HP seems on its way out, or so claims Sun. Sun’s Global information systems strategy office, Larry Singer, says Sun hears a consistent message from HP customers: “They have no idea where HP is going. If you chase the decision up the stack to the OS, you realise HP’s OS has had no major release since Fiorina took over,” he says. “People only have three legitimate OSes: Windows, Red Hat or Solaris. As soon as HP talks to 9900 or Alpha customers they have to say HP does not have an OS.”
Yeah, I’d listent to a company that hasn’t yet figured out that Linux != Red Hat.
Fools.
Yeah, I’d listen to a person that hasn’t yet figured out that Linux != an OS.
Fool.
Yeah, I’d listent to a company that hasn’t yet figured out that Linux != Red Hat.
They have a point when they say that for Sun, Linux does equal RedHat, since RedHat is the single biggest provider of server solutions based on Linux and thus Sun’s direct competitor.
Actually, Sun’s text seems to be correct anyway. Linux != Red Hat, but Linux per se is only a kernel, not an OS, and Red Hat is the most important (from Sun’s POV) Linux-based OS.
Mac OS X?
OS X Server’s share in the server world is marginal compared to the big players RedHat/Linux, Solaris and Windows.
… has a lot to answer for
I guess they haven’t realized that AIX is kicking their butts
Just because IBM’s sales have increased while Sun’s have decreased does not automatically mean that AIX is kicking their butt. There is still more Solaris than AIX. And similarly to HPUX, AIX has seen very little innovation in recent years. IBM has not been subtle about that fact that they see Linux as their future in the Unix market, not AIX.
Absolutely. IBM put it’s weight on Linux. The very fact that IBM uses the phrase “Open source” in lots of their PR material and yet could not live up to these ideals when it comes to AIX, shows that they don’t want to invest even the effort of opensourcing AIX!
So, IBM is letting AIX die in a closed source. That’s some indictment.
“Absolutely. IBM put it’s weight on Linux. The very fact that IBM uses the phrase “Open source” in lots of their PR material and yet could not live up to these ideals when it comes to AIX, shows that they don’t want to invest even the effort of opensourcing AIX!
So, IBM is letting AIX die in a closed source. That’s some indictment.”
And AIX is still outselling Solaris, an OS Sun has sunk at least half a billion (for Solaris 10 development) of valuable R&D money into.
And oh, by the way, HP-UX – that OS that hasn’t been updated since Carly took over – is also outselling Solaris.
According to IDC / Gartner (can’t remember which), the latest numbers have Solaris solidly in third.
And both IBM and HP are supposedly treating AIX and HP-UX as bastard step children and not innovating/updating, while Solaris is the be-all-to-end-all technology leader, latest and greatest OS – all of this according to Sun.
And yet AIX and HP-UX are still outselling Solaris.
That’s an even worse indictment. 🙂
And AIX is still outselling Solaris, an OS Sun has sunk at least half a billion (for Solaris 10 development) of valuable R&D money into.
And oh, by the way, HP-UX – that OS that hasn’t been updated since Carly took over – is also outselling Solaris.
Are you conlating revenue numbers with volume sales? I am sure you are.
Read those reports again and post them here for all to see.
Maybe you are talking about this:
http://news.com.com/Sun+dethroned+in+Unix+market–maybe/2100-1010_3…
IDC and gartner are conflicted on who has the crown for UNIX server revenue. Sun is ether number one or two depending on who you ask. Again in revenue not unit sales.
Sun has been heavily discounting thier products, where as IBM overpricing them. Explains the revenue part all right. Solaris still out sells the others it only sells for less because of the discounts.
Sorry but I have to disagree. Our data center went from 2 AIX servers a couple of years ago (we were mostly Sun, HP) and we now have close to 300 LPARS. IBM has spend a lot of money developing their new frames, particularly 595’s, and as far as Linux is concerned, they have no intention of replacing AIX with it, Linux is offered as an alternative. Everything they develop is designed to run on AIX first and then ported Linux within 6 months or so.
“Everything they develop is designed to run on AIX first and then ported Linux within 6 months or so.”
Erm – isn’t the OS supposed to abstract the hardware, not vice versa?
I was referring to software development
Thats not what you said, but anyway at the current time pushing vendor UNIX is extremely attractive to vendors (ahem lock-in ahem) and so this is what they always say/tout/push/market.
The market forces/trends clearly say otherwise
UNIX => Linux
So no matter what they say its only true for a small time window.
