Apple Kills the Xserve: What Does it Mean?

| Analysis

This morning, we all learned that Apple has cancelled the Xserve effective January 31, 2011. Apple will continue to support the product, but the first reaction by many will be justifiable, grave concern. However, understanding the details, Apple’s motivations, and customer tendencies is harder than just flying off the handle. Here’s my analysis.

In Ian Fleming’s novel, Goldfinger, Auric Goldfinger said, “Once is happenstance. Twice is coincidence. Three times, it’s enemy action.” It certainly seems like we’re seeing enemy action when it comes to Apple announcements regarding key technologies for the enterprise. And, in this context, for convenience, I’m using the term enterprise as a shorthand to include higher education, research, science, business, and government.

The first reactions to the cancellation of the Xserve, effective January 31, 2011 are likely to reflect serious concerns by those invested in the product and cause them to question Apple’s intentions in this space.

For context, one has to remember that Apple was seriously involved in cluster computing, supercomputing and scientific research in the 2000 through 2005 time period. Large supercomputers at Colsa (in Alabama) and Virgina Tech were built with Xserves. Many, many more smaller clusters were built as Apple sought to provide small, rack mounted solutions to the “enterprise” as I’ve abbreviated it above. Even a single Xserve in an equipment closet could provide serious capability to a small business. The combination of the Xserve and Mac OS X Server makes for an easy to use, powerful combination, manageable by a mere mortal, not a giant team.

And so one has to ask, why is Apple doing this, what’s the real impact, and what message, if any, is Apple sending?

Two Sides of the Issue

There are always two sides to an announcement like this. If there’s anything I learned at Apple, it’s that nothing is as it appears and that one has to both understand Apple’s deepest motivations as well as those of customers. Outside of that context, a lot of angst just isn’t very justified.

On the research side, there will be people who’ve invested in this hardware and Apple’s vision. Apple brought a lot of ease of use, design elegance and just plain common sense to Xserve and the former Xserve RAID. There are plenty of senior staff in any organization who subscribed to Apple’s way of doing things, have a Mac on their desks, and successfully deployed Xserves.

That said, we always worry about the message Apple is sending and that gets balanced against Apple’s business interests. Whether Apple likes it or not, this decision does send a message, but we have to have good information to interpret it correctly. Regrettably, shallow interpretations of the message often percolate through the community and is something that affects Apple, for a time, whether it likes it or not. The company has learned to live with that.

The reality is that serious researchers have long wondered about Apple’s commitment to certain markets and fitness as a long term, reliable business partner. Decisions like today’s create new cracks in one’s confidence because certain market sectors depend on long term stability and steadiness of purpose.

Drew McCormack who works with wrote me about his concern.

“Unfortunately, those of us who have worked with Apple products in the scientific and enterprise arenas have seen this coming for some time. The success of the iPod, and later iOS devices, has made the Mac a more and more marginal part of Apple’s product line.


“Apple has never been comfortable with high-end markets — it is fundamentally a consumer company — so I think they have just decided to consolidate their efforts on consumer markets, and leave the rest to the HPs and IBMs of this world. A sad day for those of us who loved the Xserve, but ultimately inevitable.”

When I was at Apple, I worked a little with the Virginia Tech supercomputer (built with Xserves). I got to know Jason Lockhart, one of the managers. He wrote me:


“It [Xserve] was a solid server platform that matured nicely. I will be sad to see it go. I knew Doug before he became product manager for the Xserve and Xserve RAID and was there when he released the product to the world. He was very proud of those products and worked very hard to refine their designs over their lifespan.”

I can attest that, at Apple, customers constantly confronted me with concerns like this, and they are legitimate. It’s a question of how Apple does business, and it’s different than how, say, IBM does business.

However, as you may have suspected, there’s more to the story.

Apple’s Position

One can’t analyze a decision like this without understanding Apple’s thinking. First of all, we have to understand that Apple was a very different kind of company when the Xserve was introduced. Apple was a smallish, one product company, the Macintosh, and it needed to branch out in every way it could in order to generate sales. That meant developing products like Xsan, the Xserve RAID, the Xserve and Mac OS X Server. These gave Apple credibility, traction and sales where it needed it in the enterprise.

Contrast that to today and look at the contribution from various Apple products to its revenue. The iPhone, iPad, iPod, and non-Xserve Macs make up more than 90 percent of Apple’s business. Meanwhile, Apple sells not so many Xserves. Michael Gartenberg, an expert industry analyst and partner with the Altimeter Group told me on the phone today:

“Sure, a small number of people are going to be very annoyed. But you have to remember, Apple isn’t getting rid of the software [Mac OS X server], they’re just changing the hardware. Right now, the Xserve isn’t selling in quantities that make sense, and, actually, the Mac mini server is doing much better. “I don’t think Apple is sending a message here. If the demand were there, Apple would keep selling them. Killing any product is hard, and someone, somewhere is going to be upset. But right now, the product just fit into Apple’s lineup.”

Mr. Gartenberg also pointed out that you can buy as many as you need right now and Apple will continue to support them, honor warranties and support contracts.

Tim Bajarin, principle Strategist with Creative Strategies, Inc, concurs.

“Remember, Apple isn’t killing the software side of this, Mac OS X Server. When I spoke with Apple, I was told that sales of the Xserve were decreasing over time. Meanwhile, more and more customers were ordering Mac minis (which are also stackable) and Mac Pros with Mac OS X Server. So this is really a case of killing one SKU that isn’t selling well, but I think Apple is committed to Mac OS X Server. I don’t think Apple is sending any kind of message”.

