back to article IBM reneges on Solaris GPFS promise

IBM has a love, hate, love, hate, hate, hate relationship with Solaris, Sun Microsystems' Unix variant. And becoming an OEM partner of Sun's to distribute Solaris has apparently changed none of that. Back in August 2007, IBM and Sun inked an OEM deal that would see IBM distribute Solaris 10 on selected models of the company's …

COMMENTS

This topic is closed for new posts.

This post has been deleted by a moderator

Anonymous Coward

'A decommitment of intent'

That's fantastic. You could use it when you're trying to weasel your way out of a relationship:

[scene: girlfriend's house, with girlfriend reading email off laptop]

"Great time last night, dadadatada, busy next week... datada... ... decommitment of intent... datada dada... DECOMMITMENT OF INTENT?!! HE'S BREAKING UP WITH ME!!!!"

0
0
Silver badge

Lack of intent or lack of market appetite?

You'd think all those old instances of Slowaris chugging away on the existing installed SPARC base would be a tempting target for IBM. Which implies IBM isn't having that hard a time enticing those customers over to AIX on Power or Linux on xSeries. Otherwise they'd be doing all they could to make the IBM hardware as attractive to them, like getting the Slowaris version of GPFS out the door ASAP.

0
1
Anonymous Coward

$ and Control

IBM is all about money and control. There may not be enough instances of Solaris running on IBM hardware yet for the bean counters to green-light the GPFS file system. It could also be that their exaggerated hopes for moving customers from Solaris to AIX is not going all that well and they see no reason to spend anymore money on things like GPFS

0
0
Thumb Down

Matt Bryant

Matt Bryant, don't you have HP-UX boxes to sell instead of pouring dirt on the competition? Very uncool I have to say. And regarding your "Slowaris" parroting - 1990's called, they want their moniker back. Wake up man, Solaris is now way ahead of AIX, HP-UX and even Linux on both pure performance and price/performance. Constantly repeating one dumb line is not going to make it true.

0
0
Pirate

Of course there's no announced support for HPUX

HPUX, so there must really be no demand for HPUX, using Matt the HP Sales Grunts reasoning... IBM must be having a really easing time convincing all those HPUX users to move over even without ever announcing even the intent to port GPFS... What a meat-head.

0
0

@Matt Bryant

"You'd think all those old instances of Slowaris chugging away on the existing installed SPARC base would be a tempting target for IBM. Which implies IBM isn't having that hard a time enticing those customers over to AIX on Power or Linux on xSeries."

Matt,

Your logic is really warped. The fact that IBM won't do the work on GPFS doesn't imply they aren't having any trouble enticing customers onto AIX. There could be a whole host of reasons they aren't doing it. Perhaps the install base is too loyal to Sun for it to be worth their while. Perhaps the SPARC install base is happy 'chugging away' and doesn't want to move. Perhaps they like the guaranteed binary compatibility of Solaris. Perhaps it is a harder job than IBM thought and they can't be arsed. Perhaps they see that ZFS is the way forward. Perhaps, perhaps, perhaps.....

Matt, I know that you have a Sun hate thing going - and that is fine - but at least hate them with some logic!

0
0
-
Boffin

small correction

IBM LS41 is a four socket opteron blade. And IBM x3852M2 is most likely IBM x3850 M2.

0
0
Silver badge
Pirate

Dear Sunshiners....

"Matt Bryant, don't you have HP-UX boxes to sell instead of pouring dirt on the competition?..." I don't sell boxes, but I do enjoy ribbing you SUnshiners! What an easy target!

"....Solaris is now way ahead of AIX, HP-UX and even Linux on both pure performance and price/performance...." Now if only the customers believed that! Ignoring the obvious stupity of your statement, a review of market figures will soon show you that even giving Slowaris away free has had negligable effect on market share, and the ongoing decline in enterprise Slowaris still in a death-dive. Now, let's just assume for a moment not all those people buying AIX, hp-ux, Windows and Linux can't all be stupid (I know you SUnshiners like to think they all are), which leaves the only conclusion as they prefer AIX, hp-ux, Windows and Linux to Slowaris. Hint - this is one reason Sun's market cap os so low, it's because the analyst think the Sun product set has the characteristics of smelly brown stuff and is not likely to make any money....

