Microsoft’s server and tools chief Bob Muglia has chided Oracle for peddling a return to ‘1960s computing’, accusing its rival of going against industry trends and backing a dying and expensive operating-system architecture. “There are some things that Oracle is doing that I just shake my head at,” Muglia told financial analysts attending the Goldman Sachs Technology and Internet Conference in San Francisco, California, “I don’t understand what’s going to happen – what they think they’re going to do with Sparc. I don’t see how Sparc can live long-term.”
…says a guy from a company, which is going Itanic. Yeah, right…
I was thinking exactly that!
I thought: “Ummm… with all the news on Microsoft’s issues of late, isn’t it better to look at your own company first?”
Either way, I hope he’s saying it based on a “watch and learn” project where they are to analyze the mistakes of other companies and make sure Microsoft doesn’t head the same direction. But I’m pessimistic…
and they’re entitled to it. But as the old saying goes,
just like assholes, everybody’s got one.
But I liked his remark about not saving money with VDI – and he’s right, if you’re using Windows, it won’t be easy.
But with virtual desktops, an organization can package and deliver just the kind of desktop that’s needed for the job – and it won’t always have to be a full-blown MS desktop.
It can be anything, including a hand-rolled, open-source one and with VDI and a fast network, you can switch from one to the other in minutes.
A nimble company could save on a lot of desktop licensing – I think ( well, that’s my opinion )
I just saw this in the News section:
http://windowsteamblog.com/blogs/windowsexperience/archive/2010/02/…
So, after lambasting Oracle about taking us back to the ’60s, MS$ decides to launch their version of “That ’70s Show”.
I guess Bobby Muglia didn’t get the memo
They’re just ticked because it doesn’t support Windows. Boo hoo…
sure the name is old but solaris is wayyyyyy advanced and still has a lot of tech, good tech, the MS doesnt have in any OS.
I get that this word he keeps using is reminiscent of the old days (like I was around then) but they could always improve hardware support for linux on sparc. If they could just make cheaper sparcs they might actually be able to sell them in desktop systems (like sunrays) and If their power scales down as well as it does up, they could do some damage to the thin client industry. They also have more advanced chip virtualization, they could make more robust virtualization stations if they would aid other operating systems in the support of their hardware. there is a lot not being done with SPARC that’s being done with POWER, embedded. putting Open Solaris on routers would make my head explode with all the possibilities. Think Crossbow on a router.
Not sure if it’s true they have more advanced “hardware” virtualization, but with AMD-V and Intel’s VTx, but unless it’s coupled with capable software, what’s the point? I don’t see Sun haven’t the equivalent of VMWare’s suite of products with xVM or whatever they are branding it these days. With the use of cheap commodity hardware and a simple VM infrastructure w/VMotion is hard to beat.
Sun still has a few niches such as really high I/O for Niagara and cheap commodity storage with the “thumper” and “thor” boxes, but aside from that, it’s not really the preferred platform for clouds.
I’m not going to invest in a platform which doesn’t have a cost/performance ratio that makes sense, doesn’t have the robust virtualization that VMWare provides. Nor would I invest in a platform that outperforms only in certain “niche” areas. I want a platform that performs well and consistently.
I’m all for Oracle-bashing, but come on. Pot, meet kettle. Somebody from MS trying to act like they don’t like vertical integration shows a severe degree of hypocrisy.
More traditional SPARC hardware is dead. But, Niagara has a good niche, that isn’t yet profitable for Intel or AMD to try to take over; and that, with continued quality support for Solaris[, and Linux], could provide some real value for stressful database, web, and general networking workloads.
Edited 2010-02-27 02:52 UTC
Guy from tech company does not have any thing nice to say about the products from a rival company. News at 11.
I have to agree with you on that point.
I worked in a shop that was a big Sun SPARC based shop.. it just didn’t cut it any longer. The one big exception was the Niagara based processors systems.. To give you an idea, we had a very intensive I/O application (Netbackup) that we had T2000 servers running at 9.2Gbps sustained for the entire backup window. CPU load was about 50% so some bright sparc decided it would be a good idea to buy some XEON based servers because they could get 4 for the same price.. the reality was that they could only do between 4 and 5Gbps tops before they would flat line despite the CPU being only 40% idle.. the workload handling and I/O on the Sun Boxes is amazing… In the end they went back to T2000 based servers running Solaris… no more xeon’s running windows doing that I/O workload.
That’s just bizarre. I Understand that Sparcs are capable of handling massive IO better than xeons, but who goes from Solaris on Sparc to windows on Xeons? THose are such completely different platforms. It would be a better comparison to go Solaris on xeons, less disruptive as well.
Standard practice over the past ten years has been to move to Linux on Intel or AMD. That’s why Sun is now no more. They never saw that coming and never accepted and came up with a solution when it did. Irconically, Oracle were a part of that. Quite what Oracle will do to reverse that state of affairs after Sun’s efforts to get the tide to go back I have no idea, so Microsoft are maybe right to have a dig at their expense. No surprise there.
The only option is to get the performance of the SPARC architecure somewhere on a par with current x86 systems and maybe push it into a high-end and lucrative niche where Power lives currently. However, the R and D and huge costs in doing that are so scaringly high as to be infeasible, and it’s impossible to see Ellison pumping tens of billions in for a completely uncertain return with no fixed timescale.
I mourn the passing of MIPS, Alpha and possibly SPARC because I just don’t think a ‘one main architecure’ landscape is good for anyone, but it’s impossible to see a way out of this.
Edited 2010-02-28 18:45 UTC
Ok, Linux is a bit more understandable. Its reasonably Unixy, but windows? Really?
