Home > Oracle and SUN > Humble Pie for Sun Humble Pie for Sun Eugenia Loli 2004-09-21 Oracle and SUN 32 Comments Jonathan Schwartz, Sun’s president, says the company has mended its ways since the days when “we didn’t listen” to customers. Can the Silicon Valley luminary brighten up its prospects? About The Author Eugenia Loli Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker. Follow me on Twitter @EugeniaLoli 32 Comments 2004-09-21 11:09 pm Sun fortunes SHOULD start look much better. Sun has got by far the most exciting and compelling product line out of all top tier vendors bar none. People out there just need to shake off the stale misperceptions about Sun in general. I’m quite bullish on Sun. 2004-09-22 12:08 am I love it. This phrase should become a Standard English fixed expression. In stead of “failure keeps me humble,” we should say “failure keeps on a-servin’ me that good old humble pie.” Please don’t mod this down. 2004-09-22 12:31 am Sun posted profit last quarter, they haven’t done that for along time. I told you that if they posted profit they’d survive and if they didn’t they’d die. they survived 2004-09-22 1:22 am Wow! What a novel idea! But seriously, big companies will never listen to customers, just continue on forcing their same old lame product down people’s throats. For example, where’s the boom box I want, the one that’ll play ogg files? 2004-09-22 1:24 am I still don’t see how SUN can turn this around. They have been getting their butt kicked for far too long. Providing Opteron servers is a good thing but they shouldn’t be the flagship product or anywhere near it. New SPARC processors or a decent proprietary replacement is in order. 2004-09-22 3:32 am > New SPARC processors or a decent proprietary replacement is in order Lagging behind the news, aren’t you? The long term replacement for the current SPARC processor line will be coming from Fujitsu SPARC64 in just a couple of years (APL), which happens to be a very well performing processor that is quite capable of kicking ass of any processor architecture out there including Power and Itanic. Even with the current state of affairs there is not a whole lot for Sun to worry about considering the fact Itanic turned out to be a complete dud and almost each release of Power requires a major forklift upgrade and/or recompilation of the software. UltraSparc has *the* best high end 64-bit processor story out there without any exceptions. UltraSparc can more than keep up with the competion in the near term by just increasing the clock speed and doubling up the number of cores in the processor. The biggest problem UltraSparc has got is a bunch of ignorant nimrods that judge performance of the system by the number of clock cycles it can churn out. 2004-09-22 4:37 am when the UltraSPARC IV+ comes out (should be within the year or close after) it will (suposidly) double the speed of the UltraSPARC IV and be 90nm (this is what i have heard if i am wrong please dont flame me ) 2004-09-22 6:11 am and Schwartz is an idiot. He is still trying to peddle expensive hardware and a proprietary OS, but what the market wants is commodity hardware and a vendor-independent OS, and good services on top of that. As for their obsession with SPARC, it’s becoming almost a kind of fetichism… SPARC is a dead-end, not because of the architecture, but simply because it’ s too expensive to develop further. 2004-09-22 6:45 am > and Schwartz is an idiot. He is still trying to peddle expensive hardware and a proprietary OS, but what the market wants is commodity hardware and a vendor-independent OS Schwartz is not an idiot, far from it. What consumers want is cheap hardware running cheap OS and this is exactly what Sun is going to give to them. Solaris will soon be opensourced and is already vendor-independent as far as the platform is concerned. Sun is doing just the right thing at the right time — offering the premier platform to the customers at the unbeatable prices (even by Linux). > SPARC is a dead-end, not because of the architecture, but simply because it’ s too expensive to develop further. Sun with Sparc has been following an incremental development cycle as opposed to Intel trying to create a revolution with Itanic and therefore doesn’t suffer from rather grotesque development costs. Furthermore in a couple years time Sun will be using Fujitsu for design and fabbing, which should considerably reduce the costs for Sun and improve the economies of scale for the entire SPARC ecosystem. If anything SPARC should have a long and prosperous life. UltraSparc chips are actually the cheapest of the bunch in the high end space compared to Itanium, Power, and PA-RISC. Itanium powered gear for example is almost twice the price of similarly configured UltraSparc gear, this is one of the reasons why Itanium is having such a hard time gaining traction. 