RE: Bill - "Of course there's no announced support for HPUX...." Well, apart from the fact HP has a very long and established relationships with Veritas (the reason the AdvFS got dumped) and Oracle (number one Oracle RAC and grid partner) and so doesn't need GPFS, HP also have the PolyServe people and could probably produce a better UNIX grid product if need be.

RE: Andy White - "....Perhaps the install base is too loyal to Sun for it to be worth their while...." Really? So you haven't seen the marketting effort hp and IBM are putting into snagging disgruntled Sun customers? Or the desperation with which Sun switched from bashing SPARC64 to praising it when Rock fell off the roadmap? Sun have a big problem and it's that old SPARC customers, even though many are very loyal (a lot of them are Sunshiners), are being forced to move by the simple expedient of CIOs turning round and saying "No, we can't wait for Sun anymore, just buy hp/IBM/whatever!"

".... Perhaps the SPARC install base is happy 'chugging away' and doesn't want to move...." Well then you have just sounded Sun's deathknell. Apart from the fact that empirical data from the market shows they are migrating off SPARC onto other vendors' platforms, unless Sun can get a product to market to keep those old SPARC custoemrs onboard then Sun is dead, fullstop. They simply can't make enough money on the other products to survive, they need the old fattener of enterprise services. Neither Niagara or x64 will give them that, and their weak storage protfolio definately won't, especially whilst Sun continues to waste money on software deadends like MySQL.

"....Perhaps they like the guaranteed binary compatibility of Solaris...." I love that old chestnut - binary compatibility between what? You can't run a SPARC Solaris app on Solaris x86, which is the migration that I see discussed most. You can barely run a SPARC Solaris 9 app in a container on SPARC Solaris 10, and then you have to run it on the Niagara box you have the problem of running an old app with strong threading demands on a system that is designed for multiple mini threads - not good, hence the Sun bleating about Niagara as an Oracle system - comic! Or Sun have to give the money to FSC if you buy an M-series system, so Sun is simply treading water at best whilst desperately hoping on Rock. And who knows what the constraints there will be if you ever try running an old SPARC app on Rock (and don't tell me you know as Rock isn't out in the wild so the reality is you haven't a clue what could change between now and the date Rock gets out, if it ever does).

I speak to plenty of other admins and they tell me one of the biggest problems is the shaky Sun roadmap and the fact that Sun have lost mindshare in the boardroom. Sunshiners may like to make decisions based on nostalgia, but busineses rarley do, and I expect IBM have simply decided the effort of GPFS for Solaris is simply not worth the dwindling return.

0
2
Silver badge
Happy

Latest nova reports!

Just for the Sunshiners - Sun's JAVA ticker is down 4.21% so far today (16:44pm GMT), which is about three times what the other tech companies are showing, and the market cap is a paltry $2.94bn. IBM's market cap is $110.48bn, and hp's is $84.47bn. Even Netapp's market cap far exceeds Sun's at $4.39bn. Maybe IBM will look at integrating GPFS with WAFL, it would seem a better bet.....

0
1
Stop

Wow, Matt ...

AYB? YSB ...

You talk about how easy it is for you to "set off" the Sun supporters (love your little cutesy nicknames, BTW) - but then you're the one who types a 1 page rant.

I hope I get an irony meter for Christmas, because the one I have is now hopelessly destroyed by this thread.

0
0
Happy

Re: Re: Bill

"Well, apart from the fact HP has a very long and established relationships with Veritas (the reason the AdvFS got dumped) and Oracle (number one Oracle RAC and grid partner) and so doesn't need GPFS, HP also have the PolyServe people and could probably produce a better UNIX grid product if need be."

Is that why HP uses Lustre for all of their Linux HPC Clusters? Because they "could probably produce a better UNIX grid product if need be."

I will try to control the urge to resort to the name calling that Matt seems inclined to do when he has no logical or valid points. Matt, please use some logic, it makes these discussions much more interesting. You say that Solaris is slow, but when the FACTS are given, you say they are just marketing. Then you show up on another thread saying that another Sun product must be slow because it has not shown any benchmarks which HP has shown. Then when someone shows you the benchmarks again, you say that they are just benchmarketing.

Your antics are too consistent and repetitive. Get some new shtick Matt. No one takes you seriously as you find a false premise and even when everyone proves you wrong, you stick to it.

