with other systems, and IRIX...well, no comment.
There's also the factor of uptime and service contract support. If you spend the bucks,
when a Sun box breaks, they'll get things back in order quickly. Sure you could run
down to Best Buy and get parts for your Linux box, but some places feel uncomfortable
with that, as they should.
Not that I like Sun all that much. They could use some of the modern polish that
Apple has put into OS X. Sun maintenance and installation feels very dated to me.
But they do run and run and run and most anything Linux can do, in the server world,
can be done (often better) with Solaris.
Netcraft's site discusses how they gather the uptime, and states that Solaris, Linux, and HP-UX (maybe more) all will max out in those charts at 497 days, due to limitations in whatever they're using to measure uptime remotely.
The *BSD boxes don't have that limitation, it seems.
I wonder how many boxes are out there where this 497-day counter has "rolled over", and if this figure is accurate given that limitation?
Apple would also gain Java as an Apple supported program and language. It would help better, faster Java come to Linux and OSX. Java could be more tightly integrated into Quicktime and thus into mobile phones where Apple is implementing it's latest builds of Quicktime.
Yell & scream & rant & rave... it's no use... you need a shaaaave ~ Bugs Bunny
I really didn't get that out of it. It was negative towards Sun, for sure, but I didn't see wired painting a doom and gloom scenario about it. They did mention that they had billions in cash at hand, which I read that as being highly unlikely to be any sort of takeover target.
And that is moving I/Os. I'm a programmer for a company that makes large scale storage arrays, and Solaris can beat most any other operating system (no matter what hardward) at moving just a bulk ammount of I/Os. We support Linux, AIX, Windows, Solaris, and many others, and Solaris is always a top performer on our tests.
Beyond that, I'm not sure what Sun machine's are good for.
Its not what it is, its something else.
The Baby Bells used their last mile monopoly to kill the "dot com" folks. The bandwith demands have grown, but not like they could have and they are concentrated in far fewer hands. This has made a glut of Sun equipment. A friend of mine bought and ultra spark, which once sold for $10,000, for less than the price of a high end home computer. There's no way he would have gotten his hands on a deal like that if a healthy and free internet market was working. Bad laws, such as those preventing me from buying California wine, and preventing me from running servers on my cable modem, have also played a part. Established interests are shining triumphant and we all suffer for it.
Microsoft has also harmed Sun's traditional scientific computing business. Microsoft has done much to blow up X compatibility and make communications with Unix difficult. One of the responses has been to move some of the calculations to M$/Intel platforms. This obviously does not work for all calcs, but consider the losses from 3D CAD and a perpetuation of the M$ as a client model. Linux can be a great aid there, so long as scientists and engineers revolt against the M$ Office lockin. They have only to realize that the pain is comming from one location and move on. Most have, but many are loath to move on yet. The support infrasturcture for free software is still growing.
In the end, Sun has much to offer. Their hardeare is first rate and they can embrace free software at any time.
DMCA, Hollings, Palladium. What might have sounded like paranoia is now common sense.
-- it's keeping them. I went to work for Sun my first job out of school. Our lab had largely Sun workstations and I really cut my unix teeth with Solaris, so the opportunity to go work for them was huge for me. When I went to work there, only 3 people in my group were over 30.
I went to interview at their brand new (at the time) Burlington, MA facility and I was simply amazed at the place; the facilities, the people and the atmosphere was so key in my going to work there (as was the salary they offered). I still think it's a great place to work, especially for people in my age group (I'm 25) who grew up getting used to flex time. That I could take a long lunch, play a few rounds of foosball and go to the gym at 4 in the afternoon made me a happy camper.
The problems began when I started sensing I ought to be moving up (or at least, around) in the company. I started in a position I liked but didn't want to stay in for more than a year or so, and as I started to make pushes to move around I was met with stiff resistance. Management claimed it was because of the economy, but I knew people who moved around and they weren't exactly examples of people who were going to save the company.
The key to this issue was that while Sun was publicly making overtures towards attracting the younger developers, the first and second level managers were only advertising positions for senior engineers and were being very inflexible in "stretching" the job prereq's for younger engineers. I often think the only reason I got a job in the first place was because I came in during one of the last "conscription"-type expansions the company did before the IT sector did its nosedive.
To this day they still have that problem; I often consider going back to Sun because the corporate culture is fast moving, fun and flexible, and I doubt I'll find that in any other company of that size. But the jobs and the people they're hiring now are all mid- or senior-level engineers.
So actually, now that I think about it, maybe it's more apt to say their problem isn't attracting young engineers -- the culture is almost geared towards them (why else would you put foosball tables and a Starbucks in your engineering centers?). The problem is that once they've attracted the young people, they have to get their managers to hire them.