You will find that IBM Software Boss has been saying for years that AIX is being transitioned to Linux. Naturally, this will probably take 10 years, so customers don’t see it. But savvy customers would ask any vendors for clear roadmap so they don’t make strategic mistakes. It makes sense for IBM to standardise on Linux has IBM has 4 incompatible server business: X86, RS6000, AS400, mainframe. They killed their X86 OS already (OS/2), and sales of AS400 has been declining. IBM has managed to merge the i and p Series under 1 chip. I will continue to merge these 4 server business from both HW and OS point of view. Unfortunately for IBM, the only choice is Linux. RedHat is basically their only saviour to merge the 4 hardware business.
If you look at AIX 5.4 road map, you will not see major improvements. Same goes for 5.3 and 5.2. The enhancements are primarily to support Power chip feature. If you take out the HAL component in the OS, they have essentially stop major enhancement at the OS itself.
And the CAPP/EAL4+ install option for AIX 5L 5.2 isn’t an enhancement?
Sorry, it is one. But my point remains valid. If you compare AIX to Linux, Windows and Solaris, and list the _major_ enhancements in the past 2-3 releases, you can tell where IBM is spending the resource. I’m concern that another enterprise UNIX is being killed by the owner. We used to have lots of UNIX, and the choice is becoming less and less. I know this will upset Linux zealots, but I consider Linux 2.6 vs AIX 5.3, I’d take AIX.
“…and sales of AS400 has been declining…”
The numbers from last quarter seems to suggest otherwise:
” Hardware revenues for the Systems and Technology Group totaled $5.0 billion for the quarter, up 7 percent. Revenue growth from S&TG eServer products was driven by iSeries midrange servers, which increased 25 percent; pSeries UNIX servers, an increase of 15 percent; and xSeries servers, which increased 11 percent. Revenues from the zSeries mainframe product decreased 4 percent compared with the year-ago period. Total delivery of zSeries computing power, which is measured in MIPS (millions of instructions per second), increased 18 percent. In addition to eServers, revenues from Storage Systems and Microelectronics increased 11 percent and 14 percent, respectively.”
http://www.ibm.com/investor/3q05/3q05earnings.phtml
One would think that the iSeries customers would be holding out from the Power5+ upgrade…guess not.
Don’t take 1 quarter, or Q-to-Q, or Y on Y, or 1 market segment, or 1 geo, etc.
Take 5 years, worldwide, all shipments. Then plot this againsts Windows and Solaris. You will see the overall decline. OS takes years to die, even decades. If you take 1 quarter, or even 1 year, you will not see the main picture.
…and as far as Linux is concerned, they have no intention of replacing AIX with it, Linux is offered as an alternative….
I am sorry sir but I believe you drank the Koolaid.
And btw, why are your running LPARS and not upars?
(genuinely interested)
Uhm, I think that is the exact definition of getting your butt kicked in the server market.
Just because there hasn’t been any updates does not by default mean that the os is bad. It could actually mean the opposite.
However I get the impression that someone at Sun has identified HP as one of the key rivals somewere because I remember reading another “HP sucks” coming from Sun. Usually I’d suspect that marketing is taking over and that actually HP isn’t as bad as they want us to belive.
Also they didn’t mention Novell, did they ?
No, and I wonder why they didn’t mention Novell… because they’ve disappeared / is disappearing from the market as well.
Well if you mention Redhat you might as well mention Novell/SUSE.
Just because there hasn’t been any updates does not by default mean that the os is bad. It could actually mean the opposite.
It’s a nice thought but sadly not in the real world(tm)
However I get the impression that someone at Sun has identified HP as one of the key rivals somewere because I remember reading another “HP sucks” coming from Sun. Usually I’d suspect that marketing is taking over and that actually HP isn’t as bad as they want us to belive.
Err, where have you been?. HP and Sun have been key
rivals always. This is just the latest installment
in the pissing contest.
I hava a feeling this guy has some kind of brain tumor that turns him into a desperate, flaming jerk once every week. This brand of marketting wins no friends. Sun just needs to bend down minding their business of making and selling great servers. If they make the right products at the right price they won’t need bad-mouth anybody else to sell them.
I do not understand why Apple is not Strongly pushing with MacOS X??
I mean, I checked it out and it really is a very nice server system, definitely on par with most of the Linux offerings (and way ahead in terms of GUI)…
I can see that MacOS X server would not be good for 100k users, but especially for the size about 1000, it should be pretty fine…
Apple could get a nice market cut from Unix guys… I guess, this is not their business….
