IBM Saves $250M Running Linux On Mainframes

coondoggie writes "Today IBM will announce it is consolidating nearly 4,000 small computer servers in six locations onto about 30 refrigerator-sized mainframes running Linux, saving $250 million in the process. The 4,000 replaced servers will be recycled by IBM Global Asset Recovery Services. The six data centers currently take up over 8 million square feet, or the size of nearly 140 football fields."

Comments

Proof of Linux's Environmentalist Friendlyness

[flayzernax]

This proves Linux has a smaller carbon footprint then other OS's!

Re:Proof of Linux's Environmentalist Friendlyness

[Anonymous Coward]

I suggest masturbation. That should keep at least one hand busy.

Re:Proof of Linux's Environmentalist Friendlyness

[Anonymous Coward]

no, we recognize a wannabe pretending to in a subculture they're not. better keep that day job with those rhymes, wigger.

Re:Proof of Linux's Environmentalist Friendlyness

[Anonymous Coward]

you little white ass bitch. come say that to my face and yo ass will get clocked. yeah you must think you all tough sittin behind your damn computer screen all day, which is the only thing that gives your skin any form of pigment. keep thinking like you are. keep eating your fat ass self to death itll make the world a better place without your damn whiny ass mouth going off over nothing.

Re:Proof of Linux's Environmentalist Friendlyness

[Anonymous Coward]

I can't believe I'm biting on any of this, but to be fair to him they aren't his rhymes. It's Public Enemy. Pretty much the finest rap/hip-hop group you're going to hear. But yeah he is a turd for flooding as much as I am for biting on this.

Re:Proof of Linux's Environmentalist Friendlyness

[Benaiah]

I really love the tags "linuxisenterprisebiyoch"
I think it should be used more often.
Linux>AllBitches! even though I don't use it, I should pretend since this is /.

Re:Proof of Linux's Environmentalist Friendlyness

[jollyreaper]

I can't believe I'm biting on any of this, but to be fair to him they aren't his rhymes. It's Public Enemy. Pretty much the finest rap/hip-hop group you're going to hear. Yeah, but that's like being the smartest 'tard in special ed, not really something to brag about.

Re:Proof of Linux's Environmentalist Friendlyness

[nasch]

Spoken like someone who's never listened to Public Enemy. If you have, I apologize.

Re:Proof of Linux's Environmentalist Friendlyness

[jollyreaper]

Spoken like someone who's never listened to Public Enemy. If you have, I apologize. Sturgeon's Law is universal. So just as 90% of the genres I like are infested with crap, a bare minimum of stuff like rap is also crap. The style of rap is lyrically dense and allows a performer to turn his performance into an essay, making a more conventional song seem sparse in comparison. Of course, because nobody is playing instruments anymore in rap, by necessity there has to be something to fill the song with. There are less words involved with rock to make room for the musicians.

The only modern rapper I've heard who's done anything interesting is Immortal Technique and even he ends up balancing out his contemplative songs with a bunch of ignorant posturing. One moment he's lamenting the state of his people, the next he's threatening to kill any nigger who fucks with him. Pick a side, dude: are you part of the problem or the solution?

I will confess that I'm not scouring the underground for good acts, I'm mostly reacting in disgust to the shit that gets on BET and MTV.

Re:Proof of Linux's Environmentalist Friendlyness

[nasch]

Your original statement seemed to indicate that all rap is stupid/lousy/bad. Now it seems you're saying most of it is garbage, with the possibility of some good stuff at the top? Or maybe you're still saying it all sucks, I'm not sure anymore. In any event, the statement of "most rap is ", which is probably true, doesn't exclude the possibility of some really good performers. And IMO PE is one of those. And you still haven't said whether you've ever listened to them. :-)

Re:Proof of Linux's Environmentalist Friendlyness

[mr_mischief]

FYI, Sturgeon's Law is, basically, "90% of everything is crud".

By saying that rap abides by Sturgeon's Law, that implies that 90% of rap is crud, while 10% is good or at least worthwhile.

It's also the same for every other genre of music, for video games (or software in general), for cars, for restaurants, for books, or for shoes. 90% of all of it is crud, and only 10% or so is worth even considering.

The only arguments to Sturgeon's law that I've heard are:

1) that the percentage might be a bit off

2) it's true, but the quantity of the better portions of crud makes up for it still being crud. As in, there's not enough within the top 10% to satisfy me or I can't afford it. So I'll gladly slum and scoop the top 10% of the crud and put up with it being crud because it's less cruddy than the rest, and I need some X amount of stuff, even if half of what I get is crud. This makes perfect sense for food and clothing if you're hungry and poor. For music, the best cure is getting some class and listening to good music in more than one genre.

Of course, there are people who have no taste or standards and who buy whatever they are told to buy by the media or by friends who likewise have no taste or standards. Therefore, some of the crud floats to the top, but it's still crud. There's room for disagreement over particular artists, companies, or products, but some are nearly universally seen as crud but still sell.

Now, I, personally, don't prefer rap as a genre to many other music genres. I'd still rather listen to the top %10 of good rap than the worst music in techno, rock, ska, rockabilly, funk, soul, or blues. While your PP didn't say, I will... I like Public Enemy, and I consider them part of the 10% of good rap groups and most of their music to be within the 10% of good rap music.

I'll put Public Enemy in my playlist along with Elvis, ZZ Top, DJ Keoki, Megadeth, The Chemical Brothers, Stray Cats, Pretenders, The Ramones, James Brown, Garth Brooks, Van Zant, George Clinton, Lynryd Skynyrd, MU330, UB40, John Lee Hooker, The Rolling Stones, 311, AC/DC, Future Sound of London, Aretha Franklin, Allison Krauss, etc. Not because I love rap, but because Public Enemy is good stuff.

Re:Proof of Linux's Environmentalist Friendlyness

[jollyreaper]

Your original statement seemed to indicate that all rap is stupid/lousy/bad. Now it seems you're saying most of it is garbage, with the possibility of some good stuff at the top? Or maybe you're still saying it all sucks, I'm not sure anymore. In any event, the statement of "most rap is ", which is probably true, doesn't exclude the possibility of some really good performers. And IMO PE is one of those. And you still haven't said whether you've ever listened to them. :-) I've heard their stuff, it didn't make much of an impression. We're at that interface between making a joke for a laugh and actually stating a position in a more serious fashion.

So, for clarification: I'm sure there are good underground rappers, just like the rock and metal I prefer has never been within shouting distance of the mainstream, at least here in the States. I'm not a fan of rap so I do not try to seek out the underground, I only encounter some of it in passing. I think that there is an inherent potential in the form of rap that is easy to overlook given who is popular today, much in the same way metal could be overlooked if your only introduction to it was Poison and Winger. And I also think that the popular rappers are underselling the medium by being a bunch of talentless hacks. The same can be said about commercial Nashville country vs. the genuine stuff that wasn't factory-farmed.

Re:Proof of Linux's Environmentalist Friendlyness

[jollyreaper]

FYI, Sturgeon's Law is, basically, "90% of everything is crud".

By saying that rap abides by Sturgeon's Law, that implies that 90% of rap is crud, while 10% is good or at least worthwhile.

That is correct. I'd say 99.9% of the rap I've encountered as someone not looking to find rap has been crap. Since I'm not looking for gems, I'm not likely to encounter the 10% which I would imagine is buried deep and underground vs. the pablum mass-marketed via the media, that stuff would be drawing solely from the 90% that is shit, thus giving us near 100% shit on the airwaves. Rock/Metal has that same problem. There are some brilliant bands out there along with others that are good but not my cup of tea. But do you hear about them or Limp Bizkit?

It's also the same for every other genre of music, for video games (or software in general), for cars, for restaurants, for books, or for shoes. 90% of all of it is crud, and only 10% or so is worth even considering.

Correct. I am not disagreeing with you. I've been disappointed with most of the thai restaurants I've eaten at but that's because the best one I've ever been to is down the street and was my introduction to thai. I promise you any thai fan would thoroughly enjoy a meal there. But if you don't like thai, you'd never seek the place out and might not even try it even if you knew about it. I'll wade through crap trying to find something I like but it would be unexpected for me to start digging in a brand new pile of crap, a pile of stuff I have no interest in, just to see if there's a nugget of goodness somewhere.

1) that the percentage might be a bit off

I think it can vary by the topic. The percentages he introduced were just conversational math to convey a general opinion rather than a scientifically accurate ratio.

2) it's true, but the quantity of the better portions of crud makes up for it still being crud. As in, there's not enough within the top 10% to satisfy me or I can't afford it. So I'll gladly slum and scoop the top 10% of the crud and put up with it being crud because it's less cruddy than the rest, and I need some X amount of stuff, even if half of what I get is crud. This makes perfect sense for food and clothing if you're hungry and poor. For music, the best cure is getting some class and listening to good music in more than one genre.

It's all a question of whether you can cherrypick. If a band only turns in a good performance half the shows they play, it's a waste of time to go. If you buy CD's at random and unheard over the net, your success ratio is going to be low. (that's what it was like before napster.) But if you can get solid recommendations, can listen to the music in advance, your odds of making a good pick are far higher than what a random pick of a 90/10 good/crap split would be predicted to net you, statistics-wise.

Of course, there are people who have no taste or standards and who buy whatever they are told to buy by the media or by friends who likewise have no taste or standards. Therefore, some of the crud floats to the top, but it's still crud. There's room for disagreement over particular artists, companies, or products, but some are nearly universally seen as crud but still sell.

That's what friends are for. That's also what makes the internet such a nice resource. "People who bought this also bought that. Will you like it? Maybe, maybe not. But it's more informed than a random guess."

Now, I, personally, don't prefer rap as a genre to many other music genres. I'd still rather listen to the top %10 of good rap than the worst music in techno, rock, ska, rockabilly, funk, soul, or blues. While your PP didn't say, I will... I like Public Enemy, and I consider them part of the 10% of good rap groups and most of their music to be within the 10% of good rap music.

Fair enough. Just like there are groups I would consider to be top 10% in genres I like that you would be indifferent to

Re:Proof of Linux's Environmentalist Friendlyness

[nasch]

Hey, what are you saying about Winger?? OK, I've tipped it back the other way. Whew.

Re:Proof of Linux's Environmentalist Friendlyness

[Anonymous Coward]

fag

No

[Anonymous Coward]

30 refrigerator-sized mainframes

The six data centers currently take up over 8 million square feet, or the size of nearly 140 football fields Something doesn't add up. That or I have some consulting time I'd like to sell IBM, I can assuredly pack 30 fridges into less space than that.

Re:No

[martinX]

The current setup takes that much space. The new setup will be (presumably) smaller.

The six data centers [with the 4000 servers] currently take up over 8 million square feet

Re:No

[pete-classic]

Sure. That makes sense. A typical PC takes up about twice as much square footage as my apartment.

-Peter

Re:No

[Anonymous Coward]

Yes. It says "six locations".

What if "location" means a building with parking spaces, atrium, lunch rooms, guard shacks, meeting rooms, offices, a floor for executives and management, etc.

That number becomes realistic.

Re:No

[jtcm]

The six data centers [with the 4000 servers] currently take up over 8 million square feet

The current setup takes that much space. The new setup will be (presumably) smaller.

I should hope so!

8,000,000 ft^2 / 4,000 servers = 2,000 ft^2 per server

My god, those are large servers! They must still be using vacuum tubes...or maybe a 65cm manufacturing process.

Re:No

[Anonymous Coward]

Not that impressive. Microsoft recently announced that they had surpassed 100,000 servers in their datacenters...

Re:No

[gtall]

Yeah, probably 100,000 Dells. Now if they'd buy some mainframes, then...M$ Malware doesn't run on mainframes.

Re:No

[BDPrime]

The story has an error. IBM operates a total of 8 million square feet of data center space. These six data centers in particular are just a portion of that. Though I couldn't figure out how much square footage is in the six data centers, IBM told me that the three U.S. sites (New York, Connecticut, Colorado) take up about 184,000 square feet. I don't know how much the other three sites (U.K., Japan, Australia) take up, but assuming that they also add up to 184,000 square feet, you're talking about 370,000 square feet total for the six sites.

Of course they're saving money

[Trailer Trash]

Because they're using all that Microsoft IP without paying for it....

(it's a joke)

Re:Of course they're saving money

[slazzy]

Wow, imagine how much money they could have saved if they had switched to Windows Server 2003, you know, since the "total cost of ownership" is so much lower? (laugh)

Must be SCO jacked up the rates...

