Time to say goodbye to Risc / Itanium Unix? Twenty years ago open systems was the battle cry that shook the absurdly profitable proprietary mainframe and minicomputer markets. The proliferation of powerful and less costly x64-based systems that can run Solaris, Linux or Windows is making more than a few Unix shops think the unthinkable: migrating away from Unix for their …
Here you go again with the same old rubbish, that article you quote is from 2003! It was a very different market 8 years ago, one in which IBM was playing second fiddle to Sun even after the dotcom crash, it is now leading the UNIX market, IBM are not about to kill AIX anytime soon.
You believe too much of what you read on the internet, including those useless benchmarks you keep quoting which as any seasoned SA will tell you mean absolutely diddly squat in the real world. They are PR/advertising, and nothing more than that!
How dare you tell a fellow UNIX professional not to worry about having work, how patronising!! We all have our preferred OS, being Ex Sun mine is obviously Solaris and always will be. But times change and you need to accept other vendors do some things better. In IBM's case, their hardware is currently far superior to Oracle/Sun's, although as an OS AIX sucks for various reasons (Get rid of the ODM to start with as it's a frickin pain when it goes wrong) Solaris is a far more elegant "proper" UNIX, imho.
Here's hoping Oracle crank it up a notch with the T4's so Solaris/SPARC stick around for some time to come.
Here's also hoping Kebabert starts to appreciate the good things other vendors can do, it's not all about SPARC/Solaris these days, as much as I wish it was.
Wow, thats the best you can come up with "you should calm down or you will get a heart attack", why don't you actually think about the points I have made and make an intelligent reponse to those. There are no "facts" or "benchmarks" to quote when it comes to making a response to my post so you may struggle I expect...
Kebabbert.....you cannot quote a misquote from 2003 as fact. In the same article IBM is saying that Linux will forever play catch-up to AIX.
What has happened since 2003? Well AIX was 3rd in the market and has now not only passed Solaris and HP but is 50% of the Unix market. I think a lot has changed since Power4 and AIX 4.3.
If anything HP-UX is in the biggest danger unless HP does something about database support. Solaris will need someone to actually buy a Exa-xxxxx before Larry wonders by he is bothering with solaris.
"....you cannot quote a misquote from 2003 as fact..."
Am I "misquoting" the official statements from IBM? How am I misquoting them? I am trying to explain to Jesper Frimann, that according to IBM, AIX will be killed off and replaced by Linux. How is this a misquote? Please cite me, and the article I linked to - and if you can point out any errors in my understanding of the article I will stop say false things. If you can not point out my misunderstandings, then maybe it is you who misunderstood the official statements from IBM?
.
.
"...What has happened since 2003? Well AIX was 3rd in the market and has now not only passed Solaris and HP but is 50% of the Unix market. I think a lot has changed since Power4 and AIX 4.3...."
So? It seems that you dont really understand me. I say that long ago, POWER6 was several times faster than x86 and it costed 5-10x more. Today, POWER7 is ~10% faster than Intel Westmere-EX but costs 3x more. Future POWER cpus will cost 1-2x more, but ideally will be cheaper than x86 or they will not sell.
Yes, a lot has changed since POWER4. The trend is clear: AIX/POWER is not a high margin business anymore. IBM had to lower the price of their best CPU yet (the mighty POWER7) because IBM is afraid of the competition from x86 and Niagara. This means something.
I predict that IBM's fear will increase, and future POWER cpus will be cheaper and cheaper. There will be a time in the future when IBM POWER can not compete with x86. x86 will catch up on POWER. You know that. I know that. This is true. The trend is clear, it will happen. Some time after, IBM will kill POWER, because IBM can not afford to loose money on expensive and slow POWER cpus, with a diminishing market share. That is when AIX will be discontinued.
Do you really think that IBM will release a future POWER cpu that is many times faster than x86, so IBM can continue to charge 5-10x more? No. Those days are gone. x86 will be faster. And AIX does not run on x86. So AIX on POWER will be no more. This is true. Even IBM executives see this trend.
Next year, the 22nm Ivy Bridge from Intel will arrive. It will be 40% faster (according to Intel) than today's Sandybridge. Thus, it will be faster than POWER7. IBM must make sure that POWER8 is much faster than x86, there will be no point in spending billions on R&D on a slower POWER8 cpu.
So, AIX and POWER is soon obsolete. Itanium is obsolete (this is a pity). Better start to learn Linux on x86. The thing is, the managers dont understand that POWER and Itanium has much better RAS than x86, and that x86 is a piece of buggy shit. They just look at the cheapest performance / money, and x86 is the future. The managers will only consider OS that runs on x86.