Actually OSX quite frankly sucks as a server due its poor threading performance inherited from the BSD’s mach kernel.
Please read the following benchmarks from Anandtech.
http://www.anandtech.com/mac/showdoc.aspx?i=2520&p=7
Nice after 2 concurrent users, a dual G5 chokes running OSX, and a dual G5 running Yellow Dog Linux kicks its butt.
But this is just MySQL, will other applications suffer?
Yes, any thread based application will suffer under OSX. OSX uses USER threads for applications. Compared to kernel threads used in Windows, more advanced Unixes (e.g. Solaris, AIX, etc), Linux kernels 2.6+
That said, Windows, Linux, and other flavors of Unixes have user threads too. Its just that they do not limit trusted apps to use a much slower USER thread.
But I’m not an expert on Kernels, so don’t take my word for it, google it, or check out that Anandtech article.
Actually OSX quite frankly sucks as a server due its poor threading performance inherited from the BSD’s mach kernel.
BSD’s mach kernel??
FYI neither FreeBSD, OpenBSD or NetBSD uses a Mach kernel…
OS X kernel (XNU) is a hybrid kernel (based on the Mach and BSD kernels)
Though it may not have much in the was of face time in the media OpenVMS is still a computing powerhouse for medical systems, and HP released 8.2 not too terribly long ago. I like Sun products, but their marketing approach lately has seemed soaked in whine. “Blah blah blah DELL sucks, HP sucks, etc.”
I hate to say it, I think Sun is correct here. I used to love the old HP, but this new HP is now a “me to” company. They don’t innovate, they just adopt.
Ever since Carly took over, they became average.
Well, the issue isn’t so much about ‘innovation’, its about a clear product road map that says, “here we are, this is where we’re going and in 5-6 years, this is what our product line up will look like, based on todays assumed scenario’s”.
If it were me, nothing wrong with the Itanium architecture, it just lacks a high profile operating system along with software – standardise the WHOLE server sales around two core operating systems, from end to end. Solaris 10 for the Itanium/x86 sales and Windows 2003; Windows 2003 for the obvious fact that its the operating system people want and Solaris because of the level of ISV support – and if enough money is thrown at the vendors, you’d start seeing Solaris Itanium versions appearing.
About the only thing HP would have to do is port Solaris 10 to Itanium.
About the only thing HP would have to do is port Solaris 10 to Itanium.
It’s just that HP would never swallow their pride, dump HP-UX and Tru64, and support Solaris instead
I am not sure about HP and Tru64. HP has basically dumped the clustering architecture of Tru64 (one of the best software products from DEC), which is truly best of breed, and decided to license it from Veritas! No, I am not making this up. HP has given up on DEC clustering and decided they’ll rather use Veritas’! When we heard about this in my departement, we were flabbergasted (best place to use this expression ever), and HP has sunk below the “joke” level in our imaginary charts. Their dumping of PA-RISC and Alpha already tarnished their reputation quite a bit, but this was too much.
Regarding PA-RISC and Alpha; I’m sure they could technologically justify it; what Itanium requires is this; huge price decreases – get workstations pumping out of the factory and selling for US$1395 per workstation – that would put it within spitting distance of a high end Dell workstation.
It isn’t the performance but the PRICE performance; if they priced around the same level as say the Opteron, Itanium would chew and spit it out, but the fact is, they aren’t willing to take a few quarters of initial loses to build up volume in the market place.
As for ISV’s; infact, SUN AND HP BOTH need to approach software companies and hand over cash to get software ported – just a sharp remark – maybe that would be a better investment than the money wasted on the number of fruitless acquisitions SUN has done.
As for Solaris – I’m sure SUN still has alot of the code left from their initial port to Itanium; it wouldn’t take much to get the ball rolling again.
As for the DEC clustering, I think they’ve dumped a large portion of DEC stuff already – why spend money on a product when you can be a whore to another? As for the issue surrounding it; sure, there would have been some fundamental technical issues relating to the different operating system architecture, but I doubt it would have been so infused that it would have been impossible to port.
“There’s no roadmap, no strategy, other than wait and see – no-one believes the story any more. They’ve lost position in storage and they’re becoming irrelevant in the enterprise space.”
The exact same thing can be said about SUN. Why are they always saying these foolish things. HP will eat their lunch if the don’t watch it.