[jkrise]

for AIX on those mainframes! After all, AIX has more Unix IP than Linux, isn't it?

Re:Must be SCO jacked up the rates...

[grantek]

or AIX on those mainframes! After all, AIX has more Unix IP than Linux, isn't it?

SCO UNIX runs on the x86 architecture, that was the basis of the claim that Linux contained copyrighted SCO code. IBM's Linux on POWER solutions run on, um, POWER :)

Really this is just a slashvertisement - it's great they're using Linux on a mainframe, but they're just IBM mainframes running multiple Linux instances, rather than multiple IBM servers running Linux. Honestly, on IBM hardware, I'd prefer IBM's OSes, but they're marketing the fact that you can have a high-powered, highly efficient, highly available consolidation solution that runs your existing Linux apps

Re:Must be SCO jacked up the rates...

[larry bagina]

SCO owns the rights to the original (AT&T) Unix code. (Maybe. Novell claims they still own the rights and the SCO was just their license collector). All *real* unixes (including AIX and Solaris) pay/paid SCO a Unix licensing fee.

Re:Must be SCO jacked up the rates...

[Airline_Sickness_Bag]

SCO owns the rights to the original (AT&T) Unix code. (Maybe. Novell claims they still own the rights and the SCO was just their license collector).

There is no "Maybe" about it. SCO does not own any of the System V code - they own Unixware and Opensewer, just like SGI owns Irix. None of the copyrights on the SysV code transferred to The Santa Cruz Operation, let alone SCO/Caldera. And SCO has not provided any documentation that shows what (code) copyrights were transferred with the specificity required by law. SCO owes Novell millions of dollars for the licences they sold to Microsoft and SUN - despite their claims that those licenses do not fall under the contract with Novell (I.e., not about SysV), their filings to the Security and Exchange Commission state that they were about SysV.

According to their contract with SCO, SCO is to collect the royalties and pass them directly to Novell. Novell then gives them 5% of it.

Not all Unix licensees pay royalties - IBM and Sun have fully paid up licences. So they never give SCO a cent.

Re:Must be SCO jacked up the rates...

[grantek]

Ok, I didn't realise that, but my point still stands - I doubt IBM are paying a per-license fee for AIX, and especially not z/OS

Re:Must be SCO jacked up the rates...

[Anonymous Coward]

AIX runs on pSeries, which in the great scheme of IBM is a "small" server. Most mainframes are running z/OS, which is simultaneously OS-like and hypervisor-like. Then they carve off what we'd call VM's, but they're not exactly the same thing, and on those partitions they run damn near any OS you'd run an enterprise application on: Windows, Linux, AIX, NetBSD, DOS.

Re:Must be SCO jacked up the rates...

[Anonymous Coward]

z/OS is not a hypervisor. The mainframe 'VM' operating is a hypervisor and you can run both z/OS and Linux under it.

More commonly, the mainframe is partitioned using a hardware feature called LPAR (logical partitions).

Ric Romero says "virtualization saves space"

[Gothmolly]

Um, hello, while this may deserve the 'neat' tag, it's hardly newsworthy.

People are consolidating lightly (and heavily!) used servers into VMs all over the place.

Re:Ric Romero says "virtualization saves space"

[Urusai]

VMs are just an excuse to keep running the same shoddy software forever. As for arguments that they help protect against crashes and the like, well, that's because you have shoddy software. Arguments that they let you run multiple environments, well, same story. You could have been paid as a software developer to rewrite crapware for the modern age, but instead all of you just cream your jeans for VMs that steal the livelihood of software developers. Dunces, all!

Re:Ric Romero says "virtualization saves space"

[jimmydevice]

I have seen applications that are well written, understood and maintained and are 30 years old. I suffered through the rewrite of a well known, commercial revision control application we used to maintain our code. Originally in C, it was rewritten in Java with horrible results. Our checkout times went from a few minutes to 10 minutes for a full checkout. All our custom tools no longer worked, but the interface looked fabulous. Frankly, I'll take a text console based application over some bloated "modern age" crapware any day.

Re:Ric Romero says "virtualization saves space"

[cyphercell]

depends on the database behind that 30 year old software for me. I've seen extremely flat databases with nothing but a text console wrapped around it. Extremely, poor integrity standards and the data falls apart like a stack of Jenga blocks (the more people playing the quicker it falls). Computers back then were (often) used to supplement paper systems and sometimes it really shows.

Re:Ric Romero says "virtualization saves space"

[Ilgaz]

Same for UI too I think. There is a British bank which their branches are relatively new here. I was there yesterday and I was amazed watching a single broker doing 5 customers work same time, million dollars stuff in 5-10 mins. There was no "mouse" or any graphics involved, I think it was some Mainframe Terminal application.

I bet the OS was virtual too running inside mainframe. :)

I bet as ex IT manager he could answer my questions but I don't think they are allowed to answer and guy was really busy moving millions of dollars like ping pong.

Re:Ric Romero says "virtualization saves space"

[mikehoskins]

I think Linux saving one company a quarter of a BILLION dollars and most of 8 MILLION square feet is hardly newsworthy. The electrical power saved (enough to run small towns, I'd guess) would hardly be worthy of attention, either.

Mainframe Linux and IBM (and virtualization) are hardly newsworthy.

Let's keep articles like these off of Slashdot and adopt Digg style "community gossip as fact," instead. Now THAT's newsworthy....

A pleasure to work with, as well..

[bigattichouse]

We (Bigattichouse's Vectorspace Database) went through their Linux certification (as well as Grid cert), and they were a pleasure to work with - providing expert advice and patience in every step of the process. Not exactly on topic, I guess, but I thought I'd share. They really seem to embrace the engineering and spirit of Linux.

Re:A pleasure to work with, as well..

[eric2hill]

Hey man! It's been a while! What's up!? Send me an email...

Re:A pleasure to work with, as well..

[etnu]

I recently worked with IBM to interop with sametime (their IM network), and my opinion of their engineering practices would probably get me fired for disparaging a partner.

Re:A pleasure to work with, as well..

[Ilgaz]

IBM always impresses with level of the support and especially their attitude of "never abandon people you sell stuff to".

OS/2 was declared dead 10000 times (even by fans) while it was getting some new graphic drivers actually purchased from Sci-Tech software.

If you are a PowerPC (G4/G5) user and in desperate need for non beta, working Java 6, you simply install PPC-Linux and install IBM supported, non beta Java 6 along with CPU acceleration to it. That is the system and OS which Nvidia/ATI refuses to ship a basic binary driver or Adobe doesn't even bother to check for Flash support.

That is how they get that image. In fact, impossible in short term but lets say Apple tells "I am not shipping any OS X updates for PowerPC, buy Intel", it would take 2-3 months for IBM to ship AIX 5L bootable install for PowerPC Macs or start coding mainboard etc. support to Linux kernel themselves.

Re:A pleasure to work with, as well..

[Nexcis]

The grammar/syntax nazi in me is screaming to get out. I, MUST, (RESIST).

Re:A pleasure to work with, as well..

[Ilgaz]

The grammar/syntax nazi in me is screaming to get out. I, MUST, (RESIST). I wondered what kind of horrible mistakes I did to cause this message by
you and installed a grammar/spell checking tool. Beside couple of
"punctuation should be in quotes" and an obviously missing comma, there
are no issues. It also says I shouldn't use "In fact" since it is not
needed.

The tool is maintained/coded for a decade, and it is fairly expensive.
People are using it instead of MS Word one without thinking.

If your standards are way beyond the capabilities a $70 program offers,
I must apologize and suggest filtering me out via adding as "Foe" or
something.

Re:A pleasure to work with, as well..

[Nexcis]

The horror.

Re:A pleasure to work with, as well..

[Maxo-Texas]

Don't loose your cool, man.
You didn't post bad.

2000 sq feet per small computer?

[Anonymous Coward]

The article says that the data centers required for the 4000 "small computer servers" aggregate to about 8 million square feet. It takes IBM 2000 square feet to house a small computer? Also, saving $250 million suggests that it costs them something over $60K per "small computer" even ignoring the price of the new mainframes. Amazing.

Re:2000 sq feet per small computer?

[RuBLed]

Vacuum tubes = costly = takes up large space = less green = it's about time

(I'm so sorry)

Re:2000 sq feet per small computer?

[irtza]

This is all easily explained. $62K and 2000 square feet can be very easily explained. Now this guy has to be a typical geek, so they need to provide him with enough soda for the day - IE one vending machine. Now, most of the time, the computers will take care of them selves so they will need something for entertainment - like a secondary system with cool games so he doesn't try to install them on the server. Now, in an age of high-def and large screens, he will require that the game is played on the largest LCD screen available and will also require an appropriately sized room to play it in. Of course he would require access to a bathroom and with the expected hygene of a geek, it may be better to not have them share. Now, we haven't even begun to talk about the food requirements of said geek. So, at the end of the day, 2K in electricity and parts for the server per year, 30K for the salary of the employee (entry level position babysitting a server) - 50K once benefits and co-pays are included. an addition 8K for the electricity for the "supportive devices". Now, you may be thinking that you can fit all this in way less than 2000 square feet. What you are forgetting is that a geek is a solitary animal. Letting them get close together would plummet their already weak work ethic. You will be more likely to end up with a contest to see who can overclock their server faster and end up with a bunch of burnt units or potentially a burnt building.

I thought all this was obvious, but I guess it takes time for people to learn... you might want to speed up your computer a little bit to keep pace.

Re:2000 sq feet per small computer?

[iagreewithmichael]

If entertainment was the problem, a blow-up doll would have saved IBM a ton of money

Re:2000 sq feet per small computer?

[Anonymous Coward]

lol.

Re:2000 sq feet per small computer?

[HAKdragon]

Hmm, so to shorten it up, "Vacuum tubes = it's about time". It's about time for vacuum tubes?

Re:2000 sq feet per small computer?

[MyLongNickName]

But what will the geek due with the other 7 hours and 58 minutes of his shift?

Re:2000 sq feet per small computer?

[flyingfsck]

That must include the office space for the hundreds of MSFTs required to apply patches and restart the 4000 servers all the time.

Re:2000 sq feet per small computer?

[slazzy]

Look at pr0n on the high speed internet connection?

Re:2000 sq feet per small computer?

[kars]

But they need those tubes for their internets!

Re:2000 sq feet per small computer?

[Scaba]

Yea, but tubes sound way better.

Re:2000 sq feet per small computer?

[Anonymous Coward]

It's called "pornography" or "porn" for short, fuckface.

Re:2000 sq feet per small computer?

[darkmeridian]

Each of the smaller facilities require lots of overhead in terms of money and space for air conditioning, security guards, power generators, walls, fire control systems, UPS, etc. Once the systems were centralized, these expenses also became centralized. Thus, less space is required, costs go down, and everyone is happy.

Re:2000 sq feet per small computer?

[Yetihehe]

No no, in his logic ( A = ( B = C ) != ( ( A = B ) = C ) so you can't really shorten like this.

Re:2000 sq feet per small computer?

[gnuman99]

Err, couldn't they centralize their 4000 small server in one place? Or at least in the same places they put their big boxes? 4000 is not that many servers considering you can have usually at least 20 per rack and then you have 20 racks.

Re:2000 sq feet per small computer?

[Anonymous Coward]

4000 is not that many servers considering you can have usually at least 20 per rack and then you have 20 racks

Ummmmm .... 20 x 20 = 400 dude. You would need 200 racks of 20 each.

Re:2000 sq feet per small computer?

[kabocox]

Vacuum tubes = costly = takes up large space = less green = it's about time

Hey, if it's not broke yet, we ain't fixing it yet.

My employer recently 'consolidated' too.

[MarcQuadra]

My employer recently 'consolidated' their server farm too. We used to have a room with fifty aging Dell PowerEdge servers, each running independently, requiring massive support, cooling, and electricity.

Now we have ten VM servers running all the migrated services, PLUS a room with about fifty aging Dell PowerEdge servers, each running independently, requiring massive support, cooling, and electricity.

I never thought 'consolidation' would require so much more space, electricity, air conditioning, and upgrades to core switches and UPS units.

Re:My employer recently 'consolidated' too.

[Anonymous Coward]

You're supposed to turn off and sell off the servers you replace. (Score -1: Obvious).