B
"I'm payin' taxes, but what am I buyin'?" -- James Brown
http://www.sun.com/smi/Press/sunflash/2003-06/sunf lash.20030605.1.html
:rolleyes:
I called Sun not too long ago inquiring about hardware so my rack. I asked them a technical question about certain software running on it. The Sales guy said he would have a tech call me back. 2 months later after I purchased intel hardware and running Linux. A survey guy called from Sun and asked how things were going. I informed him Sun lost a $10,000 sale.
even a few min can cost ungodly amounts of money
I hate it when people say stuff like that. It makes me think of all the money the RIAA lost last year due to piracy. I think we'd all be better off not knowing how a 48x CDRW or a 400 Mhz processor can cost us ten times the profits.
Sun has one thing going for them, reputation.
SUN has two problems they must sort out.
1. Java - there are too many releases and java applications seem to be too tightly tied to specific point releases. This causes huge support problems.
2. SUN supplied h/w to many of the dot.com companies. When they went bust a LOT of 2nd hand hardware appeared on the market. It's difficult to compete with your own h/w at 2nd hand prices.
(Does anybody else find that typing in the form for SlashDot submissions causes Mozilla/Redhat8 to bounce the form around the window for no apparent reason?)
Sun does not, nor will it ever, understand users. The entire company is geared toward sysadmins who are expected to be able to jump through install hoops and tweak systems to get top performance.
It will kill them.
Sun is proud of their "9's" - 99.999...% uptime. The truth is that 99.99% of businesses don't need that percentage. Sun is proud of their scaling, but I've got news - 99.99% of businesses don't need that much power. It's also clear that we'll hit 10Ghz machines with multiple gigs of RAM in just a few years, and they'll cost around $1K; if you want terabytes of disk it may cost around $2K. How is Sun going to compete with that?
When any Luser can buy a machine that competes rather well with a Sun box, and can install any amount of easy software instead of Sun's pain in the ass stuff...
RIP Sun
<rant off>
That'd just be too ironic. NeXT and Sun worked together to push the OpenStep standard, but Sun pulled out to focus on Java. Apple bought NeXT and turned OPENSTEP into Rhapsody, which begat MacOS X after widespread developer revolt. In the transitition, there was a big push within Apple to convert everything over to Java. (Even the Objective C syntax!) That initiative failed (performance was brutally awful), and Java is still supported but little-used.
Sun and NeXT/Apple just haven't been able to get along...
Funny, but true.
Sun hardware is unbelievable expensive. I know the many Sun Blade 1000s we have at work are $16,000+ each. And while they are pretty, the performance is nowhere near worth the price tag... even with the massive discounts we get from Sun. And those "low-priced" Sun Blade 100s for the desktop are nothing to get excited about, either.
While Sun may be hurting, and its hardware its over-priced and under-performs, there is A LOT of Sun hardware w/Solaris out there in the field... and there will continue to be. We can only shake our heads and speculate why...
Exactly what does IBM get by buying out Sun?
* Fast RISC chips? IBM's already got them and they're incompatible.
* High Powered Unix? AIX is pretty good, and they've already set a path towards Linux, not Solaris
* Big development force? Well, there is this, though I'm sure IBM could rehire some of its folks or some Mad Skillz linux hackers that would have a more appropriate skill set.
* Tech to do massive SMP? Maybe, but they got mad Parallel processing skillz and bought Sequent out for their killer NUMA tech which scales better anyway (Shame they can't make DB2 run on it)
* 5.5 Billion in cash? Ok, this is a nice incentive :^)
What I see as far more likely is a cross OEM deal with Apple. AIX will port to Xserves and OS X will port to p640/660/680/etc real easy since they're all CHRP boxes (more or less).
Apple get proven high end servers, IBM gets quality low end "fill" servers/laptops/desktops to round out their AIX product line.
Sun will need to embrace a similar Solaris>>Linux strategy or AIX will suck market share as folks realize AIX is the best upgrade path off of Linux (less retraining and porting, more support for advanced availability, etc.)
You are in a maze of twisted little posts, all alike.
At the end of the article McNealy says the following about linux:
"Yet when talk turns to Linux, it's as if McNealy can't help himself: He knows he should be courting the world's Linux devotees, but instead he pokes fun at them. He points out that Red Hat, the leading purveyor of Linux systems, announced revenue of $24 million for its last quarter of 2002. I don't know where this multibillion-dollar Linux business is."
however, earlier in the article, when discussing SUNs past we get to read this:
"Back in the mid-1980s, when Sun was still a startup, it had neither reputation nor intellectual property, and it faced a murderer's row of competitors. One quarter it even needed to borrow $50 million to make payroll."
yeah well, i suppose a lot of people were laughing at sun at that time too figuring out where 'the money' was. I can't believe how ignorent SUN is towards Linux.