Finally, don’t forget: Apple sales people will be explaining the deeper facts to their customers.

Looking Forward

The other thing we have to consider is that Apple is an agressive technical company. It seeks markets that hunger for a new product, like the iPad, and uses its design, technical and manufacturing capabilities to build great products that fulfill that need. Apple continues to surprise us.

Keeping that in mind, we have to suspect that Apple will continue to develop newer products that meet our needs and sell like gangbusters. In order to do that, it can’t dwell on and unnecessarily commit to products whose sales are declining, despite the incorrect conclusions we might come to. The recent revelations related to Light Peak suggest to us that Apple always has something up its sleeve, moves ahead briskly, and won’t be hamstrung by products of the past. We should keep that in mind if Light Peak comes to fruition and Apple just plain bypasses USB 3, just as it fundamentally bypassed HDMI with DisplayPort.


My first reaction to this announcement was alarm. We’ve seen Steve Jobs announce the Mac App store which could affect how businesses can manage their Mac installations. Apple cancelled the Xserve RAID (February 2008), deprecated Java recently, and now has axed the Xserve. It would be easy to draw all the wrong conclusions and throw a fit, especially if one has to explain this to non-technical upper management.

On the other hand, remember that Apple has partnered with Unisys recently to support Apple products in the enterprise. Apple has shown no signs of disbanding the enterprise and federal sales units who contribute significant sales. Macs, but especially, iPhones and iPads are enormously popular in businesses that remain dominated by PCs and Windows.

Apple is a pragmatic company that cuts it losses, doesn’t cling to low volume technologies, and constantly, relentlessly moves into the future. If you need a 1U rack server in 2011, there are plenty of companies that will sell you one. Some of them are very good at what they do, like Hewlett-Packard, some may be out of business shortly, and some may be unexpected. Apple, however, will move smartly towards US$100B in annual sales in a year or two, doing what it does best.


Postscript: For further insights, see this letter from the father of the Xserve, Alex Grossman, now the CEO of ActiveStorage. Careful reading is in order.

Popular TMO Stories


Mike Weasner

I initially had the reaction of most people; there goes the Enterprise as a Mac OS X Server customer.  But then, I had two more reactions after reading John’s article.

1) There goes Mac OS X Server (some year soon)
2) Here comes iOS Server (some year soon)

Yes, today, the Enterprise can have OS X Server on a large, space-consuming, Mac Pro, with a lot of available disk space.  Or it can have OS X Server on a less powerful with less disk space but much smaller Mac mini.  Neither may be desirable to some IT organizations.

But what about iOS Server running on some future really small but really powerful platform? Could we see a data center handing all the world’s users but housed in a closet?

The future isn’t what it used to be.


Here’s prediction:
Apple will spin Logic and Final Cut off into their own company, maybe called something like ClariX. They will be maintained for a few years, before Apple reabsorbs them, and then kills them. This will be around the same time an animatronic Steve Jobs returns to save the company by announcing a new version of iOS based on a custom version of Lisp that has been developed using the (now open source) version of Windows Vista as its core. The more things change…


Mike,you’re NOT going to see a iOS server running on ARM processors. They are not powerful enough.

I suspect this is an indication there will NOT be a server version of Lion or anything else.

Apple keeps dropping it’s pro features until there’s no pros left to make content for their iOS devices.

John Martellaro

I added a Postscript to the original article. Check it out.

Bosco (Brad Hutchings)

My reaction if I was the guy inside some shop that pushed for these things is to figure out how to pay for Christmas without a bonus. If Dell did this, you could replace them when needed with HP servers running whichever Windows or Linux you liked. But since I went with Apple, I can’t just take Mac OS X Server and run it on another vendors’ boxes.

Let this be a lesson to the enterprise… Exit costs aren’t just when you decide you want to leave. They’re also when the vendor kicks you out.

Like Java, this is another move that’s about shutting doors and throwing away the key. Maintenance and carrying costs are zero. While they aren’t making gobs of money, they aren’t losing money keeping these around. They generate bad will among the IT guys who bought it which will be played back in spades when those same IT guys get to decide about iPhones, iPads, and Macs.

And of course, the commenterati will take a deep breath and hold back any judgement—because it would be harsh if they didn’t—and just let Steve Jobs keep getting away with this crap in the name of moving forward. I bet the next thing he cuts, he’ll announce on a Tuesday and spend the rest of the week basking in praise.

Mike Weasner

Mike,you?re NOT going to see a iOS server running on ARM processors. They are not powerful enough.

I never said what processor the iOS Server platform would use some day in the future.  Yes, it won’t be today’s ARM.  Let your imagination run wild.  Think Different. (is that phrase Trademarked?)


Steve Jobs is SUCH an idiot. Why is the board even letting him make decisions about which toilet paper to use. He probably picks sandpaper, and then as sandpaper manufacturers ramp up to support the new demand, he cancels just for the hell of it. He is so dumb. He is really really dumb. Run and tell that, homeboy.


Didn’t they just make OS X Server available as an option on Mac Pro machines? Am I missing something here?

If I understand this correctly, they are just killing the pizza boxes (because nobody bought them), but the server line continues to live in a different form. Much like killing the desk lamp iMac G4; or killing the clamshell (colourfull) iBook G3s… They were all succeeded by other products.

Right now, you can by a Mac OS X Server with a Mini or a Pro. I’m not quite sure what door has been locked and who threw the key here.