2004-09-22 6:59 am Itanium powered gear for example is almost twice the price of similarly configured UltraSparc gear, this is one of the reasons why Itanium is having such a hard time gaining traction. Not only that, but if I want to sell a SPARC workstation, the cost of entry to the market is alot lower than for Itanium machines. I can ring up SUN, organise MB and processors etc. With Intel, as a small vendor, I was told to piss off by the Intel sales person. On one hand they piss and moan becuase they have insufficient vendor support, and yet on the other hand, they tell OEM’s to go screw themselves if they ring up and inquire about how one can obtain the necessary parts to assemble their own machines. 2004-09-22 8:20 am I disagree. If they really have cooled down to focus on their technology, I believe they are headed in the right direction. I would love nothing more than to watch them kick ass and create some of the best technology we’ve ever seen. Their tech may or may not be the best, but competition is a very very good thing. And they are the best competition the tech industry has. I just wanna good fair fight. Who can create the best tech and bring it to market without throwing tantrums and calling your opponents names and acting inappropriately for such a professional organization. I think its more important to create a better quality of life for your employees, but what do I know. 2004-09-22 9:42 am OK, The APL is really a stop-gap until Sun’s next gen CMT processors (Rock, N1, N2) come on line. This is based on the facts that conventional processors are getting bigger, hotter and less efficient (sucking more power). This is why USV was canned. USIV+ is much more than a clocked USIV and it should see more than a doubling of performance on real workloads. 2004-09-22 10:48 am I don’t understand Sun Microsystems, probably nobody does. 2004-09-22 11:29 am How do you make all the wrong choices like Sun did, I don’t even know how that is possible. Price cutting is a temporary thing that monopolies do in order to retain control, they raise the prices back up after the competition is drained. Sun is not an empire like Microsoft, it can’t do what they do. Sun should become a Linux vendor, it should take it’s piece of the pie and work for long term recognition in a decentralized market, look to the new paradigm, but instead Sun is fighting an up hill battle against everyone including itself, and now it’s a Microsoft customer. I say, give Microsoft back their money and lay claim to your share of the pie in the expanding Linux market. 2004-09-22 11:56 am For many years, Dell has been beating Sun in the low end server market. But it hasn’t been Dell beating Sun so much as it’s been Intel beating SPARC. Now, with the collapse of Intel’s 64 bit strategy, it’s Dell suffering from lockin to processors with poor price/performance. Sun has jumped on the best commodity processor on the market, the Opteron. The tables have turned. 2004-09-22 12:18 pm Sun’s attack on Red Hat over TCO isn’t likely to work. But hidden under the smoke and mirrors is an interesting admission: Sun knows that Red Hat has won. Solaris/SPARC was the preferred platform for applications in certain markets for years. Sun let their workstation share slip away during the dotcom boom, as they focused on servers. Red Hat is now the preferred platform for many of those applications. Solaris on Opteron has a Linux compatibility layer. Sun knows that ISVs won’t target Solaris x86. But they will target Red Hat. So commercial Linux applications will run on Solaris. But will the ISVs support it? 2004-09-22 12:41 pm Solaris on Opteron has a Linux compatibility layer. Sun knows that ISVs won’t target Solaris x86. But they will target Red Hat. So commercial Linux applications will run on Solaris. But will the ISVs support it? Well, what SUN needs to do is chuck Solaris with the system for free, and pay for the porting of workstation class applications to Solaris. Approach the Adobes and Macromedia’s, ask how much it will cost, and cut these companies a cheque to cover the cost of porting their applications over. Btw, the porting effort shouldn’t be hard considering that there are bridging tools like Mainsoft’s MFC implementation on UNIX along with a number of other win32 features. 2004-09-22 12:46 pm > Sun’s attack on Red Hat over TCO isn’t likely to work. But hidden under the smoke and mirrors is an interesting admission: Sun knows that Red Hat has won. What kind of dream world are you living in? I would not put RedHat even close to the same category with Sun. RedHat is a comparatively low tech company with close to zero R&D and pretty weak execution. It just so happened that RedHat managed to ride the Linux hype wave and end up the top of pile, it has little in terms of real technology contribution tcompared to Sun or IBM. If you’re pinning your hopes for the future computing on RedHat, you will be doomed. The biggest asset RH has at the moment is hype, once that is gone there won’t be much left. Sun should most definitely prevail over RH, it’s got an order of an magnitude better product (Solaris) which is better performing, more feature rich and cheaper at the same time. Again once the hype subsides, Sun is the one who’s got the real story going. 