Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

IBM's x86 Server Business Back On the Market

Soulskill posted about 9 months ago | from the just-throw-it-on-ebay dept.

IBM 71

itwbennett writes "It was widely reported last year (including on Slashdot) that IBM attempted to sell off its x86 server business to Lenovo, which seemed logical as Lenovo had bought out the IBM's PC business a decade ago. However, the two firms could not come to financial terms and the deal was never struck. Well, the rumors have started up again, only this time Lenovo has some competition, as Dell and Fujitsu are now being thrown into the mix as possible suitors."

cancel ×

71 comments

Sorry! There are no comments related to the filter you selected.

Does anyone check these summaries anymore? (5, Informative)

damn_registrars (1103043) | about 9 months ago | (#46032315)

the rumors have started up again, only this time Lenovo has come competition, as Dell and Fujitsu are now being throw into the mix as possible suitors.

Come one, that's just sloppy writing there. We can do better than "Lenovo has come competition" and "being thow into the mix".

Re:Does anyone check these summaries anymore? (1)

TrollstonButterbeans (2914995) | about 9 months ago | (#46032339)

Seconded.

Re:Does anyone check these summaries anymore? (0)

Anonymous Coward | about 9 months ago | (#46032369)

Come one ... being thow into the mix

Come on, that's just slopping writing there.

Re:Does anyone check these summaries anymore? (0)

Anonymous Coward | about 9 months ago | (#46034721)

Muphry strikes!

Re:Does anyone check these summaries anymore? (2)

Hangtime (19526) | about 9 months ago | (#46032415)

To be fair to the editors, this is exactly how it reads in the source. The use of quotes is warranted and allows the editors to escape my wrath....this time..

Re:Does anyone check these summaries anymore? (2)

bill_mcgonigle (4333) | about 9 months ago | (#46032583)

The use of quotes is warranted and allows the editors to escape my wrath....this time..

Wrath on, because an editor is supposed to mark such atrocities with [sic].

That being said, the summary is now corrected to proper English, so I'm terribly confused.

Re:Does anyone check these summaries anymore? (2)

wonkey_monkey (2592601) | about 9 months ago | (#46034199)

Wrath on, because an editor is supposed to mark such atrocities with [sic].

The submitter is supposed to write their own submission and not just copy-and-paste a paragraph or two from the article.

Re:Does anyone check these summaries anymore? (0)

Anonymous Coward | about 9 months ago | (#46032937)

Normal editing practice is to write (sic) next to obvious spelling and grammatical errors you intentionally keep from the original source to avoid being flamed by your readers.

Re:Does anyone check these summaries anymore? (0)

Anonymous Coward | about 9 months ago | (#46033081)

Actually you should use brackets and they should not be italicized. For example:

"Well, the rumors [sic] have started up again, only this time Lenovo has come [sic] competition, as Dell and Fujitsu are now being throw [sic] into the mix as possible suitors."

I do consider the American English spelling of the word "rumours" to be incorrect.

Re:Does anyone check these summaries anymore? (1)

x_t0ken_407 (2716535) | about 9 months ago | (#46038131)

...you should use brackets and they should not be italicized...

"...the rumors [sic] have started up again...Lenovo has come [sic] competition...now being throw [sic] into the mix..."

Huh?

Re:Does anyone check these summaries anymore? (2)

wonkey_monkey (2592601) | about 9 months ago | (#46034193)

To be fair to the editors

They're supposed to fix these things - that's why they call themselves editors.

this is exactly how it reads in the source

And submitters are supposed to write their own summaries:

Please try to use your own words; if you're quoting another source, make that clear.

Re:Does anyone check these summaries anymore? (1)

K. S. Kyosuke (729550) | about 9 months ago | (#46032423)

Come one, that's just sloppy writing there. We can do better than "Lenovo has come competition" and "being thow into the mix".

See? That's what happens when you don't buy IBM Chipkill(TM) memories for extra error correction!

Re:Does anyone check these summaries anymore? (0)

Anonymous Coward | about 9 months ago | (#46032651)

the rurnors have started up again, only this tirne Lenovo has corne cornpetition, as Dell and Fujitsu are now being throw into the rnix as possible suitors.

Corne one, that's just sloppy writing there. We can do better than "Lenovo has corne cornpetition" and "being thow into the rnix".

You both rnanaged to get it wong. Or is that wron?