If HP is such a great company and can do no wrong then why do you find it so necessary to attack the competition so much (namely Sun)?

0
0
IT Angle

Re: Latest nova reports!

On that note, HP is down 3.25% already today, which obviously means nothing, but since Matt the HP Sales Grunt felt it was important I thought I would bring it up.

What a buffoon.

0
0
Thumb Down

Matt, Matt, Matt...

If yer gonna take on Webster's role here, you need to learn to season your "I hates your computer platform" rants with more CAPITAL LETTERS and exclamation points!!! Seriously, can nobody give good old-fashioned rabid raving any more?

0
0
Silver badge
Happy

RE: Re: Re: Bill

As I recall, Lustre was just one option - hp also used SteelEye. Hp's use of Luster is historic and pre-dates the PolyServe purchase, and Veritas were very late getting into the Linux party so hp simply went with the community flow (I'd explain that to you but you Sunshiners just can't get your head around the idea of working WITH the Linux community). It also long pre-dates Sun's purchase of Lustre. Of course, I can guarantee hp will be moving off it very shortly seeing as now Sun have bought Lustre it is doomed. Sun's history shows there is no quicker way to kill a software product than for it to be bought by Sun.

As for benchmarks, I asked for recognised benchmarks for Niagara, the only ones you could provide were carefully crafted ones disowned by Oracle. Not much weight when you consider other vendors like IBM and hp post SPEC and TPC measures, and Oracle is happy to stand by their application benchmarks. But when I gave examples of our own investigations here you poo-pooed them, so pot meet keetle. So how did you exactly prove me wrong? Oh, I recall - because you whined for so long, that must make you right! Grow up.

Oh, and for the AC Sunshiner - the nova implosion is now down 6.01% on the day (16.48ET), still twice more than hp or IBM, and that puts the Sun market cap at $2.89bn. At that rate it could be three weeks and JAVA will breach the $1 shareprice, by which time the market cap will be under a $bn. Does anyone want to start a sweepstake on what point they think Ponytail will have the decency to finally fall on his sword? I'm going for the point at which JAVA gets delisted.

0
1
Happy

Re: RE: Re: Re: Bill

HP Sales Grunt, HP Sales Grunt, HP Sales Grunt... Oracle used the same legalize that they put in all white papers of the sort. The same wording is in papers that they put out for IBM, Dell, and HP. It does not make it any less important. Even IBM states that SPEC is worthless, but there are enough ignorant users that use it that they continue to use it (besides, it is extremely easy to design for). Also, there were a lot of benchmarks that were released other than this one.

The fact that you only believe in SPEC and TPC says a lot about you Matt the HP Sales Grunt. They only test very specific parts of the architecture and are not very specific to the actual end product such as web, java, peoplesoft, sap, etc. As a matter of fact SPEC and TPC say exactly nothing about how a system will work with these industry benchmarks. No I am starting to question whether you really are a Sales Grunt and perhaps you are a pointy haired manager type. That would explain your reliance on such benchmarks as SPEC and TPC to tell you what will best run your specific application. Maybe I'll change your nickname to Matt the Pointy Haired HP Manager.

0
0
Go

@Matt Bryant

Clearly you're running scared, attacking what you don't understand. You've just got to love the way that admins all the world over love you and trust you with their innermost fears and concerns.

But maybe, just maybe if you actually tried to learn Solaris, you'd find that you actually like it. Performance is WAY improved since the 400Mhz processors you obviously last used when running Solaris. What about your favourite baby? PH-UX isn't exactly a screaming perf-fest - nobody in the real world is running it! And you think that Veritas (actually Symantec btw - try and keep up) is a good thing? Your entire filesystem roadmap based on a third-party who's losing market share?

Why would I want an OS from a company that makes more money from printer ink than anything else?

Working with the community? Yes - Sun obviously doesn't know how to do that as NFS, JAVA, OpenOffice, RPC, XDR were all just figments of imagination.

HP are certainly no innovators in terms of the Linux world. Oh and by the way... where IS the PH-UX source?

0
0
Unhappy

IBM Just Can't Figure it Out

That's too bad since SUN and other open SPARC partners have figured out how to get CoolThreads into an IBM proprietary blade chassis, get Solaris onto a proprietary IBM Mainframe, get Solaris onto IBM proprietary PC's.

