Slashdot Mirror

← Back to Stories (view on slashdot.org)

IBM and OpenPower Could Mean a Fight With Intel For Chinese Server Market

Posted by timothy on from the round-the-mulberry-bust dept.

itwbennett writes With AMD's fade out from the server market and the rapid decline of RISC systems, Intel has stood atop the server market all by itself. But now IBM, through its OpenPOWER Foundation, could give Intel and its server OEMs a real fight in China, which is a massive server market. As the investor group Motley Fool notes, OpenPOWER is a threat to Intel in the Chinese server market because the government has been actively pushing homegrown solutions over foreign technology, and many of the Foundation members, like Tyan, are from China.

49 of 85 comments (clear)

  1. Tell me about POWER and IBM. by Anonymous Coward · · Score: 1

    Although I've had a long career, I've never had the chance to work with IBM's technology (I've mostly worked in Sun, HP and Linux shops).

    I'm sure that a lot of people here have worked with IBM's products before. I want you to tell me what they're like.

    What are POWER systems actually like to work with? Are they obviously better than the processors and hardware from other vendors?

    What is AIX like to work with?

    What is DB2 like to work with?

    What is Informix like to work with?

    What is Lotus like to work with?

    What is WebSphere like to work with?

    What is Tivoli like to work with?

    1. Re: Tell me about POWER and IBM. by belrick · · Score: 3, Informative

      I've been working with AIX since 1990. Prior to that a bit of SunOS. AIX is is different but generally well thought out. Most people who hate it simply aren't used to the differences. Lots of feature that we take for granted in today's Linux existed in AIX 25 years ago.

      Tivoli Storage Manager is a dream. I remember setting up a high-availability TSM (well, ADSM at the time) server and having a client backup running during fail over testing. Client connection failed, continued retrying until the server was back up on the other node, then the backup continued where is left off. Transaction backup with rollback and resumption after server fail over! Try that with NetBackup or Networker or Avamar or CommVault.

      B

    2. Re:Tell me about POWER and IBM. by tagous · · Score: 2

      All Ubuntu packages run on it with 14.04. SUSE and Redhat also have LE distributions that run on OpenPOWER systems too Take it for a spin... http://osuosl.org/services/pow...

    3. Re: Tell me about POWER and IBM. by lsatenstein · · Score: 1

      I've been working with AIX since 1990. Prior to that a bit of SunOS. AIX is is different but generally well thought out. Most people who hate it simply aren't used to the differences. Lots of feature that we take for granted in today's Linux existed in AIX 25 years ago.

      Tivoli Storage Manager is a dream. I remember setting up a high-availability TSM (well, ADSM at the time) server and having a client backup running during fail over testing. Client connection failed, continued retrying until the server was back up on the other node, then the backup continued where is left off. Transaction backup with rollback and resumption after server fail over! Try that with NetBackup or Networker or Avamar or CommVault.

      B

      And I loved VSAM, which was OS2 and which was ported to AIX. That was done in late 1980s.

      --
      Leslie Satenstein Montreal Quebec Canada
  2. We need more architectures by Anonymous Coward · · Score: 1

    OpenPOWER looks nice, but we had this OpenSPARC thing for ages and it hasn't really taken off. Somebody "liberate" Alpha, and while at it, PA-RISC, and let's build something new. We need more diversity in the datacentre and on the desktop.

    1. Re:We need more architectures by Anonymous Coward · · Score: 1

      "Alpha and PA-RISC architectures died, and why SPARC and POWER and MIPS are on their way out"

      The real reason was that they didn't perform as well as, or much better than, Intel so it wasn't worth the hassle to switch to an architecture that has a smaller software base. If one of those architectures was 100% better it might be worth it, but only 5-10% better isn't.

    2. Re:We need more architectures by ChunderDownunder · · Score: 2

      MIPS ain't extinct.

      They're about to enter the tablet market with a dual core SoC featuring a blazingly fast PowerVR GPU.