2004-09-22 2:54 pm “What consumers want is cheap hardware running cheap OS and this is exactly what Sun is going to give to them.” Not entirely true, comsuners want end to end solutions and customer service that supports the investment in equipment. Something SUN has failed miserably with in the past. “Itanium powered gear for example is almost twice the price of similarly configured UltraSparc gear, this is one of the reasons why Itanium is having such a hard time gaining traction.” Probably but business owners can run VMS,HP-UX,Linux and Windows on Itanium2 all on a single server http://www.hp.com. 2004-09-22 3:40 pm > Not entirely true, comsuners want end to end solutions and customer service that supports the investment in equipment. Something SUN has failed miserably with in the past. I could have been sort of true a few years ago, but nowadays Sun has the most complete end-to-end product line out of all vendors out there. Sun has got a pretty cohesive product line all the way from storage and servers to middleware and services, exactly where HP is failing so miserably right now and Sun will be eating HP’s lunch pretty soon. > Probably but business owners can run VMS,HP-UX,Linux and Windows on Itanium2 all on a single server http://www.hp.com. VMS and HP-UX are both dead meat and were pretty handily dead ended by HP with their rather goofy migration plan to Itanic. According to HP’s own stats about 50% of HP’s customers won’t be migrating to Itanium choosing a different processor architecture (SPARC or Power most likely). Linux on the other hand is not gaining much if any traction on Itanic, since it violates the very basic value proposition of Linux — cheap OS on cheap hardware — and Itanic makes the hardware look pretty darn expensive. Bottom line is that Linux on Itanic is not making much sense with the exception of SGI tailoring to a very specialized market with very specialized needs. The same story applies to Windows, it will takes years before Windows will be entrusted with running serious 64-bit applications — Windows is not a serious enough OS to run serious applications. In the end Itanic is left to look pretty miserable compared to more mature competition (SPARC and Power). 2004-09-22 4:52 pm What kind of dream world are you living in? I would not put RedHat even close to the same category with Sun. RedHat is a comparatively low tech company with close to zero R&D and pretty weak execution. It just so happened that RedHat managed to ride the Linux hype wave and end up the top of pile, it has little in terms of real technology contribution tcompared to Sun or IBM. If you’re pinning your hopes for the future computing on RedHat, you will be doomed. The biggest asset RH has at the moment is hype, once that is gone there won’t be much left. Red Hat, along with all other Linux distros, leverages the development efforts of hundreds of thousands of developers and companies world wide hacking away on the Linux kernel, the GNU userland utilites, Gnome and KDE, OpeonOffice, and many other open source projects. A single proprietary company, not Sun and not even Microsoft, can even come close to this kind of developemnt resource. Sun is certainly one of the biggest R&D companies out there, and they have some of the highest tech available, but even their substantial efforts can not hope to compete with the collective development efforts of the entire Linux/Open Source community. And along with the huge Linux/Open Source community’s development efforts comes huge mind share, which is extremely critical for any IT vendor. True, Red Hat itself only has a tiny fraction of the development and R&D resources hat Sun has, and if you were comparing only that, Sun kicks Red Hat’s butt. But when you are talking about a Linux distribution, you are talking about value added products and services on top of the monster development of Linux by the Open Source community, and the monster mind share that goes with it. When you look at it this way, Red Hat, along with SuSE/Novell, Mandrake, et al, all have a huge advantage over any proprietary OS, including Solaris. 2004-09-22 5:37 pm Redhat does not own the factors of production. 