(Strange thing is, I make a good living replacing aging mainframes by linux clusters. mainframes are fine when you're doing transaction processing. But for cpu-bound stuff, you're better off with a room full of opterons).

Re:My employer recently 'consolidated' too.

[Anonymous Coward]

Yeah, real life ALWAYS works like that.

I did a consolidation for Wells Fargo Home Mortgage. We took 500 servers down to a dozen beefy VMWare ESX boxen. What did management say when we were done? Cool beans guys, now leave the old servers up and we'll tell the board we doubled our capacity for 25% of what it usually costs.

Re:My employer recently 'consolidated' too.

[supremebob]

Yeah... that sounds about right. I've ordered plenty of VMWare host servers for "consolidation" as well, only to have them grabbed up my new projects that were higher priority and had more potential to make the boss look good to his managers. Sure... A few servers get shut off every year due to projects that leave, but the actual server count ends up expanding every year.

I wonder if IBM factored in the number oddball projects that require Windows systems in their server count? Windows won't run in a zSeries VM, and there is plenty of software out there that is still Windows only.

Re:My employer recently 'consolidated' too.

[LWATCDR]

"I wonder if IBM factored in the number oddball projects that require Windows systems in their server count? Windows won't run in a zSeries VM, and there is plenty of software out there that is still Windows only."
Not on servers. IBM uses Lotus Notes so no need for Exchange Servers. They probably us DB2 for any SQL systems so no MSSQL servers.
Frankly I can't think of many reasons that IBM would have to keep many Windows servers around except for testing IBM software running on a Windows server.

Re:My employer recently 'consolidated' too.

[Anonymous Coward]

I wonder if IBM factored in the number oddball projects that require Windows systems in their server count? Windows won't run in a zSeries VM, and there is plenty of software out there that is still Windows only.

Well, if you can run linux in a zSeries VM, you can run vmware on linux, then run windows on VMware...

Re:My employer recently 'consolidated' too.

[Anonymous Coward]

Many companies run a LOT of custom software, and porting that software to Linux to gain the advantages of vendor independence and support for BIG IRON is a popular choice for IBM customers.

Re:My employer recently 'consolidated' too.

[PitaBred]

So tell your boss you want to install another dozen boxes, and then they'll be able to tell the board that they doubled capacity for half of what it normally costs, more than halved electricity consumption, and freed up a lot of physical space for future expansion without having to acquire new property. It's all in how you spin it, and get your boss to realize how he'll look good for doing so.

Re:My employer recently 'consolidated' too.

[Anonymous Coward]

Uh, no.

The processor instruction sets, among other things, are radically different. For one key thing, VMWare doesn't support the s360[x] architecture, despite its running Linux. If VMWare did, Windows wouldn't run on the architecture, and if you emulated an x86 or x86_64 processor and PC-type I/O systems well enough to run Windows, it'd be really slow. As a previous poster wrote, mainframes are targeted at a specific workload, transaction processing. They make great database servers and web servers for instance.

I've run small GUI apps on a single-IFL z890 and z9 (SuSE YaST & some other small stuff, CentOS installer). It runs fine, but I don't think it would scale well. The maths processing power just isn't there, at least on an entry level system.

More info than I meant to give, especially since I'm pretty sure the parent poster was kidding. Hey, it's late! Besides, I can imagine a certain manager I know taking the idea seriously.

Re:My employer recently 'consolidated' too.

[Fizzl]

*shiver*
Please don't say that word. I get chills and shivers.
It's been already three years since I had to use it, but even seeing the name of Lotus Notes (AAH, MY FINGERS!) makes me curl up and sob on the floor.
What a great product!

Re:My employer recently 'consolidated' too.

[cerberusss]

Now we have ten VM servers running all the migrated services, PLUS a room with about fifty aging Dell PowerEdge servers

So what everybody is now wondering: "WTF are those fifty aging Dell PowerEdge servers doing besides idling?"

Re:My employer recently 'consolidated' too.

[Threni]

I would have just switched off the 50 old servers on the same day you switch on the new ones. Just `yee-har` it. What could go wrong?

Re:My employer recently 'consolidated' too.

[Citizen of Earth]

Reminds me of a government program to get the public to use energy-efficient refrigerators in Canada. It was quite successful—people replaced their inefficient kitchen refrigerators with the new ones; however, they moved their old ones to the basement and put beer in them. The environment takes another one for the team!

Re:My employer recently 'consolidated' too.

[Trashman]

Part of the cut over is you're supposed turn off the old ones at some point. Let me guess, politics? I feel your pain...

Re:My employer recently 'consolidated' too.

[LWATCDR]

Never used it but how much worse could it be than exchange?

Re:My employer recently 'consolidated' too.

[Anonymous Coward]

(Strange thing is, I make a good living replacing aging mainframes by linux clusters. mainframes are fine when you're doing transaction processing. But for cpu-bound stuff, you're better off with a room full of opterons).

What's better off? IT salaries, or the total cost?

Re:My employer recently 'consolidated' too.

[Richard Steiner]

Notes is actually cool when used properly. Notes used just as an e-mail client (like we do here) is overkill. It works, and it's okay, but it isn't fast. :-)

Re:My employer recently 'consolidated' too.

[CypherOz]

It's been already three years since I had to use it, but even seeing the name of Lotus Notes (AAH, MY FINGERS!) makes me curl up and sob on the floor.

Notes 8 is Eclipse Rich Client - the world has changed and it is actually good. Beta 3 is public and rumor has it Late Aug for release.

Imagine....

[Jrabbit05]

Imagine a Beowulf cluster of those!

Oblig.

[Ethanol-fueled]

But does it run....ah, hell. Um, imagine a beowulf cluster of...hmm.

Fuck it. Who wants pizza?

Re:this article has been up a couple minutes

[Anonymous Coward]

You fail it.

System z Mainframes

[o2sd]

It's kinda hard to find technical specifications on these mainframes beyond marketing fluff. After some looking I found this brochure, which has some interesting information on the firmware and a few details of the I/O, but not much about the processing units, and why one of these would be able to replace 133 blade servers. It does mention up to 30 superscalar processors per box, but I'm not really sure what that means. (Maybe they go next to the inverting flux capacitor).

Re:System z Mainframes

[Anonymous Coward]

newer zseries an have up to 54 processors per machine. Actually more -- 54 general purpose processors. IO, Memory management, crypto, etc are handled by dedicated processors.

Re:System z Mainframes

[BrynM]

It's kinda hard to find technical specifications on these mainframes beyond marketing fluff.

Part of that is because IBM will customize the machines to your heart's content. The sky and your budget are the only limits. They leave a good many of the loadout details (xGB/TB of RAM, DASD storage size, # of CPUs per card, # of CPU cards, even number of mainframes - they can be chained in parallel). You should look at the Z series hardware specs for the general details and look up what details you don't know.

If you're looking for benchmarks or comparisons to x86/x86-64 or other commodity architectures good luck - they are nearly impossible to find. This is due to the implementations being on entirely different scales. The best comparison you an find is the MIPS per CPU. You can find some slightly stale numbers here (BTW: an LPAR is something that's been around on mainframes for several decades - one LPAR can run up to several hundred x86 VMs concurrently).

Re:System z Mainframes

[duffbeer703]

The other part of it is that the zLinux miracle is mostly bullshit.

zLinux is great if you're consolidating mostly idle, low priority resources. The "magic" that allows you to save money while simultaneously getting raped by IBM Global Services and paying too much for hardware is thin provisioning. You might assign 10 Linux VMs 1GB RAM each, and only have 4GB of actual memory available. Same thing with CPU. This is an efficient use of resources, if your applications don't all require memory at the same time... if you're like me, your employer has lots of memory-hogging J2EE stuff. On the other hand, crypto and networking between VMs is blazing fast.

Another problem is that in a big business that has mainframes, the mainframe folks are very conservative, use much stricter change and other controls than most open systems shops and don't understand the workloads that Unix/Linux systems get. They get prickly when your linux systems start looking for lots of resources. Everything takes about 3x longer.

The other issue is that the VMs are dependent on the Linux installation on the LPAR, and you may not have many LPARs available. If you want to run Red Hat & Suse, or RHEL5 and RHEL4, you need an LPAR for each. Nobody (except for a few showboat customers) is investing in new mainframes, and you may only have a few LPARs available on an existing mainframe available.

So if you have a business model like providing lots of cheap (and mostly idle) virtual servers, and you already have a major Mainframe investment, zLinux is a great solution. Otherwise, you're probably better off looking at the hardware virtualization options that you can get from Sun or even on whatever IBM calls RS/6000's these days for 1/5 of the cost.

Just a note: I'm not a mainframe guru, and my views are slanted based on my experience in working at a particular employer about a year and a half ago. So some of the issues may have changed, or the options available to me may have been limited due to some site-specific restriction that I am not aware of.

Re:System z Mainframes

[ChrisA90278]

"one LPAR can run up to several hundred x86 VMs concurrently)."

When I started out the "hot" PC, the best you could get, was a 4Mhz Z80 running CP/M. I had one of those at home and at work, I worked the operating system of a very old (even then) CDC mainframe. It was a CDC6600. We had a Z80 emulator that ran on the 6600 and we could emulate a Z80 at about 20 times real time. Not bad, a virtual PC running on a mainframe in the late 1970's

Us software people really need to get off the ball and think of something new rather then just re-implementing 40 year old ideas on ever cheaper and faster hardware.

Re:System z Mainframes

[Anonymous Coward]

That depends on what the new idea is and whether it is sufficiently better to warrant replacing a proven approach.

Many 40 year old ideas, especially in computing, are actually better thought out than more current approaches.

'Shiny' isn't good enough.

Re:System z Mainframes

[jgardner100]

This whole thing is based on the fact that most people only read the headlines. IBM want people to think they saved a squid load of money by running Linux (and need IBM services to do so) rather than the mundane replace lots of old servers with a few new ones (which HP, DELL, Sun etc will do for you as well.)

But will it blend?

[StinkyPinky]

Sorry, I thought I would bring down the Slashdotters to the Digg level.

Re:But will it blend?

[RuBLed]

It would but Blendtec would need to consolidate their 4,000 blenders to around 30 refrigerator sized ones....

Awesome consolidation

[toby]

There's nothing new under the Sun: And Sun's offerings in hardware and software are also very much aimed at consolidation. Bring it on.

An obvious conclusion

[GFree]

Saving a lot of dough by using Linux on servers makes sense, heck it's fairly obvious to anyone here, that's where it excels.

I think Slashdotters would be more interested in stories that focus on a company switching its desktops to Linux though. Servers running Linux are pretty common. We want news about the desktop front; it would be more newsworthy at least.

Re:An obvious conclusion

[teh moges]

That would be good, but I do want at least one new story a week.

The point of this story isn't to point out that Linux servers are cheaper/better/faster, its to point out that the Linux platform got some publicity. For those trying to get everyone switching to Linux for their desktops, publicity is their one major problem (as well as many smaller problems, but thats an argument for another story).
Remember when Microsoft said that Linux infringes on their patents but they weren't going to sue? They were never going to sue, they just wanted the good publicity of "playing fair" with the competition that "steals all their hard work".
Anyway, my point is to look at the publicity, ignore the details.

Re:An obvious conclusion

[f00man]

This is the year of Linux on the desktop. :P

Re:An obvious conclusion

[mr_mischief]

Let me tell you a little true story.

This is the year of the Windows, Mac, and Linux desktops running side by side on a single geek's desk, while his mom, sister, brother-in-law, sister-in-law, and other brother-in-law and half his friends run only Windows. His grandmother and some of his aunts and uncles don't touch a computer, ever, if they can help it. His dad runs Unreal Tournament, Doom, and sometimes Axis and Allies on Windows 98 hoping he won't cause a massive crash. His wife runs Linux well enough to log in, type startx, play tuxracer, shutdown X, and log out.

Hey, it's slow, but it's progress.

Re:An obvious conclusion

[HeroreV]

His wife runs Linux well enough to log in, type startx, play tuxracer, shutdown X, and log out. What kind of desktop Linux distro is that?

Re:An obvious conclusion

[mr_mischief]

It's Mandriva 2006, which would normally be installed with 5 as the default runlevel. It's at runlevel 3 on boot because I often use it remotely or from the CLI and only start up X as needed. My wife was bored one day at the office, so it now has TuxRacer on it, and she's learned how to manually start X.org in order to play.

$250M??