AIX development in 5.3 is not flat. Note that although many of the improvements are geared towards supporting the features in POWER5 chip, that’s because moving OS logic into the CPU is the path IBM has chosen.
Also, there are more Solaris systems out there, but the vast majority are workstations, not servers. IBM doesn’t sell AIX workstations. AIX has maintained a significant lead in the data center.
One thing IBM does need to work on, however, is their maintenance level releases. They keep breaking things. And that makes AIX suck bad, mkay?
Yes, IBM does sell workstations:
http://www-03.ibm.com/servers/intellistation/power/
And where exactly did you get this idea from “but the vast majority are workstations, not servers”. I am sure a number of us would be interested in where you got that information from.
Agreed; the largest UNIX workstation vendor is actually SUN not IBM; IBM’s strength has always been in the server market, their workstation market has always been enemically small, but then again, they’d probably justify it by saying that there are more efficient ways of developing and using POWER processors than having one dedicated workstation.
IBM has great Open Systems technology and a rock solid easy to use OS(AIX). There problem is and always has been is the cost of their product vs. competeing solutions. They are the Ferrari of Unix based systems.
True, but then again, SUNs workstations are a rip off when compared to a similarly configured Opteron system from SUN, loaded with Solaris (or dumped in favour of something else) – heck, why not bypass it and get a PowerMac?
Some people just don’t get it. Customers run software and buy based on Apps and software. They don’t have a hardware perfromance fetish like most geeks.
From what I have heard the SunBlade 150 is still Sun’s largest selling workstation product. Go figure.
I find that hard to believe, given the enemically crappy graphics card installed; if one needs a system, one might as well bypass that crap and go straight for the big stuff like the Blade 1500 or 2500.
>>
“People only have three legitimate OSes: Windows, Red Hat or Solaris.
<<
I am not great fan of AIX, but I think it’s a fairly popular OS.
Also, RedHat is a company, Linux is the OS. This is something that Sun has never understood. To Sun Redhat = Linux. In particular, I think Sun may be blind-sided by Novell/SuSe.
And AIX is still outselling Solaris, an OS Sun has sunk at least half a billion (for Solaris 10 development) of valuable R&D money into.
Nice logic. IBM make more money from licensing an OS
than a competitor (Sun) does by giving theirs away for
free. How much R&D have IBM put into AIX 5.3? They
are going linux all the way, your blind if you don’t
see that.
And oh, by the way, HP-UX – that OS that hasn’t been updated since Carly took over – is also outselling Solaris.
According to IDC / Gartner (can’t remember which), the latest numbers have Solaris solidly in third.
Marginally in 3rd. And thats revenue not volume. And
being as Solaris is FREE and support costs half what
Redhat charges let alone what IBM/HP charge YOU go
figure.
And yet AIX and HP-UX are still outselling Solaris.
No they are not.
AIX and HP-UX are both great operating systems. I
just happen to think Solaris is better. Better
ENTERPRISE (forget the desktop for now) features
and opensource. HP are turning into nothing more than
a jacked up PC vendor and Intel shill. Just think
how bad their business would look if they didn’t have
all those toner sales to make everything look rosy 😉
Sun are growing faster in the low-mid range than their
competitors. The attrition is in the high end where,
unfortunately, the biggest margins are. It will be
interesting to see how things look in 6 months after
the Niagara release and when sites are able to get
USIV+.
Erm, wtf has Mcnealy got to do with the topic?
I am not great fan of AIX, but I think it’s a fairly popular OS.
Yeah, it’s very popular on IBM UNIX servers. Did you
know that Solaris 10 has nearly 3million registered
downloads? 2 million of those are x86/x64. Thats a
lot of kit not sold by Sun which can run it’s OS.
Also, RedHat is a company, Linux is the OS. This is something that Sun has never understood. To Sun Redhat = Linux. In particular, I think Sun may be blind-sided by Novell/SuSe.
Sigh. Why don’t people get this, it’s not difficult.
Linux is NOT the OS. The commericial distributions of
linux (linux kernel + user apps) which are currently
relevant in the market are: RedHat. Even in the linux
only market, Suse is still a minor player. If you are
running a small company you probably don’t care but if
you are a bigger company with some $$ to spend on real
applications and support then you are probably going
to end up with RedHat. To all intents and purposes, the
enterprise world of running linux is just as locked
in as running UNIX.