I agree. Sun should come to the same realization that Apple came to switch entirely to the x86 platform. SPARC is dying. Even if xVM came up to par with VMWare, it’s easier to find VMWare knowledge than Sun.
As for the Niagara performing faster, that might have been true 1-2 years ago. But take the most recent Nehalem platforms and I’m sure you’ll see it outperform the Niagara.
That’s the beauty of the x86 platform, it performs consistently for all applications IMO. I’d rather have a consistently performer than one that performs on two extremes of the spectrum, depending on the workload.
Talk about hypocrisy – Microsoft is everything Oracle is minus the hardware division; so why isn’t Microsoft considered vertically integrated? they’ve got HP, Dell and Lenovo by the short and curlies; they might as well lay claim onto these companies and state that they own them by default.
Personally I don’t care nor do I fear vertically integrated companies; Sun SPARC use industry standard memory, PCI cards, firmware (standardised by the IEEE 1275) and you can plonk Linux on a Niagra or SPARC64 (Fujitsu CPU which is being used now) or OpenBSD or NetBSD or Windows if Microsoft chose to port it. To call SPARC proprietary tells me that this so-called ‘analyst’ is so far out of his field of expertise he would be best to shut his trap before he makes an even bigger fool of himself.
As for what I think of Microsoft? I just hope that he is but one of many people in the organisation with those who do have the smarts, after reading what the guy wrote, universally shuddered at such blatant stupidity being put on display.
HP, Dell and Lenovo are all independent companies that sell both Linux and Windows servers.
It’s in reference to how it is non-standard compared to x64.
As much I see Sparc being praised around the net it doesn’t reflect healthy sales. I don’t doubt that Oracle can make Sparc boxes more appealing but let’s not forget that Sparce server sales have been in major decline the last few years.
That said the heavy R&D that Oracle plans to invest in Sparc will make things interesting.
Then “proprietary” is the wrong word. When talking of a proprietary technology, it means closed or otherwise locked down to one vendor or entity. E.g. Microsoft Windows is a proprietary operating system. ActiveX is a proprietary technology. Sparc doesn’t really qualify, in many ways it’s more freely open than X86. “Nonstandard” might have been a better word where today’s environments are concerned (X86 being the de facto standard).
Then again, given the rest of what this guy says I’m not surprised he can’t even be bothered to pick up a dictionary before publishing an article.
And you can purchase a x86 machine form Oracle with OpenSolaris/Solaris or Linux pre-installed or you can go off and purchase a copy of Windows and they’ll provide you with the required drivers. The hardware is already Windows tested – you’ve just got to install Windows. Considering that the vast majority of enterprise customers are on large software subscription plans form Microsoft – they don’t rely on the OEM to provide them with the software thus your point is moot.
As for independent, the fact you couldn’t get the gist of what I was saying pretty much tells me you have absolutely no skills as so far as understanding English.
Non-standard? how do *YOU* define a standard? x86 isn’t a standard, it is a bastardised proprietary piece of technology whose secrets are known because of some heavy duty legal work and reverse engineering. SPARC on the other hand is a standard:
http://www.sparc.org/
Just because there are only two or three SPARC CPU vendors in the world doesn’t make it any less of a standard.
SPARC is in the situation today because of stupid decisions made in the past by Zander (the destroyer of companies – Motorola being the best example of his incompetency) and in the future as a result will never get the volume to make up for the increasing complexity of CPU design; Intel can keep up because their volume will always make up for the increasingly complexity and cost of designing a CPU. Niagara on the other hand is an attempt to carve out a niche through a simplified CPU design, and the SPARC64 from Fujitsu being a concession that UltraSPARC is dead in the water. The question remains whether Fujitsu sells off their SPARC division to Oracle given that right now Fujitsu is straddling the fence between SPARC on one hand and Itanium/x64 on the other
Edited 2010-03-02 01:04 UTC
X64 is not a standard. I guess what he really mean is that x64 is more common than Sparc. “Common” just doesn’t have the same degrading value as “standard” though.
I think that might be the angle. It’s proprietary because you can’t run Windows on it.
Really, Microsoft’s overall strategy is no different than Oracle’s. Oracle want developers to write applications with Oracle’s tools and database, running on a platform, operating system, and hardware all produced, sold, and supported by Oracle. They even offer consulting services, and pre-built applications running on their platform.
Microsoft do exactly the same thing, minus the hardware.
The only real difference is that Microsoft had enough clout to just invent their own platform (.Net and Windows), while Oracle had to go with an existing platform (Java and Unix).
No he’s talking about how building a business model around Sparc boxes is risky since you can’t utilize commodity x64 hardware. It has nothing to do with Windows. He’s talking about non-x64 hardware.
Linux on x64 has always been more of threat to Oracle’s business plan than Windows. Oracle has stated in the past that their main competitor is IBM.
No Oracle wants to sell its database software on Oracle branded boxes. That’s the core of their business plan and it is a much different strategy than the one Microsoft uses. Microsoft wants to sell a range of software and let third parties handle the hardware.
Which has nothing to do with standard and everything to do with ubiquity. You stated that x86 hardware and components were more common and thus more readily available then you’d have to argument but the fact you’re throwing around the word ‘standard’ without the slightest understanding as to what ‘standard’ actually entails shows an epic level of ignorance on your part.
Of course, but why is that? again, you throw out a statement and never expand as to why it is the case.
How the hell can you stay with a straight face that there is a ‘massive difference’ considering that there is only one component missing?! do you really think there is a huge difference between that and the close relationships Microsoft has with OEM’s? that just because they happen to be different organisations that their business plan is so considerably different?
Can anyone say vertical integration? Microsoft’s version of it, of course…………