[evanbd]

Lets see... $250M / 4000 = $62,500 per server being consolidated? I mean, I know floor space, buildings, racks, power, AC etc cost money... but that's still a *lot*. Anyone care to chime in on how close to normal that is?

Re:$250M??

[brxndxn]

The old servers were Macs?

Re:$250M??

[Anonymous Coward]

They're probably computing cost over the expected lifetime.

Combine IT salary for 3-5 years, power over 3-5 years, etc. etc. and that number makes sense.

Re:$250M??

[thedarknite]

According to another article it is saving the $250M over 5 years, predominately from reduced running costs

Re:$250M??

[flyingfsck]

You are forgetting about the 4000 MCSE that they can now lay off.

Re:$250M??

[RockoTDF]

No, the POWER architecture is what IBM uses in their servers and high end workstations. PowerPC is a stripped down version that was used by apple prior to the intel switch.

Re:$250M??

[caluml]

You pay your servers a salary? You're doing it all wrong...

Re:$250M??

[ralphdaugherty]

They're probably computing cost over the expected lifetime.

Combine IT salary for 3-5 years, power over 3-5 years, etc. etc. and that number makes sense.

      This is the crux of IBM's virtualizations ads; you can get rid of admins and save big bucks. Also electricity and floor space, but you have to throw salaries in there to come up with the savings they suggest.

      Why does it take less IT salary as suggested above for 4000 virtual Linux servers on a mainframe than with 4000 Linux servers?

  rd

Re:$250M??

[einar2]

Yes, the number sounds realistic.

However, a lot is probably gained from licensing fees as well. Typically, this works for a year or two until the vendor realizes that you run their software now on a mainframe. Then they start to change license agreements and cost goes up again.
Currently, IBM is trying to lure software on their mainframe with huge license discounts (special JAVA cpus that are close to for free, linux on z/OS which has a reduced license fee, etc.). You can guess what wil happen once your application platform is migrated...

what does this have to do with Linux?

[pedantic bore]

The story here is about consolidation, virtualization, etc.

Linux is a small part of the technology involved here. z/OS is the real story here.

Re:what does this have to do with Linux?

[Anonymous Coward]

Actually z/OS has nothing to do with virtualization.. it's z/VM that is doing all the magic..

Re:what does this have to do with Linux?

[pedantic bore]

I've always been fuzzy on where z/OS ends and z/VM begins, so you're probably right, but the point remains the same -- this isn't about Linux, it's about something that starts with a "z".

Re:what does this have to do with Linux?

[gingerfire]

z/VM is the virtual hipervisor -- it has nothing to do with z/OS. (Which can run under z/VM in a virtual machine like linux.)

But how many LOCs?

[Anonymous Coward]

The six data centers currently take up over 8 million square feet, or the size of nearly 140 football fields. The real question is: how many Libraries of Congress can they store?

Re:But how many LOCs?

[Anonymous Coward]

A good 6 inches of chocolate rain.

Re:But how many LOCs?

[Anonymous Coward]

chocolate rain!!! WOW

$250M per how long?

[TimFreeman]

I could imagine that replacing a bunch of little computers with a few big ones would result in saved electricity, rent, and maintenance cost. Those are all ongoing expenses, so I could believe that they save $250M per year or decade or whatever, but the claim was that they saved a lump sum of $250M. If they were thinking of buying that pile of little computers, I could also believe that they saved $250M by buying a few big ones instead of a bunch of little ones; however, they said they were replacing the little ones. I can't find an interpretation of the article that makes any sense; can anyone else do better?

Re:$250M per how long?

[Dr. Smoove]

I'm not even sure this actually happened. I can make up stories like this about my non existant company.

1. "We consolidated 300 LOC's and 17 Volkswagen Beatles worth of computers by switching from Fridges to Dishwashers, in one of those wacky datacenter trailer majigs. We saved 2^64 million dollars in the process, and bought IBM."

2. ???

3. Profit!!11

IBM

[Anonymous Coward]

Considering IBM is investing a sizeable chunk of time and money into their Linux, it makes sense they'd claim using it saves millions of dollars, just as Microsoft would claim using their own server software saves millions of dollars.

Re:IBM

[Anonymous Coward]

Um, the ./ title says: IBM Saves $250M Running Linux On Mainframes

TFA title says: IBM saves $250 million consolidating Linux servers on to mainframes

It is the consolidation that is saving IBM the money not using Linux per se. IBM also has AIX they could use in lieu of Linux.

Well Duh.

[ryanisflyboy]

If you take hundreds of cabs and consolidate them down to 40 (with the associated consolidated storage) you are going to save millions. That has little to do with Linux. It is the modern mainframe that makes this kind of thing possible, which is why more people are moving to them. They must have a lot of servers spinning idle to get this done.

The reason why companies are in this pickle is because they thought more was better. They though "All we need to do is buy 4000 x86 servers and we can do tons of work." They didn't realize how HARD it is to get 4000 servers to operate in a cluster so you can take advantage of those individual systems as one body. So, they ended up with islands of computing power instead of a cluster. Naturally the mainframe consolidates these islands back to computing continents and you end up running the mainframes at near capacity all the time. Modern mainframes make this easy with dynamic CPU/RAM allocation, as well as dynamic storage. So you segment out the mainframe in to four or eight chunks. Chunk 1 is hot, chunk 3 and 5 are idle. Simply re-assign some of the CPUs from chunk 3 and 5 to 1 until the load goes down. You can take advantage of this in a big way if you segment your work load to match global demand. So chunk 1 might be data for the western USA, and chunk 7 might be EMEA. You can bounce resources between those segments much more easily. You can even script it. HP has an offering that does this automagically, I'm sure IBM has something similar.

Now, my personal opinion is why Linux? Some of the more advanced features like dynamic RAM, CPU, and IO allocation don't appear to be that solid to me. Perhaps IBM added these features to Linux or made them more robust? Maybe they run Linux inside an AIX virtualization container?

Re:Well Duh.

[Anonymous Coward]

Well with Linux you get an environment that's familiar and all the dynamic ram + CPU + IO is pretty solid. AIX is usually not in the picture it's z/VM. I believe with z/VM even the memory is virtualized so you can overcommit memory to your Linux guests and do other funky things. Of course z/VM + linux talk and have some bit of communication whether If i need this page for a long or short time. This allows z/VM to make good decisions as to determine which pages to free and to keep for the good of ALL GUESTS.

Re:Well Duh.

[dredwerker]

I thought you said take hundreds of cats and consolidate them down to 40.

I was a bit worried for a moment :)

Re:Well Duh.

[the not-troll]

Wouldn't that be more like taking hundreds of cabs and consolidating them down to 40 buses?

Re:Well Duh.

[nyclinix]

Why Linux? The organization may be a big LAMP shop, or mysql, or Oracle, or Web Logic or any of a multitude of software offerings that feel more at home on Linux/UNIX/MS than on Big Blue's Z OS.

almost enough to pay for one executive's perks

[Anonymous Coward]

great! glad to see the company is doing so well saving money.

IBM's been doing this for-ever, dude.

[crovira]

I had to maintain some software that was running on a aging 370 mainframe. The 370 was emulating a 360 which was emulating a 1401.

It was pension and payroll software and it was legally blessed.

It was such a frigging song and dance trying to get anything done that it was cheaper and faster for the company to emulate their butts off rather than trying to go through the management and the unions and the employees.

But I did learn about optimizing instruction fetches by scattering the compiled code around the circumference of a magnetic drum so that the drum would have rotated around beneath the read head in time for the next instruction.

Try and tell that to the young people of today, and they wont believe you, eh Obadiah?

Re:IBM's been doing this for-ever, dude.

[ZorinLynx]

Isn't the System/370 backwards compatible with the 360? Why would it need to "emulate" a 360 when it can just run the 360 code directly?

Just curious because I recall reading that even the latest zSeries systems can natively run code dating all the way back to the original System/360 models.

-Z

Re:IBM's been doing this for-ever, dude.

[tuomoks]

Hi, yes and no. 370 runs 360 code but, as too often even today, people coded to bypass the OS. Old devices, drums, paper / magnetic card readers, terminals, channels, etc. Even todays systems have the idea, VM especially, of 80 column cards, punches, readers, etc and if used correcly they work wonders, trust me, 360 architecture is one of the best even today. The problems is that not too many people any more want to learn the basics, i.e. Priciples of Operation ( any 3xx, a good book to read, required reading, IMHO ). Search on which OS version macro libraries Linux ( 370 HAL ) was first compiled on 360/370, you will be amazed. Emulations in 360-xxx mostly mean address space differences ( 24/31/32/64/.. ) and some added machine code / functionality, done by OS/hardware. And of course a long time trapping the floating point was/is(?) one if you didn't have the fp hardware installed.

Re:IBM's been doing this for-ever, dude.

[CodeMunch]

But I did learn about optimizing instruction fetches by scattering the compiled code around the circumference of a magnetic drum so that the drum would have rotated around beneath the read head in time for the next instruction.

Mel?? Is that you?

Re:IBM's been doing this for-ever, dude.

[mstahl]

But I did learn about optimizing instruction fetches by scattering the compiled code around the circumference of a magnetic drum so that the drum would have rotated around beneath the read head in time for the next instruction.

Mel? Is that you?

Re:IBM's been doing this for-ever, dude.

[geezer nerd]

But I did learn about optimizing instruction fetches by scattering the compiled code around the circumference of a magnetic drum so that the drum would have rotated around beneath the read head in time for the next instruction.

Given the description of a 360 emulating a 1401, I find this comment a bit difficult to follow. IIRC, none of the 370, 360, or 1401 was a drum-based computer, and the code run in each would not be optimized by consideration of location in memory.

I do remember stories back in the old days, of 360 emulating 1401 emulating 650 (an even older machine). The 650 was a drum machine, and relied greatly on SOAP (Symbolic Optimizing Assembly Program) to develop assembly language programs which were then allocated on the drum surface in optimum manner. (The 650 had the one + one addressing scheme of operand address + the next instruction address) I first learned programming writing assembly code for the 650 in my freshman "Theory of Computing" course at Carnegie Tech (1960). Ahh, those were the days! We were all taught to visualize how the machine worked in detail as we developed our programs, and when they sprung something called Fortran on us in the second semester, I found it very difficult to transition to those funny programming constructs.

And, back in those days, writing self-modifying code was a standard technique -- one which I have long-since abandoned for all the trouble it can actually cause. Real programmers debug using octal dumps! (I spent a lot of time with CDC machines, and they were octal based, not hexadecimal).

But I have to say the overall best programming experiences I ever had in 40+ years of activity have been using Java. I really, really like Java, more for all the wonderful shared libraries of useful code than anything else. What a tremendous boost to productivity! I remember when one had to reprogram every littlest algorithm to suit the environment one found oneself in, and nothing could be shared.

Re:IBM's been doing this for-ever, dude.

[TheLink]

"I really, really like Java, more for all the wonderful shared libraries of useful code than anything else"

I find Java requires just too much work for me. Sure it's faster nowadays, but still a lot more work to do simple stuff - maybe it's the sort of stuff I do. I prefer Perl for most stuff. Lots of decent modules around, e.g. Net::RawIP, Net::DHCP::Packet, LWP. One day I might even use Acme::Bleach :).

I should try to get better at Python though.

single points of failure

[192939495969798999]

Now when something goes wrong, 133 server apps go down all at once! I know, Linux is stable, but a machine hosting 133 apps just sounds like a recipe for a molly-guard type disaster.

Re:single points of failure

[ScrewMaster]

Nah ... Linux is just the icing on the cake. There's plenty of real mainframe meat underneath.

Re:single points of failure

[Strider-]

Now when something goes wrong, 133 server apps go down all at once! I know, Linux is stable, but a machine hosting 133 apps just sounds like a recipe for a molly-guard type disaster.

These are machines that don't break, period. We're talking the types of machines that run the major banking systems of the world and the like. They simply do not go down. In this situation, if one of the 133 apps buggers up, it's only that VM that's shot. You just nuke it and restart it, the rest of the machine just keeps ticking along.

Re:single points of failure

[rascher]

Not quite. They are engineered (as they have been for decades) for stability and were designed to handle that kind of load. Its CPU/RAM/storage are redundant, so that if something in the system goes down, new resources are allocated. Additionally, shops will have multiple mainframes just for that kind of redundancy. Its kind of like saying your car is a "single point of failure" - sure, it is, but they were engineered for the purpose of being reliable.