> And AIX is still outselling Solaris, an OS Sun has sunk at least half a billion (for Solaris 10 development) of jvaluable R&D money into
AIX is outselling Solaris, are you freaking stoned or just drank too much of IBM koolaid? AIX volumes (no. of systems) is a mere fraction of what Solaris gets in enterprise space. Don’t read in too much into what IBM is saying about their increase Unix sales — their Unix increased in dollar terms and not so much in number of units. The only reason IBM can brag is because they sold a number of high end systems that amount to millions of dollars each, but in unit shipments IBM is still far behind Sun. Sun is the king of the hill as far as Unix is concerned and its position should improve with Solaris 10. AIX is a dud compared with Solaris 10, so we should be seeing more of Sun kicking IBM’s ass.
> If it were me, nothing wrong with the Itanium architecture, it just lacks a high profile operating system along with software – standardise the WHOLE server sales around two core operating systems, from end to end.
Oh, there is whole lot of things wrong with Itanic architecture. Itanic performs well on highly predictable workloads such as benchmarks or HPTC, but in realworld with the loads requiring very high degree of parallelism and not being predictable by the nature (databases, DSS, etc.) Itanic will always take a back seat to the likes of Power, SPARC64 and even good old UltraSparc. It was extremely stupid of HP to bet the farm on Itanic, since if Itanium goes down, HP’s whole enterprise strategy goes along with it. And guess what, Itanic is going down by all measures — the ISV following is just abismal and Itanium is already sinonimous with “collosal failure”! HP’s only saving grace could be to bring back PA-RISC and to at least retain the customer’s that were committed to HP 9000. Otherwise according to HP’s own surveys at least 50% of HP 9000 customers are destined to jump off HP altogether (just as the news story has described).
“Actually OSX quite frankly sucks as a server due its poor threading performance inherited from the BSD’s mach kernel.
Please read the following benchmarks from Anandtech.
http://www.anandtech.com/mac/showdoc.aspx?i=2520&p=7“
You dont understand what you are talking about. Thoses test have proven to be not correct at all there was a discussion about it Anandtech did not know what they measuring and and their results are simply not true. Check out how osx is doing in clustering market, its a very performant solution. And threading is particularly important in this kind of high performamnce market, so if the threading performance is osx would be poor as you say, how do you explain that osx is doing very well in this market?????
I would advice you to read about osx kernel and how threading is implemented. First BSD and Mach do not refer to the tme same origin as you want to say, i explained in osnews forum how osx kernel XNU is built, just go, take some documentation and read it.
I really think that apple is doing quite well in entreprice space, a recent study of Gartner has shown that Apple products are attracting more and more people, their growth rate in this market is very high, Apple recently entered in the top 10 of servers makers. Thats a good performance i thinks.
Solaris is a nice os, powerful (but with some limitations) and very robust, thats a really proven os. But that’s it, i mean for a lot of users and entreprices Solaris is too difficult to manage, to set up and to maintain. Osx server is simply years ahead in this regards. Sun did not understand that people want things to operat quickly and easliy, even in technical markets. Entreprices want a plateform that its easy to deploy, they do not want to deal again and again with complex computing, they want something modern. Solaris simply (or Linux in some ways too) does not provide this kind of new requests. Soloris still works like 10 year ago.
OS X server really gives you technologies that makes things like managing a huge set of machines really straitforward. And it still built on open standarts, which is great. I think thats why apple is more and more popular in this market.
Cisco use now OSX, Xservers, Xraids and Xsan for their entire mail infrastucture, that’s a good sign of the maturity of OSX.
People at sun are stupid sometimes, because they always think that they have the best, they control the market where they are playing. Sun did not realize that they lost the workstation market with the competion of Linux and the Macintosh. People turns more and more for those systemes than using Solaris for workstation. They did not realize that Linux became a strong competitoer in the server market, and they do not realize that its not only RedHat thats all Linux, and they do not realize that OS X is coming more and more closer to them. When it will be the time to realize i guess it will be too late for them.
It isn’t the performance but the PRICE performance; if they priced around the same level as say the Opteron, Itanium would chew and spit it out, but the fact is, they aren’t willing to take a few quarters of initial loses to build up volume in the market place.
I think you would actually have to sell Itanic for
less than Opteron to get better price/performance.
Itanic can’t play in the same space as Opteron,
particularly if you consider TCO.
As for ISV’s; infact, SUN AND HP BOTH need to approach software companies and hand over cash to get software ported – just a sharp remark – maybe that would be a better investment than the money wasted on the number of fruitless acquisitions SUN has done.
Who says that is not exactly what is happening right
now 😉