Well my comment to this post seem to have been lost, so I'll quickly do a response again.
"POWER is today only ~10% faster than x86, but costs 3x or more."
What you meant to say is that low clocked 1.2Billion transistor 45 nm POWER7 chip introduced one and a half year ago is still 15 % faster than newest higest clocked 2.6 Billion transistor 32 nm Westmere-EX chip, right ?
According to your link it's 15%, as for the price Anantech is comparing a HP blade versus a IBM rack server. That will never be a fair comparison.
Try to have a look at what the HP blade costs compared to the IBM POWER blade of equal size.. let me enlighten you. The HP blade costs 75% of the IBM POWER blade, if you look at equal bare metal configurations.
You can look that up on the HP and IBM online stores. Again that is not a factor of 3 in cost. And we are not even talking TCO yet, but only cost of the actual hardware.
// Jesper
You are just evading the issue.
I am claiming that POWER gets cheaper and cheaper, and Intel gets faster and faster. Then I gave some numbers. I said POWER7 is 10% faster than Intel. It might be 13% or 15%, so what? The trend is clear: POWER6 was several times faster than Intel before (on some benchmarks) and costed 5-10x more. POWER7 is only slightly faster, and much cheaper.
Instead of disputing the exact numbers I suggest you discuss the issue instead: POWER is getting cheaper and cheaper, and Intel is getting faster and faster. This is true, it is a fact. Intel cpus are getting more performance at a quicker pace than IBM POWER is getting performance. The Intel slope is higher which means the lines will cross: Intel will catch up on POWER.
Also, IBM has said officially that AIX will be killed. This is also true. It is a fact.
.
Now, here comes the guess part (these are not facts, but opinions):
I predict that when Intel is faster, or fast as POWER cpus, then POWER need to lower the price even more. And finally when the price is too low so that IBM does not earn much money anymore, POWER will be killed off. Just like IBM killed of CELL cpu. And that is the time when AIX finally will be killed by IBM. AIX is soon no more high margin business.
I am basing this prediction on IBMs official statements and I am extrapolating the clear trend (POWER getting cheaper, and Intel getting faster). In other words, I do have some substance in this prediction: it is not evil rumours without ground (not FUD). I am not FUDing.
You don't seem to get it.. Sorry but Keb, you don't really know what you are talking about. A processor chip is just a little bit of the total cost of a server. The TCO of POWER SERVERS is getting better. The machines of a specific class of servers, cost more or less the same, as it have done for the last what 6-7 years. So the whole premises for your little theory is.. well.. wrong.
It's actually pretty simpel, the hardware of a p570 based upon POWER5 from 2004 cost more or less the same as a POWER 770 cost you today. The real difference is that a POWER 770 has a factor of 10 the raw capacity, and with it's enhanced virtualization capability perhaps even a tad more.
Furthermore what you seem to forget is that the IBM revenue coming from POWER is going up, even though their $/capacity unit is going down and down. So I would presume that the profits are at least constant. So your statement is basically not making any sense.
// Jesper
I think you're missing the point... UNIX is not Linux, Linux is probably the biggest rival to UNIX, to count them together as a whole misses the point that, in general, UNIX is run on very expensive propriatery RISC hardware and Linux is run on comodity x86 hardware. If you lump them both together you don't see the migration from UNIX/RISC to Linux/X86/Virtualised.
The bang for buck of RISC processors/servers just doesn't stack up against a modern x86_64 server. A RISC server may be bigger and faster, but the cost is disproportionate and usually the extra capacity isn't used.
My former employer (top 5 global bank) are migrating off all proprietary unix (except for tactical essentials) and are replacing with Virtualised Windows/Linux (RHEL) and z/OS and z/Linux. The reasoning is simply that the bang for buck of the RISC processors running most modern UNIX OSes doesn't stack up against a modern x86_64 chip. This will be augmented with some of the new database appliances coming onto the market.
If they need really big throughput in Linux it goes on a Z server, otherwise it's vmware.
The UNIX guys really weren't very happy about this at all (why do people working in such a fast moving industry resist change so much?) as there was the typical unix guy suspicion of Linux. However they're coming round to the idea and can't really argue that much with the seer cost savings involved.
The author writes about Linux as if it were part of the Windows product suite, and not a tentacle of Unix. Linux is the most open unix of all, and has well and truly answered that 20 year old call for for "open systems!".
The lower midrange market was there fore he taking in the late 90's. MS jumped right in there and had a great time for 5 years. Then unix sent in its Linux steamroller, which continues to rumble back and forth in the datacentre.