Re:single points of failure

[Anonymous Coward]

The power grid is what is more vulnerable. I'm sure they have backup generators, but knock out the main utility feed and the generators, and the system will last only a short period on UPS.

Re:single points of failure

[freeze128]

While I agree that IBM's mainframe systems are rock-solid (and, as a colleague is fond of saying, self-healing), accidents *DO* happen. I'm sure the mainframe is happily running its code just fine only seconds before a hurricane rips the roof off of the data center and hurls the machine into the next county....

It's those kinds of things that make disaster recovery necessary. If the apps were distributed across discrete servers, its possible that not all of them would have been destroyed. Remember the end of Twister? The barn was wasted, but the house was left intact.

Re:single points of failure

[Anonymous Coward]

But with the System z mainframe and Geographically Dispersed Parallel Sysplex (http://www-03.ibm.com/systems/z/resiliency/gdps.h tml), you could have your hot backup site located something like 300 km away, with minimal data loss (e.g. in-flight transactions).

And I think your disaster recovery example is a bit far-fetched ... if a facility is damaged, work will be moved to the backup facility. Would anyone really try to sort through the rubble to see which servers could restart and then see whether what came up represented a coherent configuration that could actually perform useful work?

Re:single points of failure

[unsubscribe]

Last week, I attended a presentation at IBM's Australian Development Lab in West Perth, where a lot of the z/OS-related code is maintained and developed.

From what we were told, IBM z/OS mainframes are the *most* reliable platform to host software services (but of course, they'd say that).

The following is from memory, as best as I can remember it, and may not be 100% accurate:

The 'z' in 'z/OS' stands for 'zero downtime'. z System mainframes are engineered for 99.999% availability, or less than 3 minutes of downtime a year (we were actually quoted 'less than 5 minutes', but (1 - 99.999%) * 365.25 * 12 * 60 = 2.63). Apparently, they quite easily meet this requirement - we were told that it is not uncommon for systems to remain online for 10 years or more without failing.

Up to 32 z System mainframes can be clustered in a 'sysplex'. Each mainframe is divided into several LPARs (Logical Partitions), each which can host several VMs. If an application fails, the automated recovery service will attempt to restart it, either on the same VM, a different VM, a different LPAR or a different mainframe in the sysplex, as appropriate in the situation. It is also possible to host a redundant sysplex in a different site, which mirrors data and which the primary sysplex can failover to in the case of failure.

IBM mainframes are used in many major corporations around the world, particularly those where the cost of downtime is very high (think thousands of dollars a second).

Re:single points of failure

[shmlco]

How is that different from the hurricane ripping the roof off of the data center and hurling 400 blades into the next county? Or as we saw in SF, a power outage taking out an entire data center?

Re:single points of failure

[unsubscribe]

Edit: Error in my calc - there's 24 hours a day, not 12.

(1 - 99.999%) * 365.25 * 24 * 60 = 5.23

Re:single points of failure

[SEE]

Yeah, if somebody hits the Big Red Switch, there's going to be a problem. But, if they don't, well, it's a mainframe.

The Linux on these machines is running under z/VM, in multiple virtual machines. When one of them has a software fault, you reboot that one's VM and keep going; the other 132 Linux-running VMs run without noticing anything happened. (It is possible for z/VM to fault, sure. But it's an OS with 40 years of refinement in the "100% uptime" mainframe culture, and its task is just managing the virtual machines.) When something goes wrong with the hardware, the fault tolerance and self-healing features keep things running, and you fix the faulty element with a hot-swap. A properly set-up datacenter is going to minimize external risks, with backup power and such. Proper choice of datacenter location will minimize natural disaster risk.

So, yeah, the big risk is human failure, and these IBM-built, IBM-owned datacenters are presumably going to have extensively trained IBM-employed mainframe personnel, which minimizes that risk.

Now, if some cable company cuts the fiber optic lines . . .

Re:single points of failure

[192939495969798999]

I know they don't break, that's why I said molly-guard type disaster. That means if someone turns the box off, it is like turning off 133 computers all at once. I'm not saying it happens much, but BRB's do get hit.

IBM shouts -- Yay Linux

[velophile]

Why is IBM so pro Linux over the last few years? I'm not going conspiracy with this, I'm just curious what their motive is. I mean, excuse me for not tossing full faith into trusty `ol "Big Blue", but for some reason every time IBM chants the virtues of Linux I cringe slightly. My guts just twist like a compass pointing due south. I don't think anybody on /. is naive enough to think IBM supports Linux for the same reasons most folks here do, but their exact motive remains a mystery to me. The question isn't so much what they gain but what Microsoft (or whomever they view as their direct competitor) loses. While there is no doubt IBM is a good friend to have, it's a friend I look to with a forced smile.

Re:IBM shouts -- Yay Linux

[ScrewMaster]

I think a good part of that can be attributed to their being able to run one more or less standard OS across dozens of disparate systems. The real question is why they picked Linux vs. a proprietary design or some other open source OS.

Re:IBM shouts -- Yay Linux

[Anonymous Coward]

Because they can make a fortune in consulting since nothing in Linux works out of the box. They'll customize to your heart's (and wallet's)content and laugh all the way to the bank.

Re:IBM shouts -- Yay Linux

[Anonymous Coward]

IMHO it is IBM's only way to Justify z/VM or z/OS. Without Linux who would even know that these operating systems existed? z/VM and z/OS isn't in the vocabulary of most /.ers. Mention the word mainframe and people think of old dinosaurs with ascii terminals. Mention Linux on a mainframe and it gives new life to old concepts. The only problem is one has to upgrade a mainframe's hardware in order to take full advantage of Linux. But go and tell a manager you want to pay through the nose for new mainframe hardware and see the look you get. Sure, IBM is doing it but is anyone else?

Re:IBM shouts -- Yay Linux

[Chandon Seldon]

IBM pushes Linux because as a hardware vendor / service provider they make the most possible money in a world with a commodity standard OS with no licensing cost. They put PR resources towards hyping Linux so that when they say "and this awesome system we've built you runs Linux" the PHBs say "awesome, that's the gold standard of operating systems" rather than "why are you using that hobbyist crap? You're IBM".

Re:IBM shouts -- Yay Linux

[Ant P.]

Maybe they just got sick of dealing with other asshole corporations.

Re:IBM shouts -- Yay Linux

[cyphercell]

IBM is a hardware company, always has been, they've been into open source software since before GNU existed.

http://en.wikipedia.org/wiki/SHARE_(computing)

Sure, they are an evil corporation with too much money on retainer, but they realized long ago, that software has an intrinsic value that crashes once the software is written.

For instance, the labor theory of value - the most influential of the intrinsic theories - holds that the value of an item comes from the amount of labor spent producing said item.

http://en.wikipedia.org/wiki/Intrinsic_theory_of_v alue

Basically, once software is written, it's value rapidly approaches zero, because the ability to replicate that work is well within the skill levels of the neophyte. IBM conceded the value of software long before Bill Gates came around floating the idea that the value of software could be upheld by government interference, essentially creating a new fiat currency, and entering the business of printing money, they hired lawyers to back it up along with becoming an extremely predatory business entity.

• Microsoft entered the business of brokering software copyrights
• IBM began brokering in software patents (primarily) and copyrights
• RMS decided to rewrite some copyrights and lay the legal groundwork for open Intellectual Property
• Sun Microsystems and Apple Computers are two companies that managed to survive IBM and Microsoft, respectively
• Every company that hasn't been purchased or destroyed by IBM or Microsoft has started moving towards FOSS, Apple and Sun are stragglers.
• Linus is an apolitical hacker, generally happy with GPLv2
• Red Hat found that the intrinsic value of software rests with the people that know how to use it, ie support
• IBM is looking at this situation and realizes that their business model was built on OSS, if they remove Microsoft they have a good shot at Dell, HP, Gateway, then finally Apple and Sun.

While IBM may have quite a bit to lose going the free software route they have a lot more to gain. Once they own all the copyrights/patents they can do whatever they want and that (currently) includes GPLvX or greater.

Ok. But How?

[Anonymous Coward]

I have been hearing more and more about all this grid computing, consolidating, VMs stuff. Please forgive my ignorance, but how would you do it? is there open source software to do it? would you stuff 10 VMWare/QEmu VMs in one server with lots of memory and hard disk? or you create a beowulf cluster and run VM's on it?

A Small Request

[Nom du Keyboard]

Can I have the old ones?

Re:Imagine....Need To Update...

[Nom du Keyboard]

Imagine a Beowulf cluster of those!

Don't you mean a virtual Beowulf cluster?

You've got to get up to date. After all, this is 2007 -- the year of Linux! (or something)

A very confusing endeavor for us

[DutchSter]

Where I work we currently run two mainframes in a sysplex environment for all the core transactions. It's a very optimized environment and handles millions of financial transactions a day. In mid-2006, IBM started giving us zLinux engines to "try out" and they gave us all of the software we needed to make a go of it. Kind of like a playground drug dealer, they hoped that by giving us a bit for free we'd get hooked and become dedicated customers. The problem was, for the type of workload that typically runs on our servers (high CPU, moderate I/O) we were experiencing poor performance on the mainframe VMs. IBM sent all their engineers out to help make tweaks and tune all sorts of things. Despite all the tuning and tweaking that took place, we could never get a single engine to perform better than a $5,000 server. Keep in mind that a single engine was retailing for around $80,000 after discounts.

We did some calculations and determined that for the price of a zLinux engine we could buy an entire rack of high-end HP servers that would outperform the single engine by a factor of 200:1. Again, maybe it was just the workload we were doing, but even IBM couldn't figure it out and our server work profile isn't exactly uncommon. Granted you can cram a lot of guests onto a host system provided that none of the guests want to use more than 10% of their CPU at any given time, but that defeats the purpose. I could probably run a VMWare host with 100 guests and call it a success, provided they all sat idle.

It was kind of funny because the IBM engineers would shake their heads and admit that for our workload it just wasn't going to work out. Then the next week the sales guy would call and ask if we were ready to buy that third mainframe since he just read the engineer's report and our visit was obviously a smashing success.

I'm not knocking the whole Linux on the mainframe concept, I'm just sharing our experience and how the whole thing seemed to be like someone in IBM Marketing declared "we need to sell Linux on the mainframe" and the Dilberts were forced to sell a product that worked about as well as a chocolate fireguard. It was a very awkward experience and even the IBM engineers seemed like they were stuck in an uncomfortable position of supporting sales for a product that even moderately demanding customers wouldn't be able to run with.

Personally I consider Linux on the mainframe to be on par with running Linux on an iPhone. Sure you probably can, but does it actually do anything uniquely useful for the business? I have a hard time selling technology to the CIO on the grounds that because it's Linux it's a good business decision regardless of the context.

Re:A very confusing endeavor for us

[1lapracer]

The blue team was in the other day and I sat through the entire 2 hours of how much money running Linux on the Z would save us. It sounded great on paper. As I was leaving with the AIX guys they could barely contain themselves so I asked what was so funny. They gave me the summary of how the last time several years ago we tried this with results that were similiar if not worse then the perivous poster. We would have spent many many millions to run our p series and or the i series servers. I'm sure we will take another shot at this but, even as a Blue supporter, buyer beware.

Re:A very confusing endeavor for us

[Anonymous Coward]

As far as the Mainframe goes it's not the CPU Performance but the IO and alot of the IBM guys will tell you that the mainframe just extremely fast with IO. Sure a single Linux guest cannot outperform a $5000 box but again the aim for the mainframe is to decrease total TCO (power maintenance and what not).

I have a few (>6) Linux guests running and all of them have the same performance regardless if the load is high on all of them and trust me on a box just running VMware even with multi core cpus (w/ VT) and gigabytes of ram fails with performance (context switching and more) because the hardware just isn't designed for what the software is trying to do.

Another factor about the mainframes is that it is designed for zero downtime. Depending on Business or Enterprise Class, you will get probably 30 - 50 something Dual Core CPUs (which all are not active) in a mainframe chassis. You'd probably think, "Woah 50 dual core CPUs that's like 100 CPUs," alas all those CPUs are not online and those dual core CPUs just pass the same instruction through both cores, and compares the result of both instructions. If they match it goes on to the next thing it needs to do, if it doesn't the CPU is varied off-line, it brings another CPU up, and the machine phone's home to IBM complaining about a broken CPU. By that time you may get a call/visit from an IBM tech about the CPU. With that all those CPUs have probably 40MB of L2 cache shared amongst the host of CPUS so bringing up another CPU to process the data another CPU failed at gives you quicker results than having things copied from memory to L2 again.