Lenovo has "come competition"? (-1)

Anonymous Coward | about 9 months ago | (#46032321)

Is this like Peter North vs. Ron Jeremy?

Re:Lenovo has "come competition"? (-1)

Anonymous Coward | about 9 months ago | (#46032387)

Peter North for sure. That guy knows how to unload!

Don't even proofread the submission, guys... (5, Insightful)

bazmonkey (555276) | about 9 months ago | (#46032325)

Editors: not being able to proofread a few sentences is telling people "I want my job taken over by a computer program".

Re:Don't even proofread the submission, guys... (4, Insightful)

TrollstonButterbeans (2914995) | about 9 months ago | (#46032367)

>Editors: not being able to proofread a few sentences is telling people "I want my job taken over by a computer program".

No, this is quite intentional. Now there will be quite a few comments to the bad grammars -- which everyone can participate in [i.e. you don't need to know physics or electrical engineering] which generates activity, participation and ad impressions on the cheap.

Oldest trick in journalism to foster activity, sadly.

Re:Don't even proofread the submission, guys... (1)

Anonymous Coward | about 9 months ago | (#46033179)

The oldest trick in journalism is a strategy that came into existence within the past several years?

I guess The Internets have not only destroyed old journalism's business models, but eradicated any memory of the 200+ years of its existence.

Re:Don't even proofread the submission, guys... (0)

Anonymous Coward | about 9 months ago | (#46032377)

"Boycot" in the headline and "Sand Francisco" in the summary yesterday was embarrassing, but stuff like the ones SS just corrected... who gives a shit. I make mistakes like that in my own posts and email messages all the time.

Re:Don't even proofread the submission, guys... (0)

Anonymous Coward | about 9 months ago | (#46033297)

"Sand Francisco" sounds like a Dirty Sanchez on the beach.

Re:Don't even proofread the submission, guys... (1)

Neo-Rio-101 (700494) | about 9 months ago | (#46032535)

As long as the computer gives me the paycheck, I'm more than happy for my job to be taken over by a computer program.

Re:Don't even proofread the submission, guys... (1)

couchslug (175151) | about 9 months ago | (#46032787)

They don't care because Dicedot makes money, and it's not as if standards matter any more since it ceased being a tech site.

IBM strategy (0)

Anonymous Coward | about 9 months ago | (#46032345)

They need to put Ginny Rometty, Steve Mills, and the rest of those Wall Street-facing PowerPoint-tweaking MBA losers out on the market.

Re:IBM strategy (5, Insightful)

stevesliva (648202) | about 9 months ago | (#46032409)

I'm kind of curious what the PowerPoint jockeys think about this sale including *all* x86 servers. Blades?? "Enterprise" servers? They will draw a line somewhere that keeps some x86 server architectures in the IBM stable, I'd guess. Just like you can buy a glorified desktop from IBM as a "tower server" rather than having to go to Lenovo, a line will be drawn somewhere, and the "enterprise" servers will likely stick around IBM. At least until the margins suck on those, too.

Re:IBM strategy (0)

Anonymous Coward | about 9 months ago | (#46032693)

They're idiots. "IBM doesn't do low margin" sounds good on the Street, but customers hear that as "IBM only sells premium price stuff." Guess what - lots of customers face pressure of their own to cut costs. When you've got IBM equipment all over the data center there's lots of cross-selling opportunties and nice branding where it matters most. Guess they'll let Lenovo, Dell, or HP walk away with that, while IBM sponsors the Masters golf tournament on TV.

Re:IBM strategy (2)

antifoidulus (807088) | about 9 months ago | (#46033229)

There were(key word being were) some other companies that "didn't do low margin" such as SGI and Sun. I wonder where they are now...

Re:IBM strategy (1)

funwithBSD (245349) | about 9 months ago | (#46033271)

Well, we really don't make much on the hardware.

Consider the lowly x3550 m4, a box we sell in large quantities. It costs around 3K basic and $22K nicely decked out to handle quite a few virtual images with 384GB of memory (you can double that).

Figure a nice 40% margin, less if you are buying more of them. Say 20%. Buy a service contract we might even take a significant loss.

Now, if you have us support that device in a datacenter, I can tell you that we will make many more times that in basic sysadmin services. Of course, it includes everything, fixes, backups, growth (replaced in 3 years) 24hr monitoring, rackspace and networking.