The proprietory unixes have of course declined, driven by technology, but that's 10 year old news. This article focusses, oddly, on Itanuim/RISC, a minority platform if ever there was one. Yes, mainframes cost more than midrange, which costs more than x64. 'Twas ever thus. Only the old names changed - mainframe, mini and micro.
Unix/Linux are technically very similar and historically intertwined. In the context of this article, they should be held the same, cutsey project names notwithstanding. The author has instead lumped Linux with Windows simply in order to make a point about the commercials. Far more interesting to group Linux more correctly with Unix, and ponder the evolution of the OSes from that viewpoint.
We build a large-ish project on both Solaris/SPARC and Windows x86.
Event though the Windows build machine is a higher spec [1] than the Solaris box, and only builds a sub-set of the UNIX product...
* On Solaris/SPARC the system build quicker, and builds in the background while other users can happily work uninterrupted (including multiple builds). This box is about 7 years old now, and has a rather dated toolchain.
* On Windows/x86 the build takes longer, much longer, on a dedicated build machine which cannot be realistically used for anything else in the meantime. This box is about a year old.
For serious jobs I still find that the SPARC/RISC architecture is far more usable than x86. Haven't tried x64 since the system is 32-bit and I really don't want to be responsible for porting to 64...
[1] by simple processor/core count & speed, memory and disks.
Commonly held misapprehension, not really true. When NT started life it had more in common with a DEC distributed realtime kernel called VAXELN (which had VMS as a development host, and some limited VMS compatibility) than it had in common with VMS itself. Source: Helen Custer, Inside Windows NT (and various other published and unpublished sources inside DEC and MS).
As time went by Gates forced more and more NT OS code into kernel mode for "performance" reasons, and thus introduced all kinds of compromises to the original OS's security and stability.
Today's NT may have good device support and lots of shiny features but in comparison with the original NT security and stability model, it sucks.
NT wasn't technically a child of OS/2 although commercially it may have looked that way at one point, and at one time there was an OS/2 compatibility subsystem in NT.
You've presumably never heard about Lasermoon, provider of the world's first POSIX compliant LINUX, maybe 15 years ago or more. Unfortunately all the PHBs that used to bleat back then about the importance of "open systems" actually wanted Windows because they thought it was cheap, and Lasermoon went out of business.
There is some Beavis and Butthead traffic going on that references a 2003 IBM document where IBM expects Linux to eventually replace all enterprise UNIX (it didn't just talk about AIX, no matter what anyone says). IBM said it in 2003. They did not say how-fast it would come.
IBM Power 7 has taken a different product-path than Intel IA32 x64. Intel IA32 (~1980) is actually a OLDER than IBM Power (~1989). IA32 has seen major changes from its 16-bit bus origins to its x64 existence today. Power was RISC and 32-bit from the start, but IBM wasn't aggressive with advancing its technology in the 1990s. Since 2001, IBM has been much more active in getting its 64-bit architecture and higher-performance aspects rolling.
#1 Power is more expensive then x64. I agree. As soon as you start talking about 8 cores-or-more, it isn't 3x. A single-socket 8-core x64 blade w/ Xeon 7500 is about $6K. A similar 8-core blade w/Power 7 is about $8K. Both can run Linux. The Power blade benchmarks out 15-20% faster. The Power blade even goes into the same enclosure and an IBM Xeon blade, if that matters.
#2 Power virtualization hardware is quite competitive with VMWare. CPU overhead is lower on PowerVM and cross-client communication is better PowerVM. VMWare handles Windows and that is just a non-starter for PowerVM. VMWare also adapts Windows to virtualization in really nice ways. If you have Windows, then VMWare must win. If you have Linux or AIX, then PowerVM provides substantial management and resource sharing that matters in a sufficiently large shop.
#3 Power 7 still has the giant-server business locked down, but at a substantial markup. This is where the 3x cost multiplier comes into play. If the application gets kludgey trying to run on x64 Linux, then running on Power 7 can be attractive (even at its cost premium)
#4 Power 7 is nearing the end of its current server-cycle. Power 7+ is due out in the Fall. IBM sells servers, not chips, with this product. A lot of bloggers find IBM's not-as-public ways off-putting. They are off-putting. That doesn't invalidate what IBM is doing, though.
... I guess that I have said enough for now...
I post links to IBM senior executives. Would you prefer me to not post links at all, so you could dismiss my posts as pure FUD and rumours? Well, IBM executives has said what I claim, I am only quoting them.
POWER6 was several times faster than x86 but costed 5-10x more.
POWER7 is 10% faster than Intel Westmere-EX and costs 3x more.