      Performance looked a tad sluggizh but that's possibly Spidermonkey needing fine tuning for MIPS

    3. Re:We need more architectures by Darinbob · · Score: 2

      Right, Intel x86 family doesn't win because it's better, but because it's a hassle to switch. Intel is good enough, nothing spectacular, no one will study x86 as a good example in CPU design courses, though I'm sure it will be taught by the faculty at business schools. MIPS and Sparc however will continue to be taught, PowerPC is still a better overall system design in every way. Even Intel is unable to climb out of the pit they are in with backwards compatibility, their own i860, i960, and Itanium chips failed not because they weren't superior designs but because they didn't have the backwards compatibility with a shitty design.

      McDonald's is the number one restaurant in the world, but you don't hear the culinary world raving about how great they are. Why then does the x86 family maintain a set of fan boys?

    4. Re: We need more architectures by ChunderDownunder · · Score: 1

      Itanic was supposed to succeed both Alpha and i386 but then AMD came along with amd64?

    5. Re:We need more architectures by Dog-Cow · · Score: 1

      I believe you are thoroughly outdated in your knowledge of Intel architecture. At the hardware level, the only vestige of compatibility is the instruction decoder. Once you get past that, an Intel chip looks much like any other architecture in terms of features and composition.

      (Yes, even x64 has fewer registers than some other architectures, but that's a design choice, not an architectural limit.)

  3. Re:RISC by Anonymous Coward · · Score: 1

    Your phone uses a RISC processor. That happened and it changed everything.

    At the time in the mid-1990s, Apple and IBM were promising to change the world, by running Mac OS and Windows NT on the same PowerPC hardware. It only took ten more years, until Apple altered reality by switching to Intel, and finally Windows XP and Mac OS X ran on the same x86 hardware.

    Today, Apple is designing their own processors again, and they're ARM-based RISC.

  4. "Ditchin' the pSeries down here, boss" by PolygamousRanchKid+ · · Score: 2

    So, is IBM going to ditch making their own POWER pSeries, and totally go for the ARM model of just licensing the technology for OpenPOWER . . . ?

    Just like in the PC world, folks stopped buying IBM built PCs, when cheap clones were available. What would be the advantage of buying an IBM built OpenPOWER system, as opposed to a much cheaper Chinese built clone . . . ? Maybe the IBM system will have some kind of "secret sauce" . . . ? Like a MicroChannel (har, har).

    At any rate, somebody is going to have to invest a lot of money to make sure that Linux runs well on OpenPOWER, in order for this to succeed.

    --
    Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
    1. Re:"Ditchin' the pSeries down here, boss" by Darinbob · · Score: 1

      Linux already runs well on POWER.

  5. Power is bigger than you think by nhtshot · · Score: 5, Informative

    I work on PPC systems every day. I also use several. I'd wager that you do as well.

    Have cable or satellite TV? 90% chance it's using a Power cpu. Drive a car with fuel injection? 65% chance your engine is run by Power, 90% chance something in the car is (ABS, nav, transmission).

    It's been around a long time (30+ years), been 64 bit much longer than x86 or ARM, has good OS support and good compilers.

    I work on and like ARM as well, but if IBM can make a value proposition in China with PPC, they actually have a chance at getting some market share outside embedded.

    1. Re:Power is bigger than you think by Blaskowicz · · Score: 1

      I looked OS/2 for PowerPC up and it indeed ran Windows 3.1.. Wow!
      If you wanted to run Windows 3.1 applications (because there's no OS/2 software), why not get a 486 with half the RAM instead. That's easier and cheaper.

    2. Re:Power is bigger than you think by lister+king+of+smeg · · Score: 1

      Have cable or satellite TV? 90% chance it's using a Power cpu.

      actually mine is has MIPS chip

      --
      ---Saying gnome 3 is better than windows 8 not so much a compliment as it is damning with light praise.
    3. Re:Power is bigger than you think by lsatenstein · · Score: 1

      I work on PPC systems every day. I also use several. I'd wager that you do as well.

      Have cable or satellite TV? 90% chance it's using a Power cpu. Drive a car with fuel injection? 65% chance your engine is run by Power, 90% chance something in the car is (ABS, nav, transmission).

      It's been around a long time (30+ years), been 64 bit much longer than x86 or ARM, has good OS support and good compilers.

      I work on and like ARM as well, but if IBM can make a value proposition in China with PPC, they actually have a chance at getting some market share outside embedded.