2004-09-22 5:48 pm “What kind of dream world are you living in?” I’m living in the dream world that has companies like Cadence making Red Hat a supported platform. Electronic Design Automation helped create the Unix workstation market. Sun started with workstations, out performed the competition, and grew until they could enter the server market. But they took their eyes off of the ball. SPARC compares quite poorly for price/performance on workstations. Most engineering applications (electronics and mechanical) now run on Linux/x86, Red Hat in particular. Solaris isn’t the problem, hardware has been the problem. Many ISVs and customers switched to Windows searching for better price/performance, and were pleased with the hardware, but not the OS. Now they are switching to Linux, with better results. I’m not claiming that Linux is better than Solaris, or that Red Hat is better than Sun. I’m pointing out a fact that you may not know, which is that for engineering applications on x86 hardware, Red Hat is the standard, just as Solaris is for non-x86 hardware. Sun knows this. ISVs don’t support Solaris-x86, and they won’t need to. If they support Red Hat, then it will run on Solaris-x86, too. Sun can’t survive with high end servers alone. They need to recapture more of the low end server market, and get back to their workstation roots. Opteron lets them do that, and I think they have a decent chance of doing well. But the low end market has become a Linux market, so Solaris needs to support running applications written for Linux. Sun may beat Red Hat for some support business, and customers may be running Solaris, but the ISVs will be selling Red Hat binaries. Market share becomes even harder to measure. I like the idea of taking applications targetted to Red Hat and running them in Solaris/Opteron partitions. Red Hat won the battle for ISV mindshare on x86, as I said before. Sun may win the support business. Those are now two different issues. I’m encouraged that someone in Sun management sees that. I’m dissappointed that Sun advocates don’t see it. Linux has been taking market share from Unix vendors for two reasons. One is lower hardware costs. Sun has addressed that with Opteron. The other reason is less lockin. Switching between proprietary platforms is costly. Switching between Unix platforms lowered that cost, but the vendors prevented convergence on to a true single platform. Linux takes us much closer to vendor independence; still not there, but closer. When vendors talk about unique features of their products, I end up thinking about how much it will cost if and when I need to switch away from those products. I don’t want applications dependent on unique features. Unique features that lower operating costs without changing applications are great. Linux as a vendor independent platform for applications is very attractive. Solaris as a better platform for running Linux applications works for me. Solaris by itself is less attractive, no matter how good the technical features are. I can move Cadence or Mentor Graphics applications built for Red Hat x86 to Solaris if it makes sense, and away in the future, should Sun decide once more that it isn’t serious about x86. 2004-09-22 5:49 pm Sun is too much of a flip flop on Linux, that’s why it isn’t more successful. Red hat does not control Linux. What a bunch of nonsense. 2004-09-22 7:38 pm To me, Sun’s biggest problem the past few years has been the simple fact that they have felt rudderless to the end user. Buggers flip flop more than Kerry. Oh, we’re gonna make Solaris free to x86 users… No, we’re not going to continue making an x86 version… Ok, we’ll keep making it, but you have to pay… But we’ll release one version back for free… All the while, their proprietary line has felt… stagnant. Newer models and OS just don’t seem to offer anything that would make someone who’s runnig a Solaris 6 era sparcserver run right out for the newest latest and greatest UltraSparc. Now, they’ve finally begun settle on a fixed path… They HAVEN’T flip-flopped on their plan for almost a year. Their x86 based ‘entry level’ Sun Fire z and x models (which even have a blade version) fit the bill for most people’s needs, and at under three grand are worth the money considering the HEAVY support Sun throws in behind it. It has taken them long enough to realize it, but the fact is the x86 based processors are an economic way to high horsepower, and on the whole the majority of die-hard techies are most comfortable on said platform. Not supporting it or worse, changing the attitude towards it repeatedly in favor of their own proprietary solution was simply alienating a hefty portion of the potential client base. If anything, it can be a stepping stone to graduate people to the more mature, stable and robust UltraSparc based solutions. 