IBM did a much worse thing to the place where I work. They used to have their own LMS product (Workplace Collaborative Learning). They simply decided to kill it and NOT provide any other option. They are now recommending an independent product (Saba), but they offer no migration tools for IBM’s users stuck on their discontinued product, and the lone support engineer will cease support in a few months. Talk about locking doors and throwing keys!


As a photographer and a long time Apple observer and user, I decided to use Adobe Lightroom instead of Apple Aperture because I just didn’t trust Apple to retain its investment in the software, leaving legacy users out in the lurch with catalogues of photographs that would not be supported on new hardware.  I saw Apple abandon the traditional approach with iMovie, which left my wife trying to use an old version of iMovie for her projects, until that is no longer supported.

I cannot see why a company would abandon a profitable unit just because it is no longer their core focus.


Clearly it wasn’t selling all that much. Instead, they’re now selling traditional tower servers (as well as Minis).

John Martellaro

vasic: No you’re not missing anything…. grin

Bosco (Brad Hutchings)

I must really be dense. Because I read Grossman’s letter twice (once at private suggestion of John) and here’s the takeaway I got (from Grossman’s perspective):

* Wow, my investors are gonna be pissed. We made our DUV being an end-to-end Apple native solution on the server/storage side for Apple content people on the client side.

* This was a huge surprise. I wouldn’t have left Apple to start this if I didn’t think I could get at least 5 years out of it and figure out something for the company to do next.

* Apple’s workaround will require 3x the space and 4x the electricity for equivalent processing power.

* What if we stick 200 Mac Minis on shelves inside a full-sized garage freezer? How would that compare as a render farm to a rack full of Xserves?


Back to *my* perspective. I obviously don’t get it. Someone explain.


How profitable could it have been? I’m guessing, but I think they sold maybe 300k of them cumulatively. Heck, even HALF that number in MacPros makes them way more money and the form factor is already proven and in use for multiple configurations.

Say goodbye to the 1U pizza box. ‘twas no performa’


The “new” server options are:
* Rack Mountable - No
* Dual Power Supply’s - No
* Lights Out Management - No

In the B&W IT Enterprise World those things matter. A Mac Pro and a Mac Mini are not servers. Sure they can run a server OS, but that doesn’t make them servers. Sorry but putting a Mac Mini on a shelf in a server rack does not make it rack mountable.

What Apple really needs is a TRUE desktop computer that could in turn be fitted with things like a dual power supply and some sort of custom bracket so that it could be rack mounted (fantasy I know). I don’t consider a Mac Pro a “desktop” because you would never put it on a desk. The iMac sits on a desk but it is an all in one computer without any expansion capability.

So what is next Apple? Going to kill off ARD and drive one more stake through the hearts of your customers who have Mac’s in Education and Business?


I appreciate the authors explanation of Apple’s view but that in no way is comfort to the customers who invested in xserve’s for longterm deployments. The enterprise market is very different than what Apple wants which is the consumer level gadgets. They don’t just toss out the old Apple stuff to get the shiney new Apple stuff.

From a business perspective it’s fine for people to say hey Apple is just focusing on profit, but that doesn’t mean they didn’t screw over clients with big deployments. 

It’s a very good lesson for everyone - think long and hard before tying yourself to a company that is the sole provider of the “whole widget”. If they decide at some point your “problem area” is not their “problem area”  - your toast.

Also the mini and powermac are not “Servers” no matter what Apple says.

other side

“That said, we always worry about the message Apple is sending”

The #1 reason enterprise won’t invest in Apple is APPLE IS NEVER IN IT FOR THE LONG-HAUL.

And here we go yet again.  Anyone who convinced their employer to invest in Xserve is likely to have a stunted (if ended) career.  Organizations that now have to rebuild their investments will rightfully be wondering if other Apple products (iPhone, iPad) will ultimately have similar fates.

Apple’s worst enemy is itself.  This just may permanently close the case (and the door) for Apple having a presence in enterprise.


Actually, I’ll take a swing at the Apple internal justification:

They couldn’t get the new Mac Pro hardware into a 1U box and they couldn’t justify building a 2U box to hold it and thus splitting their already low server sales into 1U and 2U (because you know if Apple “replaced” the 1U Xserve with the 2U Xserve, everyone would still scream).

That said, it’s too bad.  Enterprise was just beginning to trust Apple again.


The ?new? server options are:
* Rack Mountable - No
* Dual Power Supply?s - No
* Lights Out Management - No

In the B&W IT Enterprise World those things matter.

Well said. These are the minimum’s. Without them, no matter how good a Mini or Pro may work, they won’t even be considered by most data centre managers.


Apple is investing a lot of money to build two large data centre. People is guessing Apple is going to release cloud based iTunes service. I would guess Apple may follow Amazon and Google footstep to build a elastic cloud environment.


What does Apple’s termination of the Xserve mean.  It means business school 101.  Our elite business schools teach that corporate resources should be devoted to the activities that earn the highest return.  Apple, like most companies, accepts that principle, but with Apple there is another principle that is expressed in this question:  Can we innovate in the target market in a way that will capture and outsized share of the profits based on our proprietary innovation?  If the answer is yes or rather might be yes, Apples goes for it.  Unfortunately, with the Xserve, the many innovations in advanced design of hardware and software that produce greater elegance, reliability, and ease of use, just don’t register, with the exception of reliability, in the market for server geeks.