But still, we don't make money off the server itself, so why make them? The only hardware that makes money is PureFlex (x86 + RS6000 racks) RS600, AS400 and mainframe. Even then, they tend to give away the hardware as part of the service.

Re:IBM strategy (1)

Anonymous Coward | about 9 months ago | (#46033987)

But still, we don't make money off the server itself, so why make them?

Actually, we do make money on other systems. I've been told iDataplex actually came in at around 20% margin. From what I hear, x86 Flex came in at around 20% margin too, despite aspirations of 50-60% margin. The problem with IBM is that if you try to *plan* for 20% margin (which some companies would not at all mind), then you are laughed out of the room. There is no way that all the money that went into Flex would have been spent if they had known at the time how much volume and how much margin they *really* were going to get out of it. This is one reason why I think another company might do better with the assets; a willingness to realistically aim and be happy.

Even then, they tend to give away the hardware as part of the service.

And herein lies a big chunk of the problem for IBM's hardware group and the rest of IBM after that group goes away. When there is a bid involving software and/or services together with servers, the servers get cut to the bone or sold at a loss to advance the welfare of software and services. This is due to a partially self-fulfilling prophecy that being in the hardware business is low margin and therefore should be screwed over to make room for the more successful parts of IBM. Part of IBM's hardware failings is because they make it fail because they think it is failing. Keeping in mind IBM continually wants to do this and executives force STG to take the hit for the sake of SWG and GTS, what happens when STG products are no longer under the same executive umbrella? Those suicidal discounts won't happen anymore because the external vendor doesn't answer to some executives that want to see IBM succeed at all costs.

Re:IBM strategy (1)

stevesliva (648202) | about 9 months ago | (#46047523)

And herein lies a big chunk of the problem for IBM's hardware group and the rest of IBM after that group goes away. When there is a bid involving software and/or services together with servers, the servers get cut to the bone or sold at a loss to advance the welfare of software and services. This is due to a partially self-fulfilling prophecy that being in the hardware business is low margin and therefore should be screwed over to make room for the more successful parts of IBM. Part of IBM's hardware failings is because they make it fail because they think it is failing. Keeping in mind IBM continually wants to do this and executives force STG to take the hit for the sake of SWG and GTS, what happens when STG products are no longer under the same executive umbrella? Those suicidal discounts won't happen anymore because the external vendor doesn't answer to some executives that want to see IBM succeed at all costs.

YES.

I would guess their logic is that customers in the segments they divest don't actually care about the hardware, and if they do really want Thinkpads or IBM x86 hardware, they can pay a little more to get it from Lenovo. Hard to believe with Thinkpads being gone, but the strategy seems to be to keep the hardware that customers demand to the extent they're willing to pay.

Deal to lenovo is done:

Research Triangle Park, North Carolina, and Armonk, New York - 23 Jan 2014: Lenovo (HKSE: 992) (ADR: LNVGY) and IBM (NYSE: IBM) have entered into a definitive agreement in which Lenovo plans to acquire IBM’s x86 server business. This includes System x, BladeCenter and Flex System blade servers and switches, x86-based Flex integrated systems, NeXtScale and iDataPlex servers and associated software, blade networking and maintenance operations. The purchase price is approximately US$2.3 billion, approximately two billion of which will be paid in cash and the balance in Lenovo stock. IBM will retain its System z mainframes, Power Systems, Storage Systems, Power-based Flex servers, and PureApplication and PureData appliances.

$2.3billion isn't much. Sheesh.

Re:IBM strategy (1)

bored (40072) | about 9 months ago | (#46035717)

But still, we don't make money off the server itself, so why make them?

If this is what passes for IBM logic these days, then the sooner they die the better.

You sell/manufacture them, so that you can bundle higher margin products and provide a one stop shopping/support experience. How many Dells are running Tivoli software, or better yet how many places are running Dells and hiring IBM to manage them? I'm betting those numbers are far lower than the places with IBM hardware.

So cry me a river, Lenovo is doing just fine with the "low margin" PC's business they purchased from IBM. In fact they are kicking IBM's ass, their revenue numbers have been increasing ~13% per year for the last few years while IBM's revenue numbers have been flat for a decade.

If IBM management had any creativity, they would figure out how to in source some of that manufacturing and streamline it rather than giving foxconn/whoever a 5% cut then crying their isn't any margin left.