      I'm a fan of Power, and even big-Endian. Why not come out with 96bit wide processors. 32gigs ram for home and small business computers is too small.

      --
      Leslie Satenstein Montreal Quebec Canada
  6. Won't everything need to be recompiled? by enjar · · Score: 1

    Sure, you can put out a chip, but without a software stack of common applications (and operating systems) that you actually run on that stack, it's just something that consumes electricity.

    So who is going to fund the porting effort of all the tools, libraries, etc? Anyone who thinks you just grab source code and recompile on a new platform has probably never tried it. It's a pile of work.

    1. Re:Won't everything need to be recompiled? by Anonymous Coward · · Score: 1

      It's been done for Windows (Windows RT and Windows 10), OS X (ran on PowerPC a while back now on Intel), Linux (which runs on almost all CPUs) and many Unix systems. Might be a lot of work but seems to me, just recompiling for OpenPower isn't going to be THAT hard.

    2. Re:Won't everything need to be recompiled? by PolygamousRanchKid+ · · Score: 1

      A simple recompile won't do it . . . some programs, like your TCP/IP stack have dependencies on Endianess. IBM's POWER has been traditionally Big Endian. Linux is mostly Little Endian. There are C macros, ntoh() and hton() that do the required byte swapping for you . . . if you remember to use them! I have seen code that would run fine on a Big Endian machine . . . but would fall over and die on a Little Endian machine.

      Sorting out all these problems is painful grunt work. Although, at one point, IBM tried to port AIX to Intel64. It was called Monterrey, or something like that. So IBM does have experience in doing this. It's just a question if they will fork out the dollars required to do this. And if the people know how to do this have been laid off.

      --
      Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
    3. Re:Won't everything need to be recompiled? by Florian+Weimer · · Score: 1

      Future efforts are likely to concentrate on the ppc64le architecture variant, which is little endian. There are still some differences to x86_64 at the C level (chars are unsigned by default, but you can compile with -fsigned-char), but it is reportedly not too difficult to port over C/C++ application code.

    4. Re:Won't everything need to be recompiled? by enjar · · Score: 2

      Yep. In my career, I've seen the rise and fall of RISC (on both Windows and *NIX), Apple's transition between several chip families, Sun's Sparc chips and even Intel trying to out-Intel with Itanium. You get hit with major roadblocks as well as death by a thousand cuts. It's extremely difficult to get it working in the first place, and then ongoing maintenance is no small feat, either.

      I wonder if the Chinese government is "strongly favoring home grown solutions" with an ongoing infusion of funding, to do they just pay it lip service? China is a huge emerging market that plenty of vendors are trying to sell into, if they are really serious about this, it could actually provide the catalyst to make the ports happen. But no demand in the marketplace means little incentive for anything to happen.

    5. Re:Won't everything need to be recompiled? by Hollinger · · Score: 2

      Actually, POWER8 supports both big and little endian, and you can go out to Canonical's site and (as of 14.04) and get the LE version of Ubuntu for POWER8. You can read about that below. Quoting the article:

      Why is Linux on Power transitioning from big endian to little endian?

      The Power architecture is bi-endian in that it supports accessing data in both little endian and big endian modes. Although Power already has Linux distributions and supporting applications that run in big endian mode, the Linux application ecosystem for x86 platforms is much larger and Linux on x86 uses little endian mode. Numerous clients, software partners, and IBM’s own software developers have told us that porting their software to Power becomes simpler if the Linux environment on Power supports little endian mode, more closely matching the environment provided by Linux on x86. This new level of support will lower the barrier to entry for porting Linux on x86 software to Linux on Power.

      - https://www.ibm.com/developerw...

      --
      Michael C. Hollinger
    6. Re:Won't everything need to be recompiled? by Henriok · · Score: 1

      The endianess problem is a nonissue with OpenPOWER since it's little endian, just like x86. A _very_ large portion of all open source software _will_ just work with a recompile, even if the project hasn't touched Power Architecture before. So, the money is already forked out, and it's done.

      --

      - Henrik

      - when the Shadows descend -
    7. Re:Won't everything need to be recompiled? by Guy+Harris · · Score: 1

      A simple recompile won't do it . . . some programs, like your TCP/IP stack have dependencies on Endianess. IBM's POWER has been traditionally Big Endian. Linux is mostly Little Endian. There are C macros, ntoh() and hton() that do the required byte swapping for you . . . if you remember to use them!