Server geeks aren’t moved by ease of use or even the Xserve’s functional elegance.  Geeks, after all, make their living because there stuff is hard.  That is what makes the geek valuable.  So Apple’s innovations with Xserve just don’t resonate with enough of the market to give Apple an edge, so the Xserve doesn’t earn sufficient returns, which means that Apple terminated it in favor of devoting its resources to efforts, such as iOS, iOS devices, OS X devices, and other areas, where its innovation can earn outsized rents, rather than the commodity rents and declining share, which Xserve was earning.  Apple is content to leave that commodity business to HP and IBM, because it doesn’t have any innovation that can dominate the market for 1U servers. 

Apple is redirecting its resources to those markets where its innovation in elegance, ease of use, and/or, as is the case with the iPhone, with a break through technology, like multitouch, can either dominate the market or, at least, give it a superior, market leading position.  Apple’s innovation on the Xserve didn’t and apparently couldn’t do that, so it is gone.


iceycake seems to have got closer to the truth. Apple may well be moving to a cloud based service - and just maybe there is a LOT more to it than we suspect. Perhaps for small to medium business they want to offer the back end support across the net….which raises a whole plethora of privacy, bandwidth etc issues…which they figure they’ve got ironed out as they’ve axed their server hardware…
For the regular user it could be quite awesome, but for powerusers (think design/CAD/photographers/movie makers) how the heck could they work on multi-GB files across the net?
Lots of unanswered questions, but I suspect a play primarily at the consumer level.

the question is, whats in the racks in those data centres?! wink
and if they’ve got OSX server running on a dell/hp/ibm hardware perhaps they can release a hardware key that allows others to do it?!


Now’s the time for Apple to reconsider allowing 3rd parties to make clones of Xserver-capable machines only.

The 1U format is obsolete since all the competitors have gone to blade servers. They can get lots more multicore processors in the same space as multiple 1U machines, and they only need two power supplies (for redundacy) to run the whole set of blades. Upgrades of blades is easy - plug-in new blades. Eventually ARM will catch-up to Intel at much less power usage - easy to upgrade, just change them out. Could probably double or quadruple the blade and core count with ARM processors in the future.


“The iPhone, iPad, iPod, and non-Xserve Macs make up more than 90 percent of Apple?s business.”

So Xserve takes up to 10% of the business of Apple? It’s this kind of strange information that makes me doubt the whole article.

Bosco (Brad Hutchings)

Our elite business schools teach that corporate resources should be devoted to the activities that earn the highest return.

So true. Our elite business schools no longer teach the value of long-term customers or what the promise of a brand means to them. That kind of thinking is purely Main Street thinking these days, the antiquated and quaint (“antiquainted”) notions popular with people who know their skill first and somehow make the business part work.

And haywire… They won’t license the Mac OS X Server. Just as with Java, this announcement with no realistic migration plan ready is about killing it good so there is no doubt and no equivocating later on. They know good and well that nobody with a brain is going to buy Xserves between now and January.

Google is going to start getting a lot of attention for this. Just watch and take notes in 2011.


It seems clear that the return on investment in (hardware) servers is not there.  They could easily partner with any of dozens of companies making 1, 2, and 3U servers, and just host OS X server on them.  They could provide a leading edge data center product, with no hardware investment, and still make most of the “profit” from selling the OS X Server license.

it would not surprise me if the Data Center design team choose NOT to use X-Serves.  X-Serve hardware is simply not up to par.  They probably hosted OS X server on some “Intel” servers, and found them to be much better than the current Xserve.  It is very difficult to stay competitive in the server space without focusing on it and usually supporting several Operating Systems.  It is not a good fit for Apple.  Since they use the exact same chips as everybody else, they don’t get ahead, and they cannot invest enough to create a new product every 6 months.

Of course Apple is not known for working so closely with another company.  On the other hand they could pay any number of worldwide mfg’s to design and build servers for them.  And perhaps even market them through Enterprise knowledgeable channels.

They could go really mainstream and license the OS X to independent mfg’s to sell solutiions, but that is too traditional thinking.  Keep in mind I’m talking about supporting only a limited number of server designs, NOT trying to sell OSX to clients running a plethora of desktop PC’s.

Just a thought.  It does seem to me that it is in their interest to keep efficient workgroups functioning with an OS X Server like product, and the current hardware is not up to the task for Data Centers.  Maybe January will see an announcement!


Dear Bosco:  You’ll note that I said, in addition to the principle of devoting the firm’s resources to the highest earning ventures, Apple also asks whether it can innovate in a market so that it can obtain outsized rents, that is, greater than average, rents or, in the case of its IP, even legit monopoly rents.  If the answer to that questions proves to be no, Apple, unless, there is some other strategic reason to continue to invest, will abandon that market.  Apple is about innovating in markets in ways that at least obtain greater than average profits.  Average or sub-average profits are for others, such as Nokia, IBM, HP, HTC, Motorola, et al.

Most great business leaders have adopted the same rule.  Both Lou Gestner, when he headed IBM, and Jack Welch, when he headed GE, had the rule that their companies were either going to be first or no worst than second in each line of business or they weren’t going to be in that business.  Steve Jobs is no different.

As for licensing OS X, that is a dumb move.  OS X and OS X Server are close derivatives of each other, so close, in fact, that it can be fairly said that OS X Server is simply OS X modified for special needs and circumstances of servers.  That means that licensing OS X Server would expose the Apple’s crown jewels for third parties to compete against it for no more than licensing royalties.  Apple’s past experience with licensing and its present success prove the utter folly of that course, as the profit from royalties is but a tiny fraction of the profits lost from letting third parties use Apple’s own superior technology to compete against it. 