Re:IBM strategy (1)

funwithBSD (245349) | about 9 months ago | (#46036957)

I'm betting those numbers are far lower than the places with IBM hardware.

You would lose that bet.

We manage anyone's hardware in our datacenter. Totally agnostic. You want dell? Sure. HP? Sure. SUN/Oracle? No worries.

It does not matter.

Now, we give them a good discount if they want our hardware, but they are welcome to spec any hardware they want.

Re:IBM strategy (1)

maztuhblastah (745586) | about 9 months ago | (#46033111)

Just like you can buy a glorified desktop from IBM as a "tower server" rather than having to go to Lenovo

"Having to go to Lenovo"?

What exactly do you think those tower System x boxes are, anyways?

Re:IBM strategy (1)

Anonymous Coward | about 9 months ago | (#46033199)

Are they Lenovo? I knew IBM outsourced their x86 servers, but I wasn't sure if it actually was Lenovo.

If it is Lenovo then Fujitsu et al would be stupid to buy it. If Lenovo doesn't think the asking price would be profitable for them, even though they'd have the lowest costs in transferring the business, then IBM must be demanding way too much.

Re:IBM strategy (2)

Junta (36770) | about 9 months ago | (#46033867)

I knew IBM outsourced their x86 servers

Actually, currently, they don't outsource their servers (in the sense of 'rebadging') for the most part. They are usually manufactured by the likes of foxconn, and sometimes much of the fine grained design work is done outside, but at the very least IBM does high level design (if not down to the nitty gritty). They even write their own UEFI implementation where pretty much everyone else goes to AMI. This doesn't necessarily mean good or bad things about the systems, it just happens to be that way.

Exceptions include things like the x3450, the e32*, and the x3755 m3. There may be some others, but at least their most common servers are actually IBM designed. Tower servers might have more outsourcing than the rackmounts...

Re:IBM strategy (1)

Anonymous Coward | about 9 months ago | (#46035045)

IBM's UEFI implementation is a big reason I switched to Dell.

Re:IBM strategy (1)

bored (40072) | about 9 months ago | (#46035893)

No, crap.. Some of these UEFI servers are amazing...

Amazing that they can take 5+ minutes to get to the grub menu.

Of course the RS6k/pSeries machines I've used are worse. Stack a few IO drawers/FC cards in a machine and be prepared to wait a 30 minutes to boot to AIX.

I didn't realize dell was better. I'm going to have to give them another look. We have HP, IBM, and supermicro.

The supermicros we have boot in 1/10th the time with similar system configurations.

Re:IBM strategy (1)

Anonymous Coward | about 9 months ago | (#46037597)

Of course the RS6k/pSeries machines I've used are worse.

The benefit of pSeries (or in my case System z) that hardware reboots are very rare indeed. We POR our System z EC12 max twice per year for z/VM upgrades and the like and our pSeries guy's last reboot of any significance was for some fibre channel firmware or hardware replacement. I think, on the whole, it's a moot point and we could all do with less Lennart Pottering-esque conflating of server vs. desktop priorities.

Re:IBM strategy (1)

bored (40072) | about 9 months ago | (#46039319)

I think, on the whole, it's a moot point and we could all do with less Lennart Pottering-esque conflating of server vs. desktop priorities.

I think boot times matter on server gear too, and for a while so did IBM, back when they were pushing availability numbers because a single hour long reboot, once a year can put the system availability in the three nines category. Sometimes the hardware needs actual reboots be that for firmware or hardware upgrades... None of my machines have physical hotplug memory for example... And none of my PCIe x86's have hotplug PCIe, even though a fair number of the PCI-X machines did.

When my desktop takes 5 minutes to boot, one person is inconvenienced. When my database server takes 30 minutes to boot, potentially thousands are inconvenienced.

BTW: I have a zSeries too, and we don't have good power conditioning on it, and it seems to take a power hit once or twice a year. That thing is a PITA because the "DASD" takes the better part of two hours to come back up after hard power loss.

Re:IBM strategy (1)

Anonymous Coward | about 9 months ago | (#46037883)

Dell and I had a falling out around the early 2000s, and I switched to IBM. The upgrade in support was no contest, and the servers were solid. I had a not-so-good run with IBM Opteron servers, but otherwise I had no complaints. Later on I had increasing issues with the company across the board, culminating in a server which took 8 months to be delivered (that's no typo). And for the wait I got one of the IBM EUFI wonder-systems. (I was also weary of how it interacted with Redhat and didn't want to see how they did with Windows).