I guess IBM just can't figure out how to port GPFS to Solaris.

So much for IBM mindshare...

0
0
Anonymous Coward

Doomed.

FS wars are nothing new. I believe NetApp will also pitch their (unreleased) pNFS implementation as competing in the same market.

Solaris to AIX? No thanks. Maybe to Linnex. It's about personnel - and good Solaris people don't want to cross-train to AIX, but do pick up Linux rather handily. Of course, IBM would be happy either way, except that personally I'd rather specify Linux on HP blades.

And yet, Solaris is still a great DC OS. Very happy to keep it in the standards register for some time to come.

0
0
Silver badge
Happy

RE: Re: RE: Re: Re: Bill and Yet Another AC

RE: Bill:

"HP Sales Grunt, HP Sales Grunt, HP Sales Grunt..." This is just another example of Sunshiner blinkers - you can post the same fact over and over and they just don't listen. I'm not HP sales, get over it. I am one of the majority of customers Sun is failing to satisfy.

"....Oracle used the same legalize that they put in all white papers of the sort...." No they didn't. I have plenty of Oracle and vendor whitepapers without any such statement. I have even seen old Sun/Oracle papers without a statement like that.

"....The fact that you only believe in SPEC and TPC says a lot about you...." Let me explain how the enterprise works seeing as you obvioulsy haven't been there. At the top are a whole lot of people without much IT knowledge called "the board". They sit down, talk to analysts, look at the market, and try and predict where they want the business to go in X years. Then they draw up a list of business requirements to get there, and then ask us in IT what we will need to fulfill those bizz reqs. This usually means the discussion starts with an application selection, not an OS. If the application only runs on one platform (say IBM mainframe) and there is no viable alternative, then the choice is clear. But, if you go for something say like SAP, where there are many options, then you need a starting point to compare them. In such situations, using SPEC and TPC raw figures as suggested by the app vendor will get you in the right ballpark and save you having to test every option out there. Then you can use a PoC to compare (which is also a good way to get the vendors concerned into a price-war ;) ). That is why SPEC and TPC are still valid.

"....As a matter of fact SPEC and TPC say exactly nothing about how a system will work with these industry benchmarks...." Strange then that Sun used to sprout them so often when SPARC had competitive figures. But now they've fallen off the performance map they're keen to avoid that type of comparison. What a surprise - NOT!

RE: AC:

"Clearly you're running scared, attacking what you don't understand...." Running scared from what? We have less and less Slowaris in our environment every year. There is a simple reason - we do what we need to do better and cheaper on other platforms. This seems to reflect the way of the market as shown by Sun's awful market position. The only people running scared are the Sunshiners.

"....But maybe, just maybe if you actually tried to learn Solaris, you'd find that you actually like it. Performance is WAY improved since the 400Mhz processors you obviously last used when running Solaris...." We have a small contingent of Sunshiners that insist we appraise Slowaris regularly against hp-ux and RHEL. Seeing as we also have a policy of trying to avoid vendor lockin, we usually take tenders from at least two parties. This means I have tested Slowaris on x86 blades from Sun, hp and IBM, against WIndows and RHEL, and Slowaris on Niagara and M-series against hp-ux, RH and Windows on Integirty (we do still have some VMS but that is usually for apps where we don't want to port to another OS). We have in a few instances bought FSC kit for Slowaris apps where we couldn't or wouldn't port due to difficulty, but otherwise we have not bought any new Sun kit for over three years, even when Sun walked in with their pants round their ankles, because they had the least to offer. In every major project it was the board that signed off, and they have zero loyalty to any vendor, they just look at our PoC results, prices and vendor relationship. It may amuse you to know that in one instance I actually recommended some Niagara kit for a web-based project but the board went for RHEL on Xeon instead, mainly because the RHEL option was cheaper and proven.

"....PH-UX isn't exactly a screaming perf-fest - nobody in the real world is running it!..." Well we are, so I can calmly ignore the rest of your childish rant. In fact, hp-ux on Integrity is gaining share on Slowaris on SPARC/Niagara, as the IDC and Gartner market figures show. As an example, go check the telecoms billing market, which used to be a Slowaris paradise - it is now ruled by hp Integrity. And one reason is hp's wider, deeper product portfolio puts it in a better position to meet all a customer's requirements. Sun failed to diversify and failed to innovate, which is why it is now so deep in the doodoo.