Also, Apple may one day devise the right innovation that can win in the server market and then choose to reenter that market, but would find that it had to compete against its own OS X Server in others’ hands.  When that happened before, Apple had to bear the costs of buying out those licensees.  However, that might not be practically affordable if a licensee like HP had built a very successful business using OS X Server. 

So the meager profits from royalties are not worth the vast risks and potential losses from exposing Apple to competition from its own OS X, which is the core OS from which iOS and OS X Server are derived, and from the risks of precluding Apple making proprietary profits from OS X Server in the rack-server market, because it had licensed OS X Server to others, should Apple choose to reenter that market on some future day.


To follow up on my earlier comments…

The cloud will ONLY be a replacement once you can get a true high speed connection between the server and the client. Even then the cloud will NOT be a replacement for every instance of a server.

If I were still doing consulting and I had a Small Business client who wanted to buy an Xserve to replace an older Mac OS X Server I would have no problem with it. Now if a company was looking for their first “real” server then the Xserve might not be the best choice.

Anyone who has worked with both Mac and Windows servers will tell you that MS is light years ahead in some areas. The two that always come to mind are AD and DNS. Both are fairly simple to get set up and “just work” in a small environment. Anyone who has had to work with a DNS server in Mac OS X server will tell you that it is never all that simple. Add on top of that fact that OD will do screwy things if DNS is not set up 100% correctly and that means having both forward and reverse DNS working correctly. So to say that server geeks have rejected the Xserve because it is somehow less complicated than a Windows server is just untrue. If anything Server geeks rejected the Xserve because it either wasn’t a windows machine or because the server geeks really wanted to run some version of Linux. The hard core linux geeks might even have an LDAP user DB that they don’t want to mess with and/or kerberos running exactly the way they want it to be.

The ultimate solution here would have been for Apple to get into bed with Dell and allowed them to sell Dell branded servers with Mac OS X installed. I have often wished that I could customize an Xserve like I would be able to do with a Dell server. This would actually be a welcome change for most enterprises.


First of all, we have to understand that Apple was a very different kind of company when the Xserve was introduced. Apple was a smallish, one product company, the Macintosh, and it needed to branch out in every way it could in order to generate sales.

Although today Apple is one of the most branched companies which penetrated into number of domains of electronic industry and has a number of products which give them millions of dollars I think that it’s quite stupid thing to discard one of the oldest projects which exactly guarantees profit. I can’t understand motivations Apple managers especially in respect that lots of people use Xserve till this day.


“Finally, don?t forget: Apple sales people will be explaining the deeper facts to their customers.”

When said sales people have them. The regional Apple sales staff were just as blind-sided as the rest of us. They have no answers either.


The move doesn’t surprise me. Companies that have tried to invest both on the consumer side and the server side mostly end up only on the server side. Look at IBM - used to be a desktop staple but now it’s all about server side solutions. Microsoft is heading in the same direction unless it can miraculously convince people to embrace it’s mobile and eventual tablet strategy.

The typical customer that invests in Mac OS Server is the small business user that isn’t into rack mounted servers. I just saw the same thing with a local government office that is using a Mac Mini as their main office server.

The last thing Apple wants is to produce computers that don’t excite and delight the purchaser. A bunch of rack mounted servers running lights out in a locked room shuffling data isn’t in Apple’s DNA.

John Martellaro

Server. If you look at the (linked) chart, you’ll see that software and peripherals take up the other ~10 percent, not Xserve.

John Martellaro

Mark.  Read the letter from Alex Grossman (linked) in the Postscript.


Jobs is ruthless in his discipline:  Apple will either be best and highly profitable in a line of business, or it’s gone.  Even so the loss of Xserve is a sad thing.  If Xserve made enough money to pay for its costs—particularly the costs of Apple’s customary insistence on excellence in the design and pursuit of revolutionary or at least, useful innovation in its products and services—I’d have argued for Xserve’s preservation.  However, I would expect that such an argument would have been unavailing, as would it would have been with Messrs. Welch and Gerstner.

Bosco (Brad Hutchings)

The typical customer that invests in Mac OS Server is the small business user that isn?t into rack mounted servers. I just saw the same thing with a local government office that is using a Mac Mini as their main office server.

I know it’s comparing apples and oranges, but a typical company that invested in Xserves—not the only “typical”, but one “typical”—is a creative company doing video that offloads rendering tasks to servers. While being Mac end-to-end may or may not have made their computing experience easier (debatable), it certainly made their purchasing experience easier. And when you have rendering software that scales well with the number of servers and can also be used on the desktop or even a laptop for small jobs and produce exactly the same results, you have a whole scalable solution. Add in custom software for custom effects, and it’s easy to see an advantage of an end-to-end Mac creative platform.

What Apple is doing by telling the shops to migrate to Mac Pros is putting the end-to-end Mac equation into serious question. What are the costs to reconfigure the server room? What do we do without redundant power supplies and hot-swappable drive? The safe answer could very well be to transition to a an end-to-end Windows platform, because you can safely bet that in 10 years, there will still be several Windows companies making 1U rack-mountable servers, even if they stick the 2020 version of a Windows phone inside a box and cable it up.

Apple pretty much gave up on creatives yesterday because while they were a key profitable segment for Apple from 2001-2009, creatives are just a drop in the bucket now.


This is terrible news. The XServe is a great piece of kit that fits perfectly into our central server room infrastructure; it looks and is an excellent performer, with redundancy built in, as so well described in this article. The alternatives are not real servers and will undermine the march of Apple into the heart of our university’s network and user community. The Xserve is a brilliant solution that has been a fantastic performer for us for at least 4 years. It has earned Apple a great deal of kudos and credibility with the notoriously closed community of IT geeks. To be unceremoniously dumped is totally counter-productive and will re-ignite the suspicion of Apple in the enterprise as well as education service providers.