So I'm back with Dell. I haven't done comparisons, but I'm satisfied with their products and plan on sticking with them for the immediate future. Boot times are reasonable even with PERC controllers and such.

Re:IBM strategy (2)

Junta (36770) | about 9 months ago | (#46033889)

They will draw a line somewhere that keeps some x86 server architectures in the IBM stable

I'd guess the line could be somewhere around 'we can slap the IBM badge on the server on the way out the door' at best. Or else they'll have their brand completely removed from the boxes presuming that opens things up for IBM to be viewed as a logical provider for services and software on top of HP, Dell, or whatever server in x86 land.

Keeping some x86 business back from a divestiture could be pretty catastrophic for the bit that stays. When IBM sold PC, procurement got hit *hard* due to reduced volumes and had difficulty containing costs. Keeping, say, just their very expensive boxes would make that look like a walk in the park.

FRIST PSOT! (-1)

Anonymous Coward | about 9 months ago | (#46032431)

Well, almost!

What does IBM do anymore then? (0)

Psychofreak (17440) | about 9 months ago | (#46032537)

What does IBM do anymore then?

Intragalactic Bowel Movements?

http://www.schlockmercenary.co... [schlockmercenary.com]

Re:What does IBM do anymore then? (3, Informative)

JCHerbsleb (2881347) | about 9 months ago | (#46032741)

Technically, they still have mainframes, System i, System P, etc. along with the various software platforms such as DB2, RACF, and the various BMC products. I think they are attempting to transform themselves into a "service" organization (similar to what HP is attempting) and divest the hardware / software while focusing on the consulting and outsourced support services.

How is it that Lenova is able to make money (3, Funny)

Anonymous Coward | about 9 months ago | (#46032897)

off of a PC business that IBM was not able to. Is it because USian workers are all lazy? You know what we really need to do is to give China the cream of the USAian elite managers. I think we should export all the Harvard school of management graduates along with all our intellectual property lawyers to China. Think of all the exponential improvements the Chinese economy would undergo. Sure the Chinese are great at making things, but they have no management skills. We need to help these poor Chinese out. The USA may suck at most things but we have the best management in the world (just ask them). How much better would China be if it had Carley Fiorina in charge of something. It would be a real boom to the elite managers also. No longer would they have to deal with all the fat lazy USian workers. Unfortunately the workers will be directionless, and would not know how and where to go for the next meeting. So be it. That is the price of globalization.

Re:What does IBM do anymore then? (2, Informative)

Anonymous Coward | about 9 months ago | (#46033237)

IBM switched to services over a decade ago. They were the first and largest computer hardware vendor to do it, and other companies have been fruitlessly trying to emulate them. IBM's strategy is, AFAIU, actually studied in business school (anyone with a recent MBA care to chime in?), because they were one of the largest companies ever to successfully make such a fundamental switch to their business model, and their execution was near perfect.

Yes, IBM still sells mainframes and other systems, but the actual hardware is a small part of their revenue.

What IBM does is 1) invest in new research and technology, 2) productize, 3) make huge margins for a few years, then when they see commoditization coming around the corner 4) divest themselves of it and switch to integration and consulting.

Re:What does IBM do anymore then? (0)

Anonymous Coward | about 9 months ago | (#46034447)

When I was in college (which is not that long ago), my lecturer presented IBM as what he called `the service paradox': it would seem that services would generate the most revenue for IBM, but it is far from the truth. He showed us the company results of that year: over 60% of all spending was on services, less than 40% was revenue generated by services.
Surprisingly enough, their mainframe division was doing better than any other division of IBM at that time.

Re:What does IBM do anymore then? (0)

Anonymous Coward | about 9 months ago | (#46036617)

Much of IBM's services business is directly or indirectly related to their mainframe business (often outsourcing the mainframe users), so there's quite a bit of spin involved there. They still largely are a computer company.

Re:What does IBM do anymore then? (1)

unixisc (2429386) | about 9 months ago | (#46038265)

Isn't that more what Dell is doing? HP still has their usual servers, as well as their Integrity servers. But Dell is perfectly poised to transform into a pure services organization

Re:What does IBM do anymore then? (1)

TheloniousToady (3343045) | about 9 months ago | (#46032747)

It wasn't widely reported at the time, but Apple acquired their Itty Bitty Machines division several years ago. So, they don't even make phones and MP3 players anymore.