"...Why would I want an OS from a company that makes more money from printer ink than anything else?..." Oh dear, your snobbery is showing. Go check the figures, whilst hp make plenty of money from the printer bizz (a case of innovation - they didn't just get to where they are by hoping, they diversified and innovated), they also make shedloads from x86, storage, software and big iron, amongst others. The reason you want an OS from company like that is because they will have the money to invest in research and development. Unlike Sun, where they are bleeding through their reserves at such a rate I have seen analysts' reports saying they will be broke inside three years. You can sneer at print all you like, but hp's print bizz will be around long after JAVA has been delisted and Sun relegated to the Silicon Valley museums.

"...Working with the community? Yes - Sun obviously doesn't know how to do that as NFS, JAVA, OpenOffice, RPC, XDR were all just figments of imagination...." Lol, but the difference is hp and IBM work with the community an make money from Linux, whereas Sun just p*ss off the community and make vritually zero money from Linux, and SFA from Open Solaris. Sun is making a big loss on open source and the revenue from the x64 servers is sold is not going to make enough in the long run to keep Sun afloat without radical trimming of both the Sun product lines and staff. Java is a prime example of Sun failing to make any return on an expensive project.

"....HP are certainly no innovators in terms of the Linux world...." Really? So how come hp have people on a number of Linux steering commitees then? Unlike Sun. And how come hp were supporting all their x86 range with Linux when Schwartz was still mouthing off about "Solaris on SPARC and nothing else"? Take the blinkers off, take a deep breath, and then do some reading. Hp's long involvement in Linux is on record, based on the idea of using Linux as a COMPLIMENTARY product to hp-ux and Windows, rather than the Sun's history of fearing, FUDing and fighting Linux.

"....where IS the PH-UX source?..." Why should hp open source hp-ux? They don't need to. But they did open source OpenVMS, arguably a much better product than Slowaris, especially on Integrity with hp's superior storage and software offerings (gotta keep the VMS trolls happy, they're mainly old and get all grouchy round Christmas time).

But it is the season of goodwill, so I'll lay off the Sunshiners and simply wish you all a Merry Christmas and a Happy New Year, and hopefully Santa will bring you all some Linux skills!

0
2
Happy

Re: RE: Re: RE: Re: Re: Bill and Yet Another AC

It seems you missed some important facts, Matt.

"At the top are a whole lot of people without much IT knowledge called 'the board'." - yea, they buy IBM and HP... LOL!

About SPEC... "Strange then that Sun used to sprout them so often when SPARC had competitive figures. But now they've fallen off the performance map they're keen to avoid that type of comparison. What a surprise" - what a surprise that SUN led in 1, 2, and 4 CPU socket SPEC performance over the past 2 years and you didn't have a clue! What a trip!!!

"Sun failed to diversify and failed to innovate, which is why it is now so deep in the doodoo" - that is why they sell & support M$ Windoze, Linux, Solaris, SPARC, Intel, AMD? LOL!

"but the difference is hp and IBM work with the community an make money from Linux" - that is why NFS, JAVA, OpenOffice, cluster file system, etc. all came from SUN and SUN re-sells Linux servers? ha ha ha ha ha!

"open source OpenVMS, arguably a much better product than Slowaris" - I bet Solaris will fork processes faster than OpenVMS! What a trip!

You're funny, Matt!!!

HA HA HA HA!

0
0
Silver badge
Happy

RE: David Halko

And a Merry Christmas to you too, Dave. Haven't seen you post since your expansive and very informative marketeering exercise on the "Sun double teams Xeon chip" article back in August. However, I have noticed a quantity of David Halko comments on the Sun blogsites, and I was especially tickled by one at http://blogs.sun.com/bmseer/entry/new_new_news where you were complaining about a T2000 benchmark session; "....I am sure that the server scores would be much better than the competition... since performance is not everything!" Says it all, really. But back to the point.

