AIX On the Desktop Is Getting the Boot 366
flnca writes "Today, I was playing with the thought again to purchase an AIX workstation one day when I can afford them, and I was surprised to see that IBM is going to give its IntelliStation POWER Series workstations the boot in January '09. A black day for AIX on the desktop. I really wonder what's the problem there, warehouse costs? IBM has a history of burying its best stuff (like OS/2 for instance). Some years ago, I enjoyed hacking away on an RS/6000 workstation running AIX 4.2, and it was a pure joy. Not only the kernel, but also the admin tools, like smit and smitty. Their blade-centric solution uses Windows as a client for workstation application. This truly sounds like IBM wants AIX only for servers anymore. I'm not amused. Although, eXceed on Windows with an XDCMP server running on AIX might also be a viable solution ... whatever. But it can't beat a native POWER box sitting on your desk, that's for sure."
My guess. (Score:3, Interesting)
The new I7 and maybe the new 45 nm AMD cpus are probably a better solution for a workstation then a Power these days. Linux has more hardware and software support than AIX so IBM probably sees the future as an I7 running Linux.
Re:Another victim of Linux... (Score:4, Interesting)
Newer versions of *nix are killing older versions of *nix.
The exact opposite of what's happening with Windows.
A Huge Blow (Score:5, Interesting)
This is a huge blow to scientific and engineering computing. I know of thousands of POWER based Intellistations at several aerospace companies. CAD and finite element analysis software runs on these boxes, usually CATIA, NASTRAN, and some CFD codes. Engineering modeling and simulation software has been running on AIX for a while. Only now are Windows boxes near the performance that engineers need. The only good that might come of this is that hopefully the surplus market will be flooded with POWER based Intellistations and AIX CDs.
Wait? (Score:3, Interesting)
Are you saying using smit and smitty was a pure joy?
Bwahahaha!
AIX is an antique (Score:5, Interesting)
Other PowerPC options available (Score:4, Interesting)
YellowDog makes a PowerPC based Linux machine. The latest Linux Journal has a review of it:
http://us.fixstars.com/products/powerstation/ [fixstars.com]
http://www.linuxjournal.com/article/10263 [linuxjournal.com]
Not perfect, but workable.
Re:It's not just you (Score:3, Interesting)
Forgive me, because I'm a linux guy (that's all I've ever used and known), and because of that, I don't know what the benefits of having an AIX machine on the desktop would be.
I understand that on certain large hardware, AIX is preferable due to hardware or other requirements, but what is the draw on the desktop? Is there superior software, or stability? Management tools?
I manage Linux servers, and I have linux on my desktop because it seems effortless to me now, but I can't imagine that if I had one of the BSDs, Solaris, or any other unix that my experience would be different. What is the draw? I'm not flaming or trolling, I'm really interested in being educated.
Re:Don't be silly (Score:3, Interesting)
Re:It's not just you (Score:3, Interesting)
The desktop was under $10k, sits at my desk, and is mine to do what I want with it. Currently I'm testing AIX 6.1 (works great, cool new features). It'll run KDE and an ancient version of firefox, if I want, usually I just have X with multiple shells open.
Whenever I need to do something particularly major in our prod environment, it's fully vetted on this desktop first. OS upgrades, patches, Oracle upgrades, firmware, new utility scripts - I have a great little test environment for them. And alt-disk-install makes it a snap to get back to 'normal'
Do I see any use for one outside of that? Not really, except maybe 3d rendering or something.
- Tony
Re:It's a true desaster. (Score:2, Interesting)
Re:Go ahead and suck it up. (Score:2, Interesting)
While it is possible to install GNU utils and get an AIX box somewhat up to speed, what I've seen is that AIX admins/users are still stuck in the dark ages, completely unaware of FOSS alternatives and best practices. (They ONLY use telnet at my workplace)
It is the "easiest" to get things configured using smitty, which is geared towards a person who does not care to know anything about the OS underbelly. Smitty is intimately integrated with AIX, such that it is far more difficult to administer multiple machines because of smittys and AIX weird nature.
The only reason IBM can make so much money from AIX is because of clueless managers and "old" users who never adapt.
It is well known that the reason P6 chips run at 5Ghz, is because they are totally uncompetitive against Intel/AMD at lower speeds. They were delayed because IBM had a tough time getting 5Ghz yeilds and now it leaves them less headroom, unless they can get to 6 and 7Ghz quickly.
These days, many FOSS tools are available as binaries for AIX (but not latest versions).There have been several occasions where I have had to compile stuff and have simply given up because the FOSS author has not updated his build for AIX.
We just threw away several 300-700Mhz machines that were originally bought for 30K+. Why not buy good x86 stuff for 8K and throw that away later?
Re:It's not just you (Score:2, Interesting)
From someone a little bit more objective: I've managed AIX, Linux, HP-UX, Solaris, IRIX, Windows, SCO and *BSD boxes for many, many years.
Out of all of them -- Linux and AIX have by far the best command line and GUI management tools for enterprise computing. Smit and Smitty are at least as good (and in some ways better than) SuSE's YaST or Redhat's system-config-*. (HP guys, shut up. Smit/smitty blows SAM out of the water and you know it. Go back to the hole you climbed out of. ;) For example, I have yet to find a GUI LVM tool that can do what smit can do as well as smit can do it.
If you were managing an all-AIX environment, you'd definitely want an AIX machine on your desk, not a Linux machine.
OTOH, if you're handy with scripting and whatnot, nothing beats a Linux machine in a mixed envrionment. ;)
Re:Somehow this remembers me 1995 (Score:3, Interesting)
Most of the classic 8-bit CPUs have variable-length instruction coding, multiple-clock instructions controlled by state machines, and support indexed memory addressing modes and memory writes on general operations (instead of dedicated load/store instructions). All of these are hallmarks of a CISC architecture.