Re:What does IBM do anymore then? (1)

msobkow (48369) | about 9 months ago | (#46034433)

Software development and management services are their big bread and butter nowadays, not hardware.

Re:What does IBM do anymore then? (1)

PolygamousRanchKid (1290638) | about 9 months ago | (#46034593)

Lots of software: http://www-03.ibm.com/software... [ibm.com] . Click on W-Z to see all their WebSphere brand products:

WebSphere Application Server
WebSphere Business Compass
WebSphere Cast Iron Cloud integration
WebSphere Bigboote ("It's big-boo-TAY! TAY!")
WebSphere Small Berries
WebSphere Lord Whorfin
WebSphere Many Jars
WebSphere Littlejohn
WebSphere O'Connor
WebSphere Parker
WebSphere . . . .

LOL (0, Troll)

WindBourne (631190) | about 9 months ago | (#46032613)

IBM could easily keep this and be innovative. Sadly, it is ran by a bunch of MBA's who are only interested in running up stock prices.

Re:LOL (2)

rubycodez (864176) | about 9 months ago | (#46032727)

IBM does not make innovation, they make money. commodity x86 server market is not the place to make money

Re:LOL (0)

Anonymous Coward | about 9 months ago | (#46033181)

I can hardly think of another server maker that's as innovative as IBM. Watson, their mainframe series (System Z), gameframe and their recently announced Power 8 series spring to mind.

On the other hand I cannot think of any duller and less innovative business than sticking Intel manufactured CPUs into a custom enclosure and call it an X86 server

Re:LOL (0)

funwithBSD (245349) | about 9 months ago | (#46033289)

Not to mention PureFlex, which is the densest x86 (and RS600) "datacenter in cabinet" blades in the business.

Moonshot might eventually beat it when they get the Atom based cards on the market.

Re:LOL (1)

serviscope_minor (664417) | about 9 months ago | (#46033777)

Not to mention PureFlex, which is the densest x86 (and RS600) "datacenter in cabinet" blades in the business.

How sense is that?

Compared to the old single socket CPU in a 2U box that server vendors love to compare too, sure. Frequently the super-dense "solutions" hawked stack up badly compared to commodity 1U quad socked boxes, even given an extra U for a switch.

By stack up badly, I mean in terms of FLOPS/U and RAM per U.

I've worked with those commodity boxes and have been impressed. Basically, substantially over 50% of the box is taken up entirely with processor/heatsink combo and DIMMS. There's a chunk taken out for the 1kW PSU. The back end of the box is the motherboard. There's also a couple of rows of fans taking up a bit more room. Regardless of the (hot) CPUs, .5TB (they can fit 1 in now) of RAM takes quite a bit of cooling.

The empty room consists of less than 1/2 a U over the back part of the motherboard which is where there's space for a PCIe card (or two?), and there is space at the front for ful; sized hard disks. Some of them have room for a second PSU.

Basically when blades became popular, rack PCs were mostly empty space. these days good commodity ones aren't. If you are really careful, integrated the high end networking onto the motherboard dropped the redundant PSU and had no hard disks, you could perhaps squeeze an extra 15% if you're lucky.

So, how good are these devices? Do they really stack up against a cheap stack of 1U supericros in terms of density?

Re:LOL (1)

LWATCDR (28044) | about 9 months ago | (#46034459)

Love the idea of a cheap 1U quad socket commodity server. Got to love Enterprise level budgets.

Re:LOL (1)

serviscope_minor (664417) | about 9 months ago | (#46036153)

Love the idea of a cheap 1U quad socket commodity server. Got to love Enterprise level budgets.

They really are cheap. If you instead buy cheap-ass desktops in cheap cases with crap PSUs, you can get double the bang for the buck, at the cost of taking up masses of space and having no ILM or other remote management, no ECC ram (this is doable in cheap desktops with AMD) and no large system image.

The base price for such a system is GBP 1200 for the chassis and motherboard. Fully maxed out with .5T of RAM and 4 top of the line 16 core processors, it will cost about 8k. That's a heckuva lot of compute, and vastly less than the exotic blade systems.

8k is an expensive computer but by no means enterprise level budgets.