      The TCP/IP stack you're using was probably originally written by people working on a little-endian machine - VAX (if it's the BSD stack or a derivative thereof) or x86 (if it's the Linux stack) - so that's not the code to worry about; it has the relevant ntoh[sl]()/hton[sl]() calls already.

      It's your own code you'd mainly have to worry about. I.e., Linux should pretty much Just Work (it runs on 32-bit and 64-bit PowerPC, and it sounds as if support for little-endian mode is being added), but it's the third-party software that's might be an issue.

    8. Re:Won't everything need to be recompiled? by Darinbob · · Score: 1

      Any TCP/IP stack with a dependency on endianness is fundamentally broken. Seriously. Only a moron would forget to use ntohl and htonl, and only a moron product manager would allow such a stack to be sold. Every commercial vendor an embedded TCP/IP stack has ported to a variety of architectures. Even in the unlikely chance that there's a vendor of TCP/IP that has a endianness problem then you can always buy from someone else.

      But most likely the POWER users are going to use BSD or Linux anyway.

    9. Re:Won't everything need to be recompiled? by PolygamousRanchKid+ · · Score: 1

      It's your own code you'd mainly have to worry about.

      But, alas, I always seem to get called in to debug code from other folks.

      --
      Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
    10. Re:Won't everything need to be recompiled? by PolygamousRanchKid+ · · Score: 1

      Only a moron would forget to use ntohl and htonl, and only a moron product manager would allow such a stack to be sold.

      Unfortunately, if there is one thing that the IT Industry does not lack today . . . is a shortage of morons.

      --
      Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
    11. Re:Won't everything need to be recompiled? by PolygamousRanchKid+ · · Score: 2

      AIX was pretty cool way back when, when they introduced 64-bit support. The processor was 64-bit. However, you could run a 32-bit kernel or a 64-bit kernel. And you could run a 32-bit process or a 64-bit process on either of the kernels.

      So what does some poor chump (i.e. me) who is tasked with writing a device driver for AIX need to do? Well, first #ifdef the code, so you compile different stuff, depending on if you are building a 32-bit or 64-bit version of the device driver. Then you needed to add simple "if" statements in the device driver, to check if you were running a process in 32-bit or 64-bit mode. Then according to the mix, you would have to thunk the addresses, when copying the data from user space into kernel space.

      Fun stuff. I can't believe that I actually did this in a former life . . .

      --
      Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
    12. Re:Won't everything need to be recompiled? by PolygamousRanchKid+ · · Score: 1

      Yeah, but Linux on POWER today runs on Big Endian. See another post in this thread about IBM intentions, but Red Hat has not announced support for Linux on Little Endian yet. That one hurts.

      Linux on OpenPOWER doesn't exist yet . . . or does it . . . ?

      --
      Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
    13. Re:Won't everything need to be recompiled? by PolygamousRanchKid+ · · Score: 1

      AIX is full of "hacks" or "modifications" in the TCP/IP stack to greatly improve the performance on POWER architecture on MP systems. Have any of these made it into mainstream Linux? Are they even valid on Intel architecture?

      For instance, when running a benchmark on an AIX POWER system, try increasing the load, and see if your results go up. It can happen, that you increase the load, the CPU utilization climbs, but you benchmark remains the same. Well, you might be hanging in spin locks. AIX supports instrumented locks, so you can check this with the lockstat command.

      Another potential problem is that two many global variables are located in the same CPU cache line. So you can pad single variables, so that they are in separate cache lines. Or, even worse, you have one global variable that is being constantly updated by all processors, and is constantly causing cache invalidation on the memory bus. Then you need to do a hardware memory bus trace, with an HP logic analyzer that looks like something out of Hentai Porn. Then you need to write up a patent or something:

      https://patents.justia.com/patent/6430659

      So I'm just wondering if all this poop will be done for Linux on OpenPOWER . . .

      --
      Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
    14. Re:Won't everything need to be recompiled? by Guy+Harris · · Score: 1

      It's your own code you'd mainly have to worry about.

      But, alas, I always seem to get called in to debug code from other folks.