I think that the other important problem with Apple’s server business is Steve Jobs.  It is no secret that his passion and genius is in client computers and consumer devices, while he has no real understanding of and passion for the server room in the back of the house or of the customers who buy servers.  If Mr. Jobs had the same passion for and understanding of the server customer as he has for the consumer and for the user of client devices, I predict that he would have driven Apple’s server business to the same heights of excellence as we’ve seen in its client computers and consumer devices. 

What Apple’s server business needed was an senior VP who was an expert in servers and their customers, was steeped in Apple’s culture, and possessed of the burning passion to make insanely great servers, storage, and related equipment and services, and who fully had Mr. Jobs’ trust and support.  Such a person could have made Apple’s servers, with OS X as the base, the best in the industry.  Other server companies, after all, were and still are dying to get their hands on OS X Server, which is a testament to just how good Apple’s OS technology is. 

But the moment seems to have past, even if the man, such as our own former Apple employee John Martellaro, were available to run Apple’s server business.  And even if the man were available, I wonder whether he would get the same support from Jobs that Bertrand Serlet, Bob Mansfield, and Scot Forstall can count on, without which nothing much could be done.

Bosco (Brad Hutchings)

It’s not even that deep, Nemo. The Mac is on a trajectory to be a charging stand for the iPad (credit Steve Gillmor for the phrase) or whatever Apple’s pet rock of the year is in 2011 and 2012.

I wish John wasn’t so cryptic about Grossman’s memo. There is the father of the Xserve, who left Apple a couple years ago to start a company that makes products that work with the Xserve. He got investors, like Intel. He was obviously landing customers, which means Apple was obviously selling Xserves, because his investors would have shut him down much sooner if there was no business there. He was that mythical server champion you wanted Nemo, and he obviously thought that the Xserve was in good enough hands to leave and build something on top of it. At least for a few years until he could abstract the need for a server running Mac OS X away.

Anyone else starting to get the feeling that Jobs is going batshit crazy at a frantic pace for some reason other than business? Stay tuned next Friday when Jobs cancels something else we never imagined.

John Martellaro

Nemo:  We had that guy you mentioned, an expert on servers.  His name was Alex Grossman.  I won’t presume to speak to why he left, that’s his story to tell.  But I can point out that ActiveStorage is immensely successful.

Bosco: I sense that Mr. Jobs is on a furious, self-imposed schedule to make all things PC obsolete before his work is done. That’s OK. Without Mr. Jobs, we’d all still be running DOS on PCs with 1.44 MB floppies & VGA graphics.  My vote is for Mr. Jobs moving us forward at as fast a pace as we can absorb. I’ve always been a fan of the future, and we’re all having fun seeing how Apple creates it.

Bosco (Brad Hutchings)

Well, when y’all are done immanentizing the eschaton, Windows 7 and Linux should still be around to do the things you need done. So sad to watch this unfold.


John:  Apple has apparently made the decision that today, unlike yesterday, servers, beyond the Mac Pro, are no longer strategically necessary to its success.  And you seem to agree with that assessment.  And I am in no position to take issue with that assessment.

But, if that assessment is correct, I think that Apple was wise in abandoning a business that the CEO has no passion for, doesn’t understand, can’t manage, and probably doesn’t trust anyone else to manage.  Steve Jobs is not the remote delegating type.  He is intimately involved in the development and marketing of all of Apple’s major products and services.  He drives Apple to excellence, to insanely great products and services.  He takes home or otherwise test the builds of prototypes every Friday, but he can’t do that with servers.  Serves are the one product that Steve can’t look at and say:  Yes, that is it; that is what will delight the customer and not only satisfy his needs but satisfy needs that he never knew he had.  For servers, Jobs would have had to rely on someone like Mr. Grossman to drive the products and services to excellence and to foster and recognize breakthrough innovation.  However, I don’t think Steve Jobs could do that, sit back and let another man run the server division, whose products, services, and customer are too mysterious to him, whose core products he can’t test or even directly use, so the fate of Xserve was sealed once it was no longer essential, as it once was, to Apple’s success.


Finally, Apple releases a roadmap for the enterprise. Unfortunately, all it lists is a one way street right now.

For what it’s worth, I hope that this does mean that Apple moves towards licensing OS X Server on 3rd party hardware. It’s not a good comparison to look back on the Clone days of Mac OS.  Licensing server software to qualified 3rd parties and products would be a lot different for a few reasons:
1) the people running the hardware aren’t purely end uses and generally know the difference between faulty hardware and faulty software. That wasn’t he case with the Mac OS on a PowerComputing machine, and it was part of what gave Apple a bad name.
2) A lot of current 1U hardware is already compatible with Linux/Unix/Mac OS X. If you have Intel CPU’s, all it would take is Apple’s blessing for you to boot into Mac OS X Server. That wasn’t the case back in the days when Apple used Motorola processors.
3) The $500 license for OS X server could actually create some revenue, unlike licensing the client OS did back in the 90’s.

Keep in mind, Apple already allows you to legally virtualize OS X server, but not the client. That’s already a big difference between the client and server OS, and just half a step away from allowing you to do it on non Mac hardware.

I’m hoping that Apple’s rationale is that they are never going to be “all in” in the server hardware market. So maybe it makes sense to concentrate on the software, and let others step in to fill the void.