I've never worked for a big enterprise, but I have had 2 of those beasties on a job before.

Re:LOL (1)

LWATCDR (28044) | about 9 months ago | (#46041991)

I was not saying that they where not a good value. Just the idea of that much compute power being cheap is amusing. 8k for a server is still in enterprise pricing to me but well worth the money

Re:LOL (1)

jp10558 (748604) | about 9 months ago | (#46045399)

Maybe I'm stupid and haven't worked out how to make everything run in cloud like box can fail without taking anything down and you just swap the entire box, but IBM has some real value adds in the x86 space. (I have to guess other vendors do as well, but I use primarily System X).

The ILM (IPMI/ IMM whatever it's called) is really really useful if you don't want to have to be in the server room (or haul it out of a rack and then haul it back in) to do OS installs or BIOS configuration or low level troubleshooting.

Light Path diagnostics over the IMM is really useful for finding a part that's bad so you can get the new part shipped without masses of on site troubleshooting. When you go to replace the part, it makes it dead simple for a tech to see by the LED which part to replace. The "indicate" LED blink from the IMM makes it quick to find the correct server.

Enterprise reliability is noticeably better than whitebox tower chassis(at least the one's I've tried), and you're assuming space isn't an issue, which this entire thread is about.

Now, the Flex system x222 (the high density mentioned) is pretty cool (I have one I'm evaluating now). It is two 2 processor Xeon servers, each with up to 384GB RAM and multiple 10Gbit ethernet ports in a 1/2 width 1U space.

Re:LOL (0)

Anonymous Coward | about 9 months ago | (#46034769)

So density wise, x220 Flex servers are more dense than quad socket supermicros, though x220s use lower performing CPU/memory subsystem.

It compares favorably socket-density wise with x240 Flex servers, but not memory wise. For performance the Flex x240 will frequently fare better (QPI topology of E5 processors in quad socket configuration really sucks for performance). The supermicro does not provide as memory dimm slots as the processors could support.

The x440 is the same density as x240, but the performance is going to be more similar since it too uses the same QPI topology.

Also keep in mind that the ongoing costs of a bunch of 1U servers will be higher (suck down more power) and carry certain risks (cable management is not just for looks, a misstake or slip can knock out production systems). Whether the ongoing costs offset the upfront cost is a matter of debate and varies application to application. For RJ45 connections, the cable management is generally considered no big deal, but if you have SFP/QSFP in the picture, it is a non trivial thing to ignore. Because of these various debates, IBM also does things like NextScale, which offers the performance characteristics of the dual socket with the density of that quad socket and power consumption in line with a blade solution, but not with the cost of a blade solution. The cabling will be twice the challenge of a rack of 1Us, so there is a drawback as there is in all things.

Re:LOL (1)

Junta (36770) | about 9 months ago | (#46033925)

Not to mention PureFlex, which is the densest x86

I don't see how anyone could claim that. Flex is a 10U enclosure that holds, at best, 28 dual socket servers. The Dell M1000e can hold 32. IBM's own nextscale gets 12 in a 6U leading to theoretically 84 in a rack (which matches the low-speed bladecenter).

Of course, *highest* density is an overrated metric, since the densest solutions are rarely deployed since the power and cooling footprint frequently exceeds a datacenter's designed capacity. Atom based servers may mitigate that issue, but it is unclear whether their compute throughput will make it make sense compared to a well designed solution constructed out of higher power consuming hardware.

In any event, IBM's x86 business (apart from their high end x86 servers) is more about extremely thorough and detail oriented engineering and quality service and support rather than revolutionary innovation. They perhaps do some of the most in depth failure analysis in the industry and design their systems accordingly. They saw warranty costs due to corrosion of metal contacts in a dimm slot and they designed a coating to mitigate that, for example.

Re:LOL (1)

jp10558 (748604) | about 9 months ago | (#46045461)

I feel like IBM is doing something other vendors really aren't, and don't know if anyone but Lenovo would continue to do (I am going by Lenovo's continued excellence in the workstation and thinkpad hardware).

And that is make products that work well for mid-sized businesses. Of course, that's not a market IBM really wants, hence the attempted sale in my opinion. I find their price premium buys you some great engineering as you said, as well as good, no-nonsense support on hardware issues (ESC+ is a revelation over Dell or HP because it's built into the basic warranty - which itself is weird vs how IBM usually does business).