2004-09-22 9:17 pm Apple leads the field in industrial design and general appeal for the workstation market. I have a number of older Sun and SGI boxes sitting next to me right now, and looking at them I can recall a time when it was Sun and SGI who led with regard to innovative industrial design, key performance metrics (e.g. 3D) and reliability for computer hardware. People actually wanted these computers – like really lusted after them. I look at Sun’s current offerings and there is simply nothing thats affordable (e.g. under $5000) that you can’t get cheaper elsewhere. The SGI O2 did amazing things with ease that were a total pain in the ass to get done on other machines from the same era – I just don’t see anything like that from Sun. For an company that invests so much in R&D they have precious little that really differentiates them from anyone else below their flasgship 32+ CPU machines. And SGI, after trying unsuccessfully to peddle beige-box x86 internals in a tarted up ATX case quickly realised that nobody at all was impressed by this, and Sun is about to follow them down the same path with their rebadged clone Opterons. Hell I’m just waiting for the day that Schwartz follows Belluzo to go work for Microsoft. 2004-09-22 11:19 pm > And SGI, after trying unsuccessfully to peddle beige-box x86 internals in a tarted up ATX case quickly realised that nobody at all was impressed by this, and Sun is about to follow them down the same path with their rebadged clone Opterons. Drawing a parallel between SGI and Sun here does not really make a whole lot of sense. SGI problem stemmed from the fact that their they didn’t own the OS for the x86 gear and couldn’t leverage it appropriately and more importantly SGI couldn’t deliver the product at low enough price to compete with the low-brow R&D poor whitebox churners like Dell. Sun is not guilty on either of these counts having Solaris and being able to leverage the OS to its own favour. And more importantly Sun’s x86 gear is now the CHEAPEST on the market, even Dell comes out more expensive than Sun! Sun is doing all the right things and to prove it Sun has reported the highest unit shipment increase in the industry beating the likes of Dell, IBM and HP. 2004-09-23 4:17 am Sun and IBM need to merge into one company. The problem with Sun is that it can’t see where to fit in, well, I just looked at this situation just now, and this IS the answer. What also should happen is that a portion (a department) of the workforce of Sun and ideally SCO will become a new Linux vendor. They will enter the market as a competitor to Redhat and Novell. 2004-09-24 2:44 am “Apple leads the field in industrial design and general appeal for the workstation market. ” No, they don’t. That would be Dell’s Precision line (running Linux or Windows), not the PowerMac. 2004-09-24 4:52 am Hello? Its a Dell. commodity everything, no attention to detail, doesnt perform significantly better than a machine of the same spec i could put together in my spare room for less mmoney. The Dell ‘Precision Workstation’ is a budget dell desktop with a 3rd party graphics card, and doesnt even have the level of R&D that SGI put into their x86 workstations to qualify it as anything but a budget desktop with a fancy graphics card. Maybe a piece of shit PC with a ‘Precision Workstation’ sticker on it is what you consider a workstation, but i certainly dont class it as one. 2004-09-24 5:06 am “Hello? Its a Dell. commodity everything, no attention to detail, doesnt perform significantly better than a machine of the same spec i could put together in my spare room for less mmoney. ” You’ve never used one, have you? They are quiet, reliable, affordable and fast workstations. Out of that list, only “reliable” applies to my Octane. “The Dell ‘Precision Workstation’ is a budget dell desktop with a 3rd party graphics card, and doesnt even have the level of R&D that SGI put into their x86 workstations to qualify it as anything but a budget desktop with a fancy graphics card. ” The claim I questioned was the “Apple is the best workstation” one. Pay attention. What does SGI have to do with this? BTW, the later SGI x86 workstations were stock ATX hardware, unlike the earily 320 and 540 (the later models had *3rd party graphics cards* from Nvidia). Dell’s hardware more certainly does have a significant R&D investment behind it – just not all funded by Dell. “Maybe a piece of shit PC with a ‘Precision Workstation’ sticker on it is what you consider a workstation, but i certainly dont class it as one.” You very likely are not the target market for technical workstations. 2004-09-24 7:32 pm “And more importantly Sun’s x86 gear is now the CHEAPEST on the market, even Dell comes out more expensive than Sun!” After seeing this quote and others similar I decided to price a Sun server like the Dell that I just bought. Dual 2.8 Xeon 4 146GB drives, RAID 10 1GB RAM 24×7 support Rack mount rails Standard cd/floppy No OS Dell 2650: $4,721.00 Sun V65X: $8,872.00 Am I missing something? The price difference is close to double. Is there a more comparable model than the V65X?