The 6502 may have had a few RISC like features if you pretend that the first 256 bytes of memory are registers, but it still doesn't really qualify as genuine RISC.
And the surprise is it's taken this long... (Score:1, Interesting)
In late 2000, there were rumblings of getting rid of the desktop RS/6Ks.
In 2003, the rush was to grab any chip design tool that ran on Linux instead of AIX because an extra PC intended to be used by an administrative assistant could run the chip simulations and layout faster than any pSeries hardware we could afford in the budget.
It's now 2008, corporate policy now dictates that if it even remotely resembles a remotely accessible server it shall be locked up in a controlled area. We could kludge together network KVMs to a pile of AIX desktops in the server room, or we can yank the video card from a 550 and replace it with a fibre channel adapter and break it into several LPARs for the developers and tell them "Have fun using ssh!"
I enjoyed having an AIX box as a primary or secondary workstation for about 8 years. I enjoyed a few periods where the video adapter in my AIX box blew away anything readily available for the PC. But now that it's 2008, the AIX box at my desk is reached via ssh. The last time I had to reinstall it, the reinstall was done from within minicom. The graphical programs I use run on the corporate approved Linux desktop, not on AIX.
Then there's the minor matter that linux KVM for PPC turns non-lparable Power hardware into a direct competitor for the lparable hardware...
Re:The march towards Linux (Score:4, Interesting)
AIX still has significant advantages over Linux for us. A lot (all?) of the stuff that is new in AIX has come from the AS/400 390 mainframe stuff, and the hardware for AIX line is now the same as that for the AS/400 line (or whatever they're all called this week).
For our business, AIX is still rock solid, has excellent support (as you'd expect for the cost) and can dynamically switch resources between virtual systems. The CPU allocation is wonderful. It can automatically assign spare CPU to any system that needs it, giving preference to production systems.
The virtual networking and hardware self-monitoring is also far superior to what little I've seen in the Windows area.
While I can't comment on other systems, AIX has given us a lot of flexibility and reliability that the Intel team here (mostly Windows) don't get in their virtual server environment.
Of course all this is changing, and the smaller systems are getting the bigger system stuff.
So the real question is not "how ready is Linux to AIX?" but rather "can Linux do what I want now?" because all the mainframe technology is filtering down to be accessible to consumer grade stuff.
please stop the "hi didn't say it" troll (Score:2, Interesting)
Please can we let the "Bill never said that" troll die ?
Yes, there are quote of him telling that he didn't utter these precise words. :
On the other hand
I have to say that in 1981, making those decisions, I felt like I was providing enough freedom for 10 years. {...} that would last a great deal of time. Well, it didn't
and :
I laid out memory so the bottom 640K was general purpose RAM and the upper 384 I reserved for video and ROM, and things like that.
{...} It was ten times what we had before. But to my surprise, we ran out of that address base for applications withinâ"oh five or six years people were complaining.
He didn't say the precise words. On the other hand he admitted having a hand in the design and thus he is the short-sighted we have to blame for not taking enough headroom for letting further evolution, at a time when contemporary processors like the 68k, where making special provisions for future 32bits computing.
You're asked to help design the OS and the memory layout for a machine made by the maker of most big iron machines of that time (IBM !). This machine is probably going to stay around for a lot of time (again it's an IBM - not an obscure asian maker who might not be still in business). It will probably stay mainly in the business sector, where customer have money to throw at memory upgrades (ok, he got lucky with that : the machine enjoyed lot of popularity in small business and homes which are much more financially limited and won't necessarily throw that much money at it).
And what do you do ? You split the memory in a way which will limit further evolution. Just when computers are switching from 8bit to 16bit because the memory was too much cramped, you make a design which will create problems down the line again, too.
He could have laid the reserved memory in the beginning thus not putting any top limit to the memory.
He could have let the position of reserved memory open, (with an interrupt telling where the actual reserved memory lies) and thus letting machine use different memory layout as needs for memory increase.
Well, he was short-sighted, he made bad decisions and in consequence of these, programming on ms-dos has been a world of pain for several years.
Therefor he deserves to be ridiculed about the 640k limit. Even if he never said it literally.
Re:No, (Score:3, Interesting)
A black day for AIX on the desktop. I really wonder what's the problem there, warehouse costs?
The first sentence is incomplete, and the second one is a run on. I didn't even notice it until you asked because the construction is not unusual at all, even though it is technically wrong. There's a couple of these in your summary, but I wouldn't have immediately pegged you for a non-native English writer since most of us write like that anyways in informal postings.
Sun did the same thing (Score:3, Interesting)
Re:2009: Year of AIX on the desktop (Score:3, Interesting)
Was it simply that the AIX box was a multiproc power system?
No, it was the system architecture of the RS/6000 and the clarity of AIX that I enjoyed. It was like "playing with the real toys". At that time, Linux for instance, wasn't half as far as it is now.
Re:I recommend Xming instead of Exceed for X (Score:1, Interesting)
I've liked Xming because for the last few years it was being maintained and Cygwin/X wasn't. But in recent days Yaakov Selkowitz and Jon Turney have updated Cygwin/X to X.Org X11R7.4, which is harder for Colin to do for Xming because he is not using the Cygwin library and not yet using xorg's modular build system. Colin graciously gave permission for Cygwin/X to use his Xming patches. Xming is currently at 7.3, and it is apparently necessary to donate to the Xming effort to get the latest version rather than 6.9 (not that there's anything wrong with that). With Cygwin/X, you can easily install just the base system, xserver, and fonts and you will hardly notice any unneeded apps. Plus I find it very nice to have a whole Unix environment there to download when I need it. Xming is still much nicer to configure, though.