The slight premium price doesn't make sense if you're a huge Enterprise because you'll probably be doing internal + external cloud, so I think RAID for whole computers - so you want super cheap, swap it out if it breaks. It also doesn't make sense if your scale is 5 servers - you don't have the budget, and can just walk up to them to do all the troubleshooting. But it saves my bacon in the middle scale.

Re:LOL (1)

evilviper (135110) | about 9 months ago | (#46035843)

commodity x86 server market is not the place to make money

The x86 server market doesn't have to be all "commodity". Look at a company like HP, who's NonStop and Tandem systems look likely to switch over to x86-64 processors in the near future.

As long as you can maintain some simple value-add on your servers, the CPU architecture really doesn't matter. Apple figured that out on their desktops/workstations. Cray has that figured out in their supercomputers. And IBM... Wants as much vendor lock-in as they can get, to keep people from fleeing from their support contracts, that amount to permanent hardware and software rental agreements.

Re:LOL (1)

LWATCDR (28044) | about 9 months ago | (#46034441)

Not a whole lot of room for innovation in the sever space.
Case design? Not really everyone uses racks.
CPU? Not in the X86 space.
Software? Linux, Windows, maybe BSD or Solaris.
Servers == following standards.
Once you get to the big iron running Power you have room for innovation.

Maybe reduced 'innovation'. (0)

Anonymous Coward | about 9 months ago | (#46035133)

There is much truth to what you say, but there is still engineering.

One, you oversimplify the engineering of the mechanicals as 'case design'. There is a good amount of thermal engineering toward the end of optimal power consumption. Multi-server enclosures segmenting cooling zones the right way and ramping the fans just right.

Component selection plays a critical piece too. IBM has kicked out voltage regulator vendors and similar things.

The other piece is about manageablity. The 'innovation' (if any) is mostly at the software layer, but it does require a very thorough and disciplined instrumentation of the components in hardware and firmware.

Re:Maybe reduced 'innovation'. (1)

LWATCDR (28044) | about 9 months ago | (#46042089)

That is more evolution the innovation IMHO.
Component selection is just good QA.
There is nothing wrong with good solid engineering and frankly that is what you want and need in the x86 space.

Of course when you get to blades, HPC with custom interconnects, and CPU GPU compute systems you are in a different category and one I believe offers a lot of room for innovation.

Is that configuration is same or upgraded one? (0)

Anonymous Coward | about 9 months ago | (#46033207)

Is that configuration is same or upgraded one?

http://www.vistasadindia.com

I have a hard time seeing Dell... (0)

Anonymous Coward | about 9 months ago | (#46034609)

First, obviously there is the fact that they just spent a boatload of money buying themselves, it seems a stretch they would be able to extend themselves to acquire such a large thing. Moreover, what they stand to gain seems insufficient to go out on a limb. Dell already is well positioned in the x86 server market, and an acquisition would probably drive much of the customer base to HP (a good chunk of their business is loyal to *IBM*, if forced to choose between Dell and HP I think they'd choose HP). If they are looking to be the provider of servers to IBM's solutions, then that would be counter to Dell's ambitions to grow their own. Of course, the scope of the sale is unknown so it might actually include some of the groups that integrate solutions (notably, much of the Top500). Of course, one thing it would do is see IBM's server presence go away without some other company getting the benefit of a jump into the market. One thing is for sure though, IBM and maybe Dell employees best hope if a sale happens, it is *not* to Dell. The overlap between Dell today and what they would acquire means there would be a bloodbath of eliminating redundancy. Probably the best hope is Fujitsu. The least overlap and a chance at retaining presence at US government sites (pretty big chunk) that would not purchase Lenovo. Fujitsu is probably in a stronger position than Dell in terms of willingness of southeast asia to buy.

Fujitsu and IBM enterprise a good fit (1)

shuz (706678) | about 9 months ago | (#46034899)

Fujitsu is known for making some solid never fail tank style servers. I admin a few of these myself and didn't even realize the hardware vendor for many years until a cluster failover card failed and needed to be replaced. In this case it was a Fujitsu Sun system. I can only assume Fujitsu IBM systems would carry on the overbuilt stability minded servers you come to expect from an enterprise server like IBM.

I can't say the same for the other two contenders.

Check for New Comments
Slashdot Login

Need an Account?

Forgot your password?