      You probably don't get called in to debug endianness issues in any TCP stacks in portable general-purpose non-hobby OSes; if so, I'd look a bit askance at the developer team for that OS.

      My point is that worrying about porting server OSes for PPC machines is not something worth worrying about, as they've either already been ported (Linux) or, if they ever get ported, are likely only to be ported to little-endian machines using compilers that aren't going to differ between platforms on the signedness of char (Windows). What you worry about there is third-party applications.

    15. Re:Won't everything need to be recompiled? by Henriok · · Score: 1

      Little Endian Linux for OpenPOWER exists, and have for some time. SuSE (SLES 12), Debian (kernel v.3.13 and 94% of the software repository), Canonical (Ubuntu Server 14.04) and RedHat (RHEL 7 and RHEV) have distros ready to go. Canonical is a platinum level member of OpenPOWER Foundation and the poster child for compatibility, and Shuttleworth is traveling the world showcasing this.

      --

      - Henrik

      - when the Shadows descend -
  7. Mill Architecture by Anonymous Coward · · Score: 2, Interesting

    The Mill architecture is around the corner now, and promises immense potential. It elegantly addresses many deficiencies of conventional architectures, and enables substantially increased efficiency while also simplifying system software and compilers. It is a fascinating and compelling design, which re-abstracts the hardware and software in a fundamentally superior way.

    While the Alpha is a nice RISC design, at heart it is more similar to an x86 than not. The paradigm introduced by the Mill architecture is a world apart.

    1. Re:Mill Architecture by bdcrazy · · Score: 1

      While also simplifying system software and compilers.

      Every few years we hear about this. What makes this attempt any more likely than the previous attempt? We have to deal with the software we have now, not what we'd like it to be.

      --
      Tonights forecast: Dark. Continued dark throughout most of the evening, with some widely-scattered light towards morning
  8. RISC-V seems nice by Anonymous Coward · · Score: 1

    RISC-V is an open source architecture. It is royalty-free and very modern.

  9. Tyan is from Taiwan. by shess · · Score: 1

    Don't get me wrong, they'll be as happy to sell into China as into the US, but if anything China seems likely to trust their hardware _less_.

    1. Re:Tyan is from Taiwan. by Henriok · · Score: 1

      That's why China has set up an own consortium based on OpenPOWER, the China Power Technology Alliance, CPTA. They are building a purely chinese OpenPOWER ecosystem, with all aspects of hardware and software.. like the CP1, a POWER8 clone with a Chinese crypto engine (since they didn't' want the american version, and wasn't allowed anyway).

      --

      - Henrik

      - when the Shadows descend -
  10. Re:Why ? by Henriok · · Score: 1

    Why OpenPOWER as a separate entity from Power.org? I think it's because Freescale has all but quit developing Power Architecture. So there's essentially only IBM left doing active development of Power Architecture. And, I think that Freescale and IBM really have different goals for the future. Freescale is aiming at low performance (compared to POWER8) embedded systems, where ARM is gaining more and more ground. IBM isn't interested in going in that direction, and saw an opportunity to write a new chapter with POWER8 and forward, being able to ignore and break backwards compatibility with the legacy of Power Architecture. IBM isn't making money selling low margin hardware, they are in the business selling high margin technology and services. It probably won't matter to them if you in the future buys a OpenPOWER box from some white-box OEM vendor i Taiwan, with an Chinese designed OpenPOWER processor, fabbed by TSMC.. if they can charge you for using their applications, services and consultancy hours.

    --

    - Henrik

    - when the Shadows descend -
  11. AMD Fade Out? by Anonymous Coward · · Score: 2, Interesting

    Uh what?

    Last I heard AMD was going balls to the wall with an ARM server chip and 'Zen' server cores in Q1 2016.
    Come to think of it, outside of GPUs, server chips are the only thing I've heard of that AMD's working on down the road.

    1. Re:AMD Fade Out? by serviscope_minor · · Score: 2

      Was thinking the same. Plus a fully loaded 1U quad socket AMD server is pretty cheap now as those things go.

      --
      SJW n. One who posts facts.
  12. Tht elephant in the room by Okian+Warrior · · Score: 1

    The elephant in the room, of course, is security.