The real problem with licensing OS X and any of its derivatives that Steve Jobs recognized immediately on his return to Apple is that the parties that you license OS X to can successfully compete against Apple on price, because their costs don’t include the margins need to fund innovation.  The licensees simply take Apple’s innovation and sell computers for less than Apple can sell them for, because Apple’s costs include the costs of innovation in software and hardware, elegant design, advance materials research and implementation of those materials, high quality construction and manufacturing, marketing, ect.  And if Apple tried to include those costs in the royalty for the license, it wouldn’t work, because those costs are impossible to anticipate, and even if Apple could include those costs in the royalty, no one would be interested a license that is that expensive and/or that required a licensee to design its product to Apple’s exacting specs.

So licensing for Apple, which needs to make good margins from hardware, is a losing proposition that would, if adopted, destroy Apple, because it would either deprive Apple of the revenues needs to innovate, as Apple competed on price, or it would result in Apple losing too much market share, because people, both enterprise and consumers always go for cheaper, even if cheaper is a bad value.

Frank Lowney

The fact is that w/o a rack mounted server, Apple will soon disappear from most data centers.  MacOS X Server will first lack upgrades and then be discontinued.  The promise of the MacOS X Server UI has never been and will never be realized. 

That’s not to say that Apple won’t offer a cloud-based MacOS X Server.  This would be ground breaking in the finest Apple tradition.

Failing that, folks in higher ed and research at least can get along quite well with open source software and commodity Linus hardware and OS.


Apple will ultimately kill all server products because everything is moving to the cloud. Ten years from now, it won’t make sense for any small or medium sized business to operate their own servers. Everything will be provided as a service, and even companies who have all their own proprietary stuff will develop and access it globally from the cloud. Crazy? We’ll see.


I know from experience that Apple has regarded the Mac Mini as a robust Mac OS X server platform for several years. For redundancy, two Mac Minis are more cost-effective than any dual-corded 1U rack-mount from Apple, Dell or others. Since then, Apple has produced a dedicated server version Mac Mini with dual-HDD. At around three times the price, is it surprising that the X-serve is not selling so well?

More important, and more interesting, one of the features of the original X-serve was that Apple used the X-serve to power its own data centres, including the phenomenally successful iTunes store. Now that Apple is about to bring online its own billion-dollar server farm, what equipment will they be putting into it?

Let me guess. Mac Minis without enclosures and with SSD’s in place of the dual-HDD’s, mounted in racks, blade-style. Now or later, they may be more like Apple TV’s. Speculating further, it would be possible to get 12 stripped-down Mac Mini servers, or possibly 48 Apple TV’s, on a single 1U chassis.

Put it another way - why would Apple abandon its server market, when Apple itself is possibly its own best customer?



Privacy concerns, security concerns, and availability will make that a very hard sell.


Unisys. License OSX server for use on Unisys-provided hardware. They don’t play in the consumer space, so little chance of being cannibalized as in the past.

Small business = Mac Pro server or Mac Mini server
Enterprise = Unisys licensed hardware/software combo

More businesses are willing to let Unisys in the door than they are willing to allow Apple in the door.

Otherwise, why do you think they announced that Unisys will be handling Apple’s Enterprise effort?

//just a thought.

Just a reader

your toast

Should be “you’re toast.”

Just a reader

?The iPhone, iPad, iPod, and non-Xserve Macs make up more than 90 percent of Apple?s business.?

So Xserve takes up to 10% of the business of Apple? It?s this kind of strange information that makes me doubt the whole article.

There’s still the iTunes store and app store plus licensing fees for patented IP.


Jobs is ruthless in his discipline:  Apple will either be best and highly profitable in a line of business, or it?s gone.

Jobs needs to realize a server room isn’t an iPod.  You don’t just throw it out & move your library to another device…

A lot of people had doubts about XServe in the first place.  To play in the server business you need the commitment and support that many thought Apple couldn’t deliver.  Those people have now been proven right.

What especially hurts here is Apple has burned people who were just getting over what happened with Newton and OpenDoc.  “Fool me once, shame on you…”.  And of course these are the same people who make corporate IT buying decisions.

If I were an IT manager, I would have no choice but to conclude that Apple truly is a consumer toy maker. For serious corporate investment and deployment, I would definitely look elsewhere.

Bosco (Brad Hutchings)

Let me guess. Mac Minis without enclosures and with SSD?s in place of the dual-HDD?s, mounted in racks, blade-style.

That depends on what the servers do. If they are processing bandwidth heavy transactions, sure that will work great. They’ll want lots of low power CPUs and network connections crammed into the smallest space. If they are doing heavy computation, they’ll want lots of higher power CPUs, adequate RAM and storage for the tasks. If they’re leaning toward the former, there are OTS mini-ITX configurations that would save them considerable money versus cannibalizing Mac MIni boards.

The data-center-centric view of this decision isn’t right. Neither is the suggestion that we’re beyond need for businesses to have servers because everything will happen in the cloud. More likely that there will be a “fog” as well—servers and P2P apps scattered all over the place using excess last-mile bandwidth—and that the balance between “cloud” and “fog” will be in flux for eternity.

Just Me

The discontinuation of the XServe is a real drag. However, my major concern is the future of OS X Server software. It seems obvious that GUI in OS X 7 is going to be a departure from the past. That might be fine for the client version of OS X, but it’s not going to add anything to OS X server. It’s time for OS X server to diverge from OS X client. I would feel much better if there was any information about the future of OS X server.


Interesting article. Even more interesting comments.