What is your predictions about future POWER cpus? What does the trend tell you? That IBM will continue to lower the price of POWER? That POWER is soon not a high margin business anymore? And you know what IBM does with low margin business? IBM kills it. As any sane capitalist company with greedy share holders would do. IBM is an american company. IBM gets rid of low margin business, when IBM spends billions on R&D and can not get the expends back. x86 is the future. (Even though x86 is a buggy piece of shit, it is cheaper and soon faster)
You aren't making much sense...
Again your claim that POWER is 3 times more expensive is a very broad and so absolutely not always, if ever, the case. Look at my previous post or try to look up some figures up yourself.
Sure a POWER 780 is an expensive machine, but it's also a fast machine. For example have a look at this link to the best price/performance TPC-C benchmarks
http://www.tpc.org/tpcc/results/tpcc_results.asp?print=false&orderby=priceperf&sortby=asc
Then you'll find that the POWER 780 the 'higher end' version of the POWER 770 8 socket box, comes in at 13th place, with the best x86 windows result that doesn't use an enterprise database, being only 29% cheaper, even though the POWER 780 is basically a highend machine, with dual service processors, being able to hot repair of nodes and and ...
So yes POWER and other UNIX'es are more expensive than x86.. but it's not like it was in the old days. Now it's two digit %'es not a factor that is greater than 2.
You are saying that POWER is to expensive.. then you say IBM is killing off POWER development while POWER is having the best makedshare that it have ever had ?
You are not making sense.
// Jesper
The question is not the numbers, never mind them. It does not matter if it is 10% or 13%.
The question is: POWER is getting cheaper and cheaper. And Intel is getting faster and faster. Soon, POWER will be very cheap, to be able to compete with Intel, and that is when POWER will stop being high margin. And why would IBM do low margin business, when POWER costs billions in R&D? Why should IBM not kill a low margin business when you get a off the shelf Intel CPU at a lower price, and higher performance?
@Kebabbert I never said anything about FUD, so you aren't reading what I wrote. I did imply that the 2003 document was a long term prediction and that citing is as a fait accompli is superficial.
If you keep it in the future tense, then it isn't superficial. Your choice.
Your 10% number has been challenged with citations several times. It really is 15-20%. I infer from Jesper that giant servers (32+ cores) favor Power much more. He has examples.
OPINION
Power is behind x64 in manufacturing process (INTC has 22 micron NOW), but ahead of x64 in GHz (IBM has better heat control). IBM will get to 12+ cores per socket this fall. x64 has been battling with its more-complex ISA because it makes it harder to layout tighter/faster logic.
IF you believe that x64's ISA will ultimately hog-tie improvements in 1-4 years, then you would favor Power (some do). IF you believe that INTC will 'overcome', then it is hard to see Power being viable in 7-10 years. INTC did get dragged kicking-and-screaming (by AMD) into doing that much needed x64 function-enhancement. INTC is also battling ARM for low-end Windows sales (ARM already is strong in handheld smartdevices). INTC's reputation for 'vision' and 'intensity' are key elements for success in this challenging time.
IBM doesn't want to put Power in the commodity server business, so your warnings about IBM's appetite for low-margin sales are about-right. The hyperbole about them getting scared off the moment the bloggers find a disadvantage is not right.
FWIW, my guess is that INTC won't find an easy way out of the heat dissipation issue with x64. In the worst-case scenario, they will go massively parallel (>16 cores/socket) on highly complex chips for large servers in the next 3 years. This worst case would keep the door open to Power on non-Windows environments. INTC also faces ARM on Windows, now that MSFT has opened the door on that. INTC is a very resourceful company, though. I don't count them out to find an imaginative approach.
FWIW, my guess is that IBM will only push Power enough to maintain control of the high-end market. They probably can push GHz up faster than INTC, because of their heat-dissipation advantage. Power isn't going to break-out past the Linux/AIX market though (that was easy), so unless Linux makes major inroads on Windows, Power is playing in a facet of the marketplace.
Great discussion going on here - all driven by hardcore IT experts with their respective OS preferences and a deep desire for even higher clockspeeds, combined with a remarkable pennypinching mentality ...
Anybody asking about the concerns of those stupid business users - those who just understand their business but have no clue about the latest trends in IT ? Are they really screaming about more bang for the buck, sophisticated virtualization and cloud computing ... or wouldn't they come up with other priorities if they only were asked ?
Things like I don't need even more complexity and I can do without the latest fancy stuff. Rather, I want my application to run reliable, I don't want my database to get corrupted, I don't want to get hacked, and I don't like hunting for even more of those expensive experts in pursuit of those rather simple goals.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