"...."Sun failed to diversify and failed to innovate, which is why it is now so deep in the doodoo" - that is why they sell & support M$ Windoze, Linux, Solaris, SPARC, Intel, AMD? LOL!" <Sigh> Like a lot of Sunshiners, you fail to understand that diversity does not mean just staying in one quarter of the datacenter. Hp, IBM and Dell all have products that span both inside and outside the datacenter, with hp probably being the most diverse. What this means is that if you fall behind in one area, returns from others will still allow you to invest in research and develop back to parity or being the market leader. Sun's problem is the narrow range of products means it is not generating enough money to invest in research and marketting activities. Hence it is poorly placed to meet the downturn, poorly placed to compete with competitors able to loss-lead and make the money back in other areas, and poorly placed to innovate with new products. Sun joined the Wintel/Lintel party so late and after decades of abusing both Windows and Linux communities is it any surprise they lag so far behind in sales?

"...SUN led in 1, 2, and 4 CPU socket SPEC performance over the past 2 years..." What SPEC bench was that? I wasn't aware there was a SPEC_fastest-loss-of-market-cap. Of course, the market seems to have missed your mystery SPEC results too, seeing as Sun's marketshare is rapidly declining.

"....that is why NFS, JAVA, OpenOffice, cluster file system, etc. all came from SUN and SUN re-sells Linux servers?...." NFS was developed for UNIX, not Linux, originally in 1983. Sun had no intention of it being used anywhere other than UNIX and with the intent of embedding Sun OS in the commercial datacenter to allow it to attach Sun workstations to mainframes and other vendors' UNIX servers. Linux didn't kick off until 1991, so nothing to do with Sun "working with the Linux community".

Java was originally released in 1995 in an attempt by Sun to stop Microsoft taking over the webserving bizz Sun had come to depend on. Sun had to give it away free as they were being killed by Windows. Despite claiming it was "open source", Sun retained close control and didn't release the whole code when it finally GPL'd some of it in 2007, which exposes Sun's claims of openess as just so much baloney. Sun has still made zero money from Java and is unlikely to. Instead, competitors like hp have used better Java performance to sell hp-ux servers in the datacenter, and cheaper Lintel/Wintel kit at the edge, accellerating Sun's decline. Again, this had nothing to do with Sun "working with the Linux community", just another failed Sun survival strategem.

"...OpenOffice...." You mean the product Sun bought from Germany's Star Division Corp? The product where Sun signed a deal with Micorosft to protect it's own StarOffice bizz but hoping that M$ would then sue the OpenOffice community out of bizz (see http://www.infoworld.com/article/04/09/16/HNopenoffice_1.html). You actually want to show Sun's supposed support for Linux by highlighting a product where Sun cynically tried to shaft them? Oooh, you must be a whizz at chess - not!

"....cluster file system...." Which cluster file system? There are plenty out there. And don't even go there with Slowaris clustering, a product so unreliable Veritas made a fortune from selling Veritas Cluster Server to Sun shops.

"....SUN re-sells Linux servers?..." Yes, Sun resells Windows and Linux, at a much slower rate than hp or IBM. And only because they have to as SPARC/Slowaris is a dead product and Sun - after years of FUDing Linux, Windows and x86 - had no other option for survival. The difference is hp and IBM have years of experience, integrated products and solid services offerings, whilst Sun has a few boxes and a lot of ill-will.

"....I bet Solaris will fork processes faster than OpenVMS! What a trip!...." I know, but the VMS trolls were looking glum. ;)

Here's hoping Santa kept you on the good boys' list and bought you http://www.amazon.com/Linux-Administration-Beginners-Guide-Fifth/dp/0071545883/ref=sr_1_3?ie=UTF8&s=books&qid=1230215519&sr=1-3

/point, laugh in a festive manner, you get the idea.

0
2
Happy

Re: Re: David Halko

Wow Matt, I can't imagine that a mere HP Sales Grunt would do the level of research on the competitions web site that you appear to do. You must actually be in HP marketing. That would explain your complete and utter fear and distaste for anything Sun. I could understand your constant trolling on Sun related Register articles, but to spend the rest of your obviously busy admin life (that's sarcasm if you missed it) you mist be in HP marketing... I'm not sure how the industry as a whole would take the ethics of one vendor putting out such obvious FUD on public forums, but perhaps you're doing it without your employers permission.

0
0
This topic is closed for new posts.

Forums