    With NSA "upgrade factories" - where spyware is installed by the NSA before delivery - China and everyone else is looking for alternatives to American products.

    (And note that the spyware can be implanted in the BIOS, and even the hard drive firmware, and will persist even if the system is wiped, or the BIOS is replaced.)

    The scope of economic damage this has done is astonishing. I've never believed in trickle-down economics, but once China starts making servers my guess is our IT industry will tank from the top down.

    Expect an economic crisis in, oh... about 5 years.

    (The solution would appear to be a complete open-source ecosystem including BIOS and hard drive firmware. Just as I can verify my linux installation, there should be verifiable BIOS and hard drive firmware, so that any country can purchase any computer, and be confident of its security.)

    1. Re:Tht elephant in the room by tagous · · Score: 1

      OpenPOWER is Open. Here's the source code for the OpenPOWER BIOS. Enjoy... https://github.com/open-power

  13. Re:Keyword: *SOFTWARE* by Bert64 · · Score: 4, Informative

    The POWER architecture has been around longer than X64, the vast majority of linux software comes with source code and compiles fine on power (and arm, mips and anything else) so it doesn't matter what the underlying processor is. A lot of the software that doesn't come with source these days is java based, which will run just fine on power too.

    Except for a small number of fairly niche apps, most linux based server loads will work fine on a power system.

    --
    http://spamdecoy.net - free throwaway anonymous email - avoid spam!
  14. turning an aircraft carrier by globaljustin · · Score: 1

    if IBM can make a value proposition in China with PPC, they actually have a chance at getting some market share outside embedded.

    yes i agree

    from TFA summary: "because the government has been actively pushing homegrown solutions over foreign technology, "

    China is serious about this.

    They are wise to the level of embedded spyware and also the way companies will lock you into proprietary everything.

    Also, it's a wise move from an IT perspective. Especially for something as huge as China, pushing "homegrown solutions" on that economy of scale is a major change and it will have a noticalbly positive effect.

    "Like turning an Aircraft Carrier"...that's China's IT infrastructure...it's so huge, it takes a long time to turn, so your criteria for deciding to change course is different...also the consequences of when you do change course are orders of magnitude different

    --
    Thank you Dave Raggett
  15. Re:Power or Freescale? by Darinbob · · Score: 1

    I've never used a 68K embedded system. Those don't really have a good niche is the problem. 32-bit SoCs meant for high performance typically use PowerPC, and 32-bit SoCs meant for lower power or economy typically use ARM. My experience only. Sure it may have big numbers of sales, mostly with ColdFire in the automotive market.

  16. Will I be able to buy one for my desktop? by Required+Snark · · Score: 1
    It would be nice if Intel had some competition to keep them honest.

    AMD used to fill that slot, but they don't count for much any more. So far Arm is not much of a player outside of tablets/smartphones.

    I want meaningful choices.

    --
    Why is Snark Required?
  17. Re:Keyword: *SOFTWARE* by lsatenstein · · Score: 1

    The POWER architecture has been around longer than X64, the vast majority of linux software comes with source code and compiles fine on power (and arm, mips and anything else) so it doesn't matter what the underlying processor is. A lot of the software that doesn't come with source these days is java based, which will run just fine on power too.

    Except for a small number of fairly niche apps, most linux based server loads will work fine on a power system.

    I wonder if IBM would produce a Power cpu for the desktop at less than Intel I7 pricing. I would not mind if the chip was made in China.
    If they do, hopefully it will be a 96bit version, with programmable little Endien/Big Endien mode.

    --
    Leslie Satenstein Montreal Quebec Canada
  18. Re:Keyword: *SOFTWARE* by rdnetto · · Score: 1

    I wonder if IBM would produce a Power cpu for the desktop at less than Intel I7 pricing. I would not mind if the chip was made in China.
    If they do, hopefully it will be a 96bit version, with programmable little Endien/Big Endien mode.

    Why on earth would you want a 96-bit CPU? Even the current 64-bit ones can 'only' address 48-bits of memory (i.e. 281 TB).
    And if you want it for a certain computationally expensive load, 128-bit would make more sense (or just doing the computation over 2 -bit words).

    --
    Most human behaviour can be explained in terms of identity.