In my non-expert opinion, this is, to quote the Bard, much ado about nothing.

Most intriguing are the comments suggesting that SJ is incompetent or that the business community will look at Apple’s decision on the XServe and rethink their use of Apple’s other products.

People fear what they do not understand, which underlies their fear of change, as it represents a step into the unknown. If the past decade has shown anything (apart from Jobs taking Apple from the brink of extinction to the most profitable, influential and respected - by peer review - tech company on the planet - so much for his incompetence) it is that Apple have a grasp of the industry, and its position within it, that is uncanny, almost to the point of being alien. Their thinking is clearly alien to the competition and pundits, in that their conventional wisdom has ridiculed or inveighed against Apple’s choices, often accompanied by predictions of its impending doom as a consequence of those choices, yet Apple has gone from strength to strength and now sets the industry standards. What Apple does today, others will be doing by 3 years.

Not that ‘every little thing they do is magic’ (Cube anyone?) or that they are above mistakes (hockey-puck mouse ring any bells?), but the company learns from its mistakes, and clearly does not commit itself to a course of action without study.

I would argue that this decision is not about profit per se, but business strategy. More than one commenter above has argued that Apple pulling the plug on the 1U XServe will shake enterprise confidence, however enterprise has already expressed its confidence in the XServe with its wallet, and the verdict is ‘no thanks’.

At the same time, that same sector has expressed its confidence in Apple’s other offerings, not only the iOS devices but the Mac (of late) also with its wallet. In other words, enterprise has demonstrated its ability to treat each of Apple’s offerings on their independent merits. Apple’s action, in this regard, is an acknowledgement of reality.

For enterprise to be shaken by the termination of a product that it has emphatically rejected, from a company whose other products it has eagerly embraced makes little sense. For the market to punish Apple for dropping a non-selling product line also makes little sense, and would be an aberration of precedent. This is the sign of a company that is responsive to its user base, but also pays attention to its bottom line. If it is going to engage, there has to be a market for the service, and it has to be profitable.

Apple will either engage that service (note: service, not product) the XServe was intended to address with a new vehicle or it will not, but if it does, it will do so with something that people will want to use, as demonstrated by uptake and gains in market share and profit to the company. That service, and any new product(s) it requires, It will help drive and shape the industry, the way Apple’s other product lines do. If it cannot play at that level, the level that Apple clearly expects for itself, then Apple is unlikely to continue that product. But to devote resources to a product or service that the majority of users of such services are unwilling to purchase makes no sense for a business that sees itself as an industry leader, and not a follower.

Bosco (Brad Hutchings)

El Reg weighs in, with more quotes from Alex Grossman. Note the comparative space requirements of Mac Pros.

@wab95—Your analysis misses the value of end-to-end single OS environments, very common in mid-sized creative companies. They will quickly see the value of migrating to Windows, where there are multiple suppliers for everything they need.


Your analysis misses the value of end-to-end single OS environments, very common in mid-sized creative companies


You are undoubtedly correct that the end-to-end single OS market has value, and your point is well taken. Indeed, my point was that Apple have recognised that they do not compete effectively in that space with the 1U XServe, certainly not among medium and larger businesses. Apple’s performance in other areas have raised the bar, I believe, across their portfolio, and if they are to compete for that market, they will do so with a product and/or service that garners more than 1% market share and contributes to more than 1% of gross profits, where I suspect the XServe fails miserably on both counts.

If critics are correct, that this current move hurts Apple’s business with enterprise, then I posit that the place to look for empirical evidence is not the blogosphere, but Apple’s bottom line, specifically two quarters hence, allowing for inbuilt market inertia and the upcoming holiday season. This would show up as a secular trend, a population-based change over time, due primarily to contraction in sales to enterprise.

The other point, and here I respectfully disagree with Timothy Prickett Morgan’s piece you cite above, is that rather than SJ and/or Apple acting with the impetuosity of a three year old, Apple takes a long view and plays a long game. If Apple decide that they want into the enterprise market, it will not be as a bit but principal player, and that they will as carefully set the board for the enterprise as they did for the private consumer market such that, when they make their move, it will be emphatic and from a position of product/service strength.

Then again, I’m no expert, just a bloke with an opinion.


I have a hard time believing that a company (any company) would survive, not to mention ascend to the position of a second most valuable company on the planet, by being run by an impetuous, childish CEO who makes foolish, irrational decisions without thinking of consequences.

As we saw in recent quarter, the Mac business is still a very sizable chunk of Apple’s profits, and creative professionals fuel a pretty sizable chunk of that Mac business. To believe that Apple would cut the Xserve business, consequently undermining the Mac business (by alienating creative professionals) is to ignore the situation. Apple is where it is because it carefully weighs and analyses each segment of their business.

I firmly believe that the percentage of creative professional shops (photo, video, design, audio, music) that are wall-to-wall Apple (including Xserve in data centres) is fairly low to negligible. Those that have data centres that can take advantage of a rack-mountable hardware more than likely have at least hybrid (multi-OS) environments. The smaller ones (with ‘server closets’ rather than data centres) were likely running tower-based servers already, and won’t really get affected by the discontinued Xserve. Loss of the few creative shops that were all Apple, and will now (or in the next 3-4 years) switch to all-non-Apple hardware will be as meaningful as loss of professionals who left because they couldn’t get a non-glossy display on an iMac.

Bosco (Brad Hutchings)

@wab95: This won’t hurt Apple’s bottom line. A complete evaporation of Mac sales won’t (by itself) hurt Apple’s bottom line