Not that I work at IBM I just know a bit about the newer Z mainframes and I usually find if you can save your company a few dollars a quarter and tried to change the way you do things for the hardware you have, you usually end up in a winning situation (unless you work for scumbags).

Go with the solution that is both cost-effective (over time) and provides a decent performance.

Re:A very confusing endeavor for us

[timeOday]

Personally I consider Linux on the mainframe to be on par with running Linux on an iPhone. Sure you probably can, but does it actually do anything uniquely useful for the business?

Is it really linux on the mainframe that was the problem, or rather virtualization and IBM pricing? In other words, if you ran one instance of linux hosting 1,000 server processes, might it not work well?

If so, linux might have advantages over a special-purpose mainframe OS. Presumably there are more apps for linux and mainframe developers are harder to come by. (Or does Java make the host OS irrelevant anyways?)

Re:A very confusing endeavor for us

[phoebe]

Isn't this the entire crux, extreme reliability costs a lot of money and resources. Yes, a single node is slower than a regular server but you can also run many nodes in parallel: so you are only limited by software design.

To implement remotely similar reliability with regular hardware you are going to need a redundant SAN, redundant switching, redundant NICs, redundant CPUs, redundant memory, etc, and a lot of cabling. Running something like VMWare ESX will allow you to bump VMs in realtime between hosts at the expense of licensing and VMware preferred hardware, but you would have to be lucky enough to have a predictable hardware error in order for migration to proceed. How likely is that, and are all regular server components fully fault tolerant like a mainframe? Hence VMWare HA, it will restart your VM on another server automatically, but you still suffer downtime.

If you take a moment out to think about it, this is also a software design limitation, demanding each node has a high uptime, ideally the software should be distributed and be able to recover for node failures.

So the choice of platform depends on how clever your software is or how much downtime you are willing to accept.

Re:A very confusing endeavor for us

[dintech]

I'm not knocking the whole Linux on the mainframe concept, I'm just sharing our experience and how the whole thing seemed to be like someone in IBM Marketing declared "we need to sell Linux on the mainframe" and the Dilberts were forced to sell a product that worked about as well as a chocolate fireguard.

I hope you're wearing that chocolate fireguard now because there's a lot of flame coming your way. :)

Re:A very confusing endeavor for us

[einar2]

We spoke with several bigger companies in Switzerland who made the same experience you did. You install linux on the mainframe and the system behaves like a very expensive Unix box.

There is nothing magical about mainframes. They are build for a special task (transaction processing) which is rather IO intensive. A transaction request comes in, a small processing footprint touches the database and out goes the reply. Due to the characteristic of this processing style, you can run such machines under high load using 80% of their capacity. If you install linux on them then they behave like Unix machines: idling around at 10-20% to have enough resources left for the next spike of processing.
Virtualization helps to distribute the load better. However, this has nothing to do with the mainframe. It is the ability of virtualization itself.

Re:A very confusing endeavor for us

[RedHat Rocky]

Again with the mainframe diatribe. TCO! Uptime! Scorez!

My environment involves running several "guests" to do basic services (DNS, print, web). Here's the deal:

1. zVM is really really expensive. And I mean just the licensing. We started on it just like the OP, IBM came along handing out the crack for free.
2. Some workloads work okay, some don't. And no, it's not obvious which is which.
3. Mainframe people are just as expensive as the Mainframe and usually don't have the depth to handle "guest" issues.
4. All eggs in one basket. Guess what, upgrading zVM is an outage. For ALL guests. At the same time!
5. Go to any mall and buy memory for a PC, it's cheap. Mainframe? Riiiight. Lemme cut a PO.

Re:A very confusing endeavor for us

[metamatic]

Mainframes aren't designed for CPU performance. They're designed for reliability and I/O performance.

If your workload is high CPU, low to moderate I/O, then I'm guessing what you need is a Grid, not a mainframe.

(Opinions mine, not IBM's.)

Re:A very confusing endeavor for us

[Anonymous Coward]

You don't get it, it's designed for 0 downtime. Everything is hotpluggable - CPU, RAM, DISKs. Need to upgrade the OS, oh sure no problem create another LPAR move your running guests over to it. Bring down the Lpar you want to upgrade and done. Rinse and repeat.

Re:A very confusing endeavor for us

[RedHat Rocky]

You don't get it.

$100,000 a year is $100,000 a year too much.

Re:A very confusing endeavor for us

[Anonymous Coward]

We had nearly the exact same experience (large financial services company with 5k+ servers). We went so far as to compare a VMware configuration and a Xen configuration with the same on Z. Both VMware and Xen performed and scaled significantly better, with a much better TCO. Sure, the Z hardware had some capabilities that couldn't be replicated on x86 hardware today. But down the road, who knows. Bottom line - a $15,000 x86 server crushed the $100,000 Z processor. We were left wondering exactly what workloads do run well on the IFL when consoidated.

Re:A very confusing endeavor for us

[Hugo Graffiti]

Just out of interest, what language was your workload written in, was it Java by any chance?

And another thing ....

[o2sd]

8 million square feet for 4000 servers??!!! WTF? That's 2000 sq.ft per server.

Why would a server need a 3 bedroom apartment to run?

Re:And another thing ....

[freeze128]

Why would a server need a 3 bedroom apartment to run?

To house the support staff!

Re:And another thing ....

[flyingfsck]

It's the office space required for all the MCSEs, Tape Trolls and their Foozball tables, plus the accompanying Hell Desk for all the support people with the requisite funny accents. Backing up, patching and rebooting 4000 MS servers is a huge job...

Re:And another thing ....

[Anonymous Coward]

no, no, no that's the donut storage area! Obviously, the support staff lives in the basement =P

Re:And another thing ....

[o2sd]

...all the support people with the requisite funny accents

Oooh, those mid-west accents are required .... always wondered about that.

OS/2

[Anonymous Coward]

Because after OS/2 died IBM didn't want to get fucked by Microsoft again and Apple didn't have a business solution. Creating another ground up OS was definitely not a solution so they were left with two options: Linux and Unix. The only reason Linux is run on as many servers as it is today is because of IBM. They do Linux R&D, produce software, and provide B2B services for the OS. Linux is the official OS of IBM and without their support Microsoft would currently have a monopoly on server software. So, a simple answer to your question would be that IBM has a financial interest in Linux. But it goes deeper than that: IBM has committed to Linux - it's a marriage.

I think you're hesitant to accept IBM because of the whole 70's/80's "Big Blue" stuff, but after Microsoft swept the rug out from under their feet the company's strength was permanently compromised. The consumer market rejected them (hence the sale of the PC division to Lenovo) and until they committed to Linux software was a major vulnerability for them. The openess of Linux enabled them to get back in the game - their customers didn't have to worry about the future of the platform while their immense contributions to Linux enabled the OS to really threaten Microsoft. So yeah - as a Slashdotter, IBM are the good guys. They support Linux and they don't aggressively protect their many, many patents (they use their patents to protect themselves rather than trying to sue everyone they can for $$$). Personally, I think IBM is the most important tech company in the world.

Save space with project Black Box

[HOTTILA.COM]

http://www.sciam.com/article.cfm?chanID=sa006&arti cleID=B1027B68-E7F2-99DF-352186A04761EB7F&ref=rss You can form a cube and save space...

It still doesn't make sense

[Anonymous Coward]

If it takes 8 million square feet to house 4,000 servers, that's 2,000 square feet per server.

Re:It still doesn't make sense

[dwarfsoft]

The extra space is for sufficient ventilation for the sweaty Sysadmin. Nobody wants to go near him.

It's just an endless cycle

[ogren]

I still haven't seen any conclusive evidence that Linux on mainframe is a good idea. I'm sure running 30 new mainframes is going to cost less than 4000 aging servers. Just about anything would be less expensive than 4000 aging servers.

But I bet that a small farm of modern medium sized servers running Linux on VMWare would be even less expensive. Or Solaris/Niagara. Why would you want to run an open source operating system, whose major benefits are openness and affordability on the what is literally the most expensive and most proprietary computing platform in the world!

These server consolidation projects are just giant boondoggles spawned because the server sprawl finally got insane. It's an endless cycle:

A. Giant server consolidation project that takes 4000 servers down to 30 servers.
B. Department B complains that Department A's application keeps hanging and consuming all of the CPU. They demand their own hardware "for availability reasons".
C. Vendor C demands dedicated hardware for licensing/capacity planning/supportability reasons. Rather than constantly bicker with the vendor over supportability they get dedicated hardware.
D. Department D complains that the IT department is charging outrageous prices for time sharing on the mainframe. After all a dedicated server only costs $XXX.
E. Suddenly there are 4000 servers again.
F. IT department spends some insane amount of money on infrastructure to manage the 4000 servers.
G. IT department budget gets insanely large trying to manage that much stuff.
H. Some CIO gets the idea that all of this money managing servers is ridiculous and we should do a server consolidation project.
I. IT department spends an even larger amount of money on the latest super high availability gear and consulting services so that the can run 4000 commodity servers inside a few big servers. All because it will "cost less to maintain".
J. Go back to A.

Re:It's just an endless cycle

[rainhill]

Well said, u spoke like you'd been in my shoes.

Re:It's just an endless cycle

[Mxyzptlk]

From reading what you've written - it seems that you are working in my department, but I don't recognize your name - strange ;-)

But seriously, this consolidation loop you are referring to is a bit like outsourcing trends: the higher managers can only see the cost of IT, and not the benefits, and consequently have a hard time justifying why they should keep their jobs to the top level management. By consolidating, they create a smoke cloud that hides all the costs - they are instead spread out over other departments - and get a pat on the back from top level management. After a while, the projects demand their own hardware, and the pendulum swings back for yet another go...

The root cause is that it is easier to describe the costs of IT in terms of money, than it is to describe the benefits; the benefits are much more fuzzy and long term, so when it comes to making a quick and easy decision - they choose the short term easy win.

Re:Imagine....Need To Update...

[kriebz]

actually, I think we need a Parallel Sysplex of these.

4000 servers doing nothing

[flyingfsck]

What bugs me is why IBM had 4000 servers that were evidently doing nothing?

In my experience, consolidation using virtualization only works if the servers in question don't have anything to do and only runs a zoo of defunct web sites for example.

Six 1.2 million square foot data centers.

[twitter]

The article says that the data centers required for the 4000 "small computer servers" aggregate to about 8 million square feet. It takes IBM 2000 square feet to house a small computer?

It's less amazing when you think of it as six 300x400 yard warehouses run by clients. Those are big buildings but "data centers" are usually large. There are not enough details from the article to figure it all out but four thousand computers to 30 boxes is an impressive feat that will save lots of electricity.

2003 Server, Dude!

[twitter]

Lets see... $250M / 4000 = $62,500 per server being consolidated? I mean, I know floor space, buildings, racks, power, AC etc cost money... but that's still a *lot*. Anyone care to chime in on how close to normal that is?

Running 2003 server, each needs an "admin", licensing fees for software, you know it gets expensive.

Really, I think the cost was in 130 football fields of mostly unused server warehouse.

Re:6 Locations

[Dare nMc]

"location" means a building with

Translation, linux is responsible for 5000 jobs lost. All of a sudden B.Gates might have a point about that sucking sound.

125 Acres/6 =20 acres. so 20 Acres of office space would likely require a good 600 employees, correct.

seriously though, freeing that many physical resources for more productive uses sounds like a win/win.

1920 square feet per server? WTF?

[jerryasher]

140 football fields for 4000 servers? That's about 30 servers to a football field, or about 40'x40' per server.

Google tells me that a football field is 300'x160' or 57600 square feet.

If every server had 14 square feet you could put all 140 of them on one football field.

140 Football fields

[thegoofeedude]

"The six data centers currently take up over 8 million square feet, or the size of nearly 140 football fields."

Yeah, but how many Libraries of Congress is that?

Failure: A cautionary tale

[Maxo-Texas]

Before, .33% failure rate = 13 failures a day. You had well understood procedures for dealing with failures.

After, .33% failure rate = 1 failure per thousand days. This is a recipe for hell.

But wait...
When you do have one machine fail- it takes down 133 virtual servers at the same time. You raised your risk enormously.

IBM will tell you all about fail-over just like they did our executives.

Half the country down for three days is the reality.

---

Still it is interesting to see a return to the centralized mainframe farm. Sure hope those multiply redundant communication lines don't go down.