On a long enough timeline, the survival rate for everyone drops to zero.
This is where Sun is screwing itself over. Sun is a hardware company, and they make their money selling hardware, and support for that hardware. Their business model seems to be based on selling proprietary hardware which costs a lot, and is worth a lot. They need, among other things, scads of college students who are familiar with their OS and willing to recommend it over things like Linux and *BSD.
Sun is charging $20 for the x86 Cd for Solaris 9. This is the option for the poor college student, who definitely WON'T have a Sun workstation in his dorm room, unless it's an ancient sparcstation from the campus surplus (probably with Linux on it!). The problem is that for a poor college student, $20 for Solaris 9 is non-trivial. If you know you want to learn Solaris, you can easily do it. If you aren't sure, you surely aren't going to try it.
By trying to get a few bucks from selling CD's, they are completely ceding the ``recent college grads push our stuff'' effect to Linux and the *BSDs. Talk about being stupidly shortsighted! If a company puts in a Linux server, they can be reasonably sure that any computer-savvy recent graduate can assist a competent administrator. If they put in a Solaris server, not only do they pay more for the hardware, but they will have a much harder time finding cheap labor who are semi-capable of administering it.
See what I've been reading.
I don't think that's why he's saying. I'll use an example:
In my office there is an SGI system we bought ~6 year ago for doing visualization work. After about 1 year of using it, we needed more power. Did we go with a new SGI? No we went with a Linux box and an NVidia graphics card. The system was over 8 times as fast. It also cost 1/4 as much.
I think what the grandparent was says was that Sun should use Linux and make cheep, powerful x86 based workstations much like they are doing with servers.
The policy of the United States is worse than bad---it is insane. -- Ludwig von Mises, Economic Policy(1959)
Perhaps Slashdot should learn from Sun's mistakes.
There's a Mercedes gap too. I want one and can't afford one, but it's not government's job to do anything about it.
I would automatically question any comparison between Sun and any Intel based platform. However even Sun recognizes the future looks to be Linux on Intel.
We recently performed an internal cost/benefit analysis of Sun vs. Linux on Intel. Our study showed that while Intel platforms are very competetive, they fall behind on supportability. Intel machines require VGA port, BIOS, and keyboard/mouse ports. To provide remote OOB management you end up spending a fortune in cards and/or console managers, that Sun has built-in to their low end equipmnt. By low end I'm talking about a 1U $995 machine.
In fact we recently had a conference call with Sun about their Linux boxes... I told them that if they wanted us to buy Sun Intel Linux machines they would need to dispense with the VGA port and provide the same Light Out Management console port that their Sparc machines have. Which effectively means they need to build an OpenFirmware/OpenBoot machine with a RJ45 console port. Sun's rep stated that they are working on incorporating those technologies into the Intel platform.
So I think if Sun can deliver such a machine, in the sub $1000 category they will end up as the trendsetter for Intel based Linux boxes.
In 2002, everyone lost sales, except for Dell's dekstops. Sun lost fewer sales than IBM and HP and everyone else in the industry, so it picked up marketshare, a critical yardstick of competitive performance in down markets. So, while Sun is fuxxored, it is less fuxxored than IBM, HP, SGI, Unisys, etc.
The piece was a hatchetjob that displayed very little deep understanding of the IT market.
SoupIsGood Food
I find this discussion and article interesting as Sun has never been the high performance leader, even in their own machine class. When it comes to RISC, Sun's SPARC line and decendents has always been slower than the competition. DEC's alpha, IBM's PowerPC, HP's PA-RISC all were always ahead.
What Sun provided was a platform on which more software was available sooner than any other platform. Then, it became more software than any other platform except Microsoft. I am sure this is the origin of the pre-occupation with Microsoft. Yet, while Sun was regularly able to pummel its better performing competitors with its wider and earlier software availability, it just can not rival Microsoft in the breadth and timing of software available. Note that I am not refering just to the software produced by the system manufacturers. In fact, if that were the sole measure, then HP and IBM would have given Sun a much greater challenge. Sun's key to success was getting ISVs to use their platform as their native development platform, ensuring it was the first platform everyone released on. All the others were ports, and thus were released months later. This was a huge edge for Sun that was terribly difficult for competitors to remedy. Simply building faster, "better" hardware would not lure ISVs to shift their development platform to another hardware vendor's product.
But, Micrsoft is far ahead of Sun in exactly those things that allowed Sun to beat its competition. I don't see Sun ever being able to succeed using that strategy, and they sure don't seem to be interested in any other. Though, with the other RISC platforms dropping like flies, being replaced by Itanium with all of its performance and acceptance problems, and sudden Sun's hardware looks like it may become king of that hill. Of course, no one is paying for that class of hardware any longer... if they do, they now go buy IBM's tREX and run piles of virtual Linux machines on it.