Re:Failure: A cautionary tale

[Anonymous Coward]

You think IBM has wasted its decades of experience in the field of ultra-high-reliability? You think they don't have answers for every one of the issues you raised, and then some? Go read the other postings on this thread, such as this one. Then you will begin to realize how stupid you look.

Re:Failure: A cautionary tale

[Maxo-Texas]

Mr. AC.

The three days I'm talking about are reality. I lived through it. IBM completely failed at their promise of zero downtime failure. We had to make heroic efforts to cover their failure.

Re:System z Mainframe Specs

[Anonymous Coward]

There are a few reasons why the specs for mainframes are so hard to find.
One is that the things you find on IBM's website are designed for CEOs and CIOs who don't really care about technical details -- only "solutions"
The second is that the specs themselves aren't well-defined. As an earlier poster pointed out, you don't buy one of these things off the shelf. You tell IBM what you want to do with it, and you work with them to construct not just a mainframe, but all of the storage and other add-ons that come along.
And finally, the third reason is that the specs don't line up with anything you likely work with normally... (If they did, you'd know where to find them.)

Here are some specs for the z9 Enterprise Class:
http://www-03.ibm.com/systems/z/z9ec/specification s.html

Simplified you are looking at 54 CPUs with 512GB of memory.
The CPUs themselves are basically Power6 processors, but thats really simplifying everything down.

Each CPU is actually a "book" of CPUs. Several run at once on the same data. If any disagree, the instruction is rerun on a different CPU. Entire backup books (in addition to the 54) kick-in if a problem is detected.

Additionally, the z/Series comes with a bunch of "Specialty" CPUs. You can get 27 CPUs that do nothing but process Java work natively. Or ones that handle DB2 workload. Or even special processors optimized for the linux kernel. Oh and don't forget the built-in hardware crypto CPUs.

Memory and I/O and Power and everything else works pretty much the same way on a mainframe. And all of it is hot-swapable. (Even the Emergency Power Off switch can be replaced while the system is running).

The hardware specs are impressive, but the biggest deal about these boxes is that they don't go down. Most people I talk to question the idea of consolidating servers into one box because of "single point of failure" concerns. This is where the mainframe shines. These things have MTBF of decades, and will just churn away forever.

Big Iron. Right concept, wrong platform.

[HockeyPuck]

I work in a longtime "blue shop" perspective, not always from a software/OS perspective. While I like the 'concept' of running linux on zSeries, I think you could take a look at the requirements and choose a platform that can run the same Apps.

For example, for email we run Lotus Notes on a couple of BIG pSeries (AIX) servers. We could have run it on a farm (technical term) of windows boxes.

For webservers, which you could run on AIX, or linux on zSeries. We have multiple (read: many) x86 servers running linux+apache. Why? They connect to a backend app server (pSeries) which connects to a backend zSeries DB2 (I'd prefer Oracle however, to run Oracle on zSeries requires it to be run in a linux VM).

We definitely subscribe to the school of using VMs whether they are zSeries, pSeries, or VMWare on x86. Even if the x86 server is running ONE application, we still put vmware underneath, as it allows for us to move the image to a newer hardware platform when it's time to upgrade. Even some of the larger x86 servers run vmware but in each partition there is a single instance of apache. Makes for managing storage that much easier (fewer zones, cabling etc).

Would I consider moving our apache on linux on x86 to apache on linux on zSeries? Not really. It's a waste of CPU cycles (MIPS). I'd rather use zSeries MIPS for something a bit more critical like keeping my database up and running than serving out webpages (static or dynamic).

IT isn't not about religion, it's about finding the best tool considering your requirements. I have no problems telling IBM that product XYZ is trash. While my servers are IBM, you won't see IBM disk, or IBM tape, and atleast once a quarter some salesman from IBMs storage group is at my door. He buys me lunch and every quarter he is sent packing. You won't see ibm bladecenters as the thought of hundreds of additional servers to manage isn't appealing (but I'll gladly take 100s of VMs across larger x86/pseries boxes).

I know many of you were expecting to hear me say 5000 linux servers, but there are options for my requirements that did not lead to big "google style" linux farms.

BTW: I have no problems kicking out IBM on x86 if HP/Dell/Sun have a better product, and knowing this and letting IBM know this gives me a great advantage over them, as they very well know I'm capable of bringing in something more suitable. (I *used* to have IBM storage).

Commisar mentality

[Anonymous Coward]

Who cares if their motives aren't as...."pure" as yours? They're using it because they deemed it the best tool within the parameters they selected. That's it, 'nuff said. They're no fanboys.

Re:System z Mainframes - tech details

[Anonymous Coward]

Try looking at the following redbook: http://www.redbooks.ibm.com/redbooks/pdfs/sg247333 .pdf. It gives more technical details on the underlying hardware (see chapter 2).

Those of us who do software development for z/OS have access to some pretty impressive 'screamer' hardware.

Consider this...

[Krozy]

Yes, 4000 "small computer servers" times 2000 square feet equals 8 million square feet. But this is unlikely the arrangement. Consider instead a few buildings of data centers, each with 1 or more relatively small rooms. Within a room, there may be a few racks, all surrounded by walking space, and other perhipherals like AC units. Then outside of those rooms, more walking space for hallways. When you factor in all the human space and simple space for ventilation, and then cubicles and monitoring for support personnel it could average around 2000 square feet (40x50).

The same logic can be applied to costs.. $250 million / 4000 machines = $62.5K. Some of that is actual hardware, and software licenses. Some of that is ongoing support from their full time employees on staff to maintain the things.

Re:Wait till 2008

[Anonymous Coward]

Next year is the Year of the Linux Desktop.

Again.

WARNING! THIS GUY WORKS FOR SUN!

[Anonymous Coward]

Warning - this guy works for Sun.

What was before?

[DavidD_CA]

The headline and article imply that they are switching to Linux (among other things).

What operating system were they previously using?

Re:What was before?

[simong]

Probably a mixture of AIX, Solaris and Linux with a sprinkling of HP-UX thrown in. IBM are nothing if not eclectic and indeed pragmatic about the services they provide.

Re:What was before?

[Ilgaz]

I am sure Windows server flavors are involved too. IBM is "Big blue"- they take decisions based on the job, not whoever makes it or codes it. While they were advertising OS/2 like crazy, I have seen lots and lots of Windows clients in their buildings doing stuff. That was the day I started to question future of OS/2.

Re:Where the FUCK is iLife '07???

[Anonymous Coward]

It's just for cool people this time. Sorry, buddy! Just use Outlook like the rest of your sub-average looking friends.

Re:Big Iron. Right concept, wrong platform.

[curious.corn]

What did you get to replace the Shark? Just curious... not Hitachi, I guess... e

Recycling!!!!

[axia777]

My favorite part of the article was this:

"The 4,000 replaced servers will be recycled by IBM Global Asset Recovery Services. The six data centers currently take up over 8 million square feet, or the size of nearly 140 football fields."

Recycling 4,000 servers? That is awesome. I am glad that a major corporation like IBM an take on the idea of recovering such a large pile of servers. Making more new servers from old servers? Now if every corporation on the planet would do follow suit and do the same with everything ever produced.

"Football fields"?

[darnok]

"The six data centers currently take up over 8 million square feet, or the size of nearly 140 football fields."

I suppose when the US finally goes metric, they'll have to deal with units of area such as "millifields", "centifields" and "kilofields". In time, the measure will have to formalised e.g. "the distance a 100kg, 190cm man is able to kick a leather-encased rubber bladder...".

Or maybe the current generation of writers that thinks "140 football fields" is a meaningful substitute for "a really big chunk of space" will have died off by then.

Re:"Football fields"?

[badzilla]

Does anyone have the conversion factor from Football Fields to Libraries of Congress?

Re:"Football fields"?

[DigitalSorceress]

I've always hated the media's tendency to measure things in football fields {area} (American football of course), Football Stadiums {numbers of people and/or volume} (again, American football), Olympic swimming pools, Empire State Buildings, Statues of Liberty*, feet of water covering a given US State, and trips to the moon.

Ok, maybe I can see that some folks would have a problem understanding volume (because our poor educational system means that we are barely able to manage two dimensions), but why must they insist on measuring things in terms of these other objects?

Ahh, must be because our poor educational system means that we really CAN'T manage a two dimensions, and even have trouble with one.

~sigh~

* I for one was quite surprised at the actual size of the Statue of Liberty the first time I saw it in person... I somehow imagined it being quite a bit taller.

Re:6 Locations

[Anonymous Coward]

They aren't closing the sites, they are just making room for other boxes. Certainly in Portsmouth,UK they have spent a huge amount of money recently on the data center part of the site, so they aren't going to close it down any time soon.

OT: Feeding the Troll

[cp.tar]

Well, this has been the first /. flooding I've ever witnessed...

It is rather interesting that you should flood flood like you do, then bemoan cultural intolerance... I participate in a forum where several users (or "morons", as I dub them) demand and exercise their "right to flood", claiming that cleaning up their flood is denying them their right to free speech.

I just don't understand how do you find the time to do things like that...

Re:OT: Feeding the Troll

[yada21]

Rather impressive. I'm wandering how he got round the time limits and also managed to do it without someone else getting one inbetween. Open multiple browsers via different proxy's? Some kind of bot? A group effort?

Re:OT: Feeding the Troll

[cp.tar]

Umm... Torpark?

Re:OT: Feeding the Troll

[Anonymous Coward]

A bot + multiple proxies. Tor is way to slow for this, plus many of the exit nodes are already blocked so that wouldn't work out.

Re:6 Locations

[packeteer]

Such is progress. Unfortunately people lose jobs with almost any form of real progress made. If all the progress we ever made was simply needing more things, and people to work on those things, we would never get anywhere. By cutting the number of people IBM needs it is giving the economy more workers. As you said, this is a win/win situation, the only problem is the short term for the people who lost their jobs.

Right now they are unemployed and this is why we need things like socialized unemployment help. The system will temporarily hurt people by making progress and we as a society need to protect those people until they can get a job again.

The buggy whip industry used to be booming, imagine if we had tried to keep that around...

IBM Saves $250 by playing OSS devs for suckers

[Anonymous Coward]

The savings isn't from Linux itself, after all IBM doesn't have to pay for AIX since it's already theirs.
The savings comes from IBM being able to lay off their AIX devs. It's easy to save money when there's a free labor force that's been deluded into thinking that it's virtuous to work for IBM for free.

No, You Don't Need Different LPARs for RHEL 4 &

[BBCWatcher]

There are a lot of errors in your comments, unfortunately. Of course you can run Red Hat and SuSE concurrently in a single LPAR under z/VM, and multiple versions thereof. This has always been true, ever since Linux began running on mainframes many years ago. You might want to have more than one LPAR to run more than one version of (first level) z/VM, but you don't need many. Two or three for z/VM and Linux is typical and just fine. And it's not as if LPARs are in short supply on mainframes: up to 60 are available on a single machine (30 on the smaller model), so "spending" 1 to 3 is no big deal.

Re: Investing in new mainframes, come on, get real. It's so easy to find market data because companies like Gartner and IDC publish it, and IBM just announced its 8th straight quarter of mainframe hardware growth, something that hasn't happened since before Y2K. It's impossible to do that with "a few showboat customers."

And no, you simply cannot approach the level of virtualization these machines offer on any other system, at least for typical business computing, and still offer reliable service to users. In fact, in IBM's case many of the software licenses are presumably "free," and they still found big cost savings by taking 4,000 machines down to 30. For the rest of the world the mathematics in such situations are even more compelling.

I was that soldier

[simong]

I was involved in a migration to the zOS architecture three years ago. I am currently involved in a similar exercise for a British telecoms company whose name escapes me. In both cases the principle was perfectly sound, but the reality rapidly starts to come down to what can be migrated, when, and why. At IBM application compatibility was a major consideration, and ultimately prevented key parts of the system from being migrated. At the current site, surprise surprise, the problems are the same, plus reluctance to do the work (upgrades, work required on the client's part, age of applications and Plain Old Politics). I wish IBM good luck, and perhaps because there is a better integration of operations and systems they might succeed, but I would be willing to bet that by the end of the process, they will have reached about 80% of their target.

A Little More on Availability

[BBCWatcher]

Yes, System z mainframes are engineered for 99.999+% availability. But it's important to define availability here very precisely. IBM defines this as business service: what the user gets. Therefore, planned downtime is just as bad as unplanned downtime. A lot of IT people get confused by this point, but it's very important. "Excuse me while I shut down credit card approvals for a couple hours to upgrade the database" and you'll be escorted out of the building promptly.

Now, there's nothing in the original story that talks about how the highest levels of availability are important to this particular collection of applications that IBM is consolidating. Indeed, though Linux is mighty fine on the mainframe -- the most highly available Linux -- it's no z/OS, the flagship mainframe operating system. Still, one would hope IBM would include in its cost calculations the cost of downtime. It's likely user interruptions will decrease as a result of moving this work to the mainframe. That means the users will be more productive, and there's a financial benefit to that. How big or how small depends on the applications and the users.

Re:A Little More on Availability

[Anonymous Coward]

So how exactly do you upgrade the database with zero downtime? I mean upgrade to version N+1 of the
database server code.

You transfer the execution context of every session into a new process? Yeah right.

They also saved 20% on their car insurance...

[XnavxeMiyyep]

...by switching to Geico.

Amplification re: CPU Sparing

[BBCWatcher]

Actually, on a System z9 EC (Enterprise Class), a single CPU chip failure is not a "Call Home" repair event. Only the second CPU chip failure would result in an automatic call, while your business keeps running of course. (There are a minimum of two spares in each machine.) The average time to first failure for a particular machine is somewhere in the many decades range.

OK, just for fun (because it never actually happens in the real world), what happens with a triple failure? If you happen to have a "fully configured" mainframe -- all processors turned on -- then.... your business still keeps running. Yes, the system might lose some processing capacity, but it keeps running. The higher priority stuff (from a business view) takes precedence automatically, and life goes on. This is all on a single machine still.

If you've got an S18, S28, S38, or S54 model, then, at your business's convenience, the faulty hardware can be replaced. (You might do this at night, for example.) The repair technician tells the mainframe to "evacuate" memory on a portion of the machine while the OS and applications keep chugging along, possibly with reduced capacity, often not. (Depends on what configuration you choose.) When the evacuation is complete, the technician can pull a processor/memory group (called a "book"), insert the new one, bring the new one online, and... everything still keeps running. Again, this is all on a single machine -- no clusters required for any of this.

Oracle Correction

[BBCWatcher]

Oracle Database is available for both Linux on z and z/OS....

But it'd be extremely unlikely you'd want to switch from DB2 for z/OS. Even Larry Ellison concedes that.

Your sig is broken.

[porkchop_d_clown]

Didn't the ",1" suffix cause the program to autostart?

Re:Your sig is broken.

[Cobralisk]

No, the ,1 will cause them program to load to the same address as it was in when it was saved. Without that argument it will load into the default BASIC location, which may be different for different cbm machines. Generally machine code programs need to be loaded into the correct address. Some programs do have bootstrappers that will cause the program to immediately run upon loading, but it is not default behavior.

Re:Your sig is broken.

[Single GNU Theory]

ISTR that the ,1 caused the program to be loaded at an address other than the default, start-of-BASIC address. You could load machine language programs in there and then use the SYS command with the address of the start of the program. You could, if you wanted, also use the ,1 to load BASIC programs, but they'd end up at the same address as they always did without the ,1. The specific address was somehow stored with the file.

Autostarting was (again, if I recall correctly) done by loading the program with a ,1 into the address of the next instruction to be executed, but I never programmed the Commodore 8-bit computers in machine language.

Only IBM could save money though

[jocknerd]

They don't have to buy the mainframes. If any other company tried to do this, it would cost them $250 million.

It still amazes me

[Anonymous Coward]

You really need to take these articles with a grain of salt. How do you save money when you are buying new hardware? Were they really paying 250$Million in a year on energy? 5 years? 10 years? The thing that drives these projects are the fact that horsepower per square inch goes up every year and ultimately you need to squeeze more computing power into the same space rather than expand the existing data center. And in the long run, that computer room will fill up again because every inch of empty space is begging to run more software at this point.

I also hate how IBM tries to shove big iron in their huge pSeries complexes and zSeries Mainframes and they just don't do everything that everyone needs, they don't always save space and they're expensive per CPU than smaller pSeries machines. I find it hard to believe they're going to completely replace 4000 servers with 30.

mainframe dead, film at 11....

[whitroth]

Every decade, we hear "the mainframe is dead", and a year or two later, IBM announces record sales....

And for those that have no long term memory from watching too many videos and commercials that are cut for the attention span of a three-year-old, let me remind y'all that it was four or five years ago that they maxed out a mainframe running 48,000 virtual machines (using VM) on one mainframe, and easily ran 32,000.

            mark

140 football fields

[Zaatxe]

American Footbal or English Football?

They also saved 20% on their server space...

[Necreia]

... by switching to Gentoo

And at this rate

[Nybble's Byte]

How long before all that computing power can fit into your shirt pocket?
Gimme!

Something tells me...

[LeedsSideStreets]

the "Highly Reliable Times" won't be covering this one.

No disassemble!

[scoobrs]

The 4,000 replaced servers will be recycled by IBM Global Asset Recovery Services.

Global Asset Recovery Services sounds like the same, evil organization which tried to recycle Johnny Five. No disassemble!

Windows on z mainframes?

[Stephen Ma]

Now, my personal opinion is why Linux?

There is no version of Windows for z Series mainframes. If not Windows or Linux, what other operating system (1) runs on the z Series, and (2) is popular enough to be worth consolidating? No, it has to be Linux. Besides, Linux is the hottest thing in the server world now.

I found their problem...

[jgoemat]

Each of their 4,000 servers cost $62,500 more than the equivalent VM running on a mainframe... How the heck do they get that figure? Is it support and maintenance costs? Over what time frame are we talking about saving $62,500 per computer. Many administrative tasks will still probably have to be done on the individual computers (back-ups, user administration, etc.), you would basically be saving costs with people having to work with the actual hardware. Having centralized administration can save some time, but $62,500 per year for each computer? I'm sure the 30 mainframes aren't cheap to buy or run themselves...

wouldn't it have cost less...

[Anonymous Coward]

Wouldn't it have been even more cost effective to use AIX? It's not like they have to pay for the software, and I would imagine that shutting off 4000 servers means they have a LOT of qualified people with spare time to run them.

I know the real reason

[Anonymous Coward]

IBM has often been a company with no viable low end equipment. This caused them to lose a lot of sales to Sun. The original IBM Linux strategy was aimed directly at stealing low end customers from Sun.

Amusingly, a quote I heard (a couple of years after the strategy was started, when it was obviously successful) went along the lines of "we sold three billion dollars worth of mainframes for virtualizing Linux last year; I don't know why anyone would do that, but we sure make a lot of money off of it." I don't know what the driver of the expansion of IBM's Linux strategy beyond the low end was, but I expect it was just running with a good (i.e. profitable) thing.

Metric?

[vidnet]

The six data centers currently take up over 8 million square feet, or the size of nearly 140 football fields.

In metric, that would be around 104 soccer fields.

Mainframes and Virtualization

[BASICman]

Anywhere you see mainframes and virtualization, you're talking about z/VM (http://www.vm.ibm.com/). Its been kicking around since the early 1970s (I remembered it from a mention in one of Tannenbaum's OS textbooks).

I'll bet you that there were a lot of happy people in Armonk the day they managed to get it running Linux (aka: new workloads).

A Peek Inside the Economics

[BBCWatcher]

First of all, that figure is $12,500 per year. Which isn't a lot of money.

OK, let's look at staffing first. Let's assume IBM got a very good ratio of server administrators to servers and could run those 4,000 servers with 100 administrators (40 to 1 ratio, which is very high). Let's assume IBM can get that number down to 30, i.e. one person per new mainframe server. Assuming a fully burdened staff expense of $100,000 per worker (salary, benefits, office space, telephone, ID badges, whatever), that's $1,750 of the savings right there. We have $10,750 to go.

Now let's look at electricity. Let's assume (conservatively) that each old server requires 300 watts. That's 0.3*8760 = 2628 Kwh. At $0.10 per Kwh that's $263 per year. Triple that to include the cooling costs. That's $3.1M in annual electricity for 4,000 servers. IBM says they can reduce that by 80%. That's another $631 in savings. We now have $10,119 to go.

Next up: network taps. You don't need most of that redundant Cisco gear, cabling, etc. any more. Figure $800 per year (full cost) for that stuff per server. We now have $9,319 to go.

Data center space costs money. (Osaka real estate is not cheap.) A System z9 EC requires about 58 square feet of space, including clearance. Let's double that for good measure. Let's assume IBM does a really good job and gets 20 old servers into the space one of these new mainframes requires. So 200 "footprints" are reduced to 30. That saves 19,720 square feet. Figure a couple hundred bucks a square foot for raised floor data center space (conservative again), and you take another $1,000 off. $8,319 to go.

If these servers are running Oracle Database, the annual software subscription and support is about $8,000 per CPU. Let's assume uni-processor CPUs (very conservative assumption). That means reducing 4,000 CPUs down to about 900 CPUs (assuming 30 processors in each of the new mainframes). That results in a savings of $6,200, and we now have $2,119 to go. Moreover, these 3,100 Oracle licenses can now be redeployed onto other servers somewhere else at IBM. (It's 3,100 fewer Oracle licenses that IBM has to buy.) Also, undoubtedly there is other software, like monitoring software. Let's make the assumption IBM gets this free.

Now I get to the stuff which is harder to estimate without more information but is no less real. Disaster recovery is comparatively cheaper on mainframes. (Mainframes have something called CBU -- Capacity Backup.) And it works better, too. Service Level Agreement penalties to IBM's customers are likely reduced with the new machines. IBM gets some money back by recycling the old machines (residual value). The technology refreshes (hardware replacements) are much cheaper on mainframes, both in terms of physical equipment and installation costs. (In a 5 year business case you'd have to include hardware refreshes.) They're probably also doing some disk and tape consolidation, and they're going to have SAN savings (fewer storage connections and switches). Disk space might go down at least slightly. There might be hardware maintenance expense (warranties), but perhaps IBM gets that for "free" on all its servers. All of these other factors could easily get past the remaining $2,119 figure.

Anyway, this is a very back-of-the-envelope analysis, and there are some transition costs to include. (Although there are also transition costs when simply replacing hardware during a normal refresh.) But yes, it appears IBM can easily make the case for the financial savings here. It's a pretty easy case, in fact.

What's really interesting is that IBM competes against other companies in the data center hosting business. If they can trim $250M off their data center expenses, some of that will go into reducing their bid price. And that means they'll be much more formidable as a competitor in this market and make more profits. This is a secondary effect, but it could be very big. Another secondary effect is that data center operators (outsourcers, private companies, and governments) will see these cost savings and want them also, buying more IBM mainframes. IBM wins again. Really smart move on IBM's part doing this 4000 to 30 consolidation, I'd say.

Re:A Peek Inside the Economics

[jgoemat]

First of all, that figure is $12,500 per year. Which isn't a lot of money.

I have to agree that isn't a lot of money for managing a server for a year taking into account all of your considerations. The problem is that you are wrong, the number is actually $62,500. I used calc.exe originally, but Google confirms it. So unless all of your numbers are off by a factor of at least 5, it's still a large amount of money :)

Re:A Peek Inside the Economics

[jgoemat]

So unless all of your numbers are off by a factor of at least 5

Which I guess they would be if you amortized over five years like the article... Doh!

DB2 for z/OS: Yeah, Right!

[BBCWatcher]

DB2 for z/OS pulls off this feat, when you configure it in a DB2 data sharing group (i.e. in a Parallel Sysplex). Yes, you can perform complete N+1 upgrades of the database server code, one member of the group at a time, while maintaining continuous business service.

Yes, it's an amazing piece of engineering. And it works.

Re:System z Mainframe Specs

[o2sd]

Sorry for the late reply.

That was a really informative post!

Each CPU is actually a "book" of CPUs. Several run at once on the same data. If any disagree, the instruction is rerun on a different CPU. Entire backup books (in addition to the 54) kick-in if a problem is detected.

That is VERY cool. I guess that forms part of the five nines uptime.

The hardware specs are impressive, but the biggest deal about these boxes is that they don't go down. Most people I talk to question the idea of consolidating servers into one box because of "single point of failure" concerns. This is where the mainframe shines. These things have MTBF of decades, and will just churn away forever.


I want one for my team (dre-e-e-eam, dream dream dream, dre-e-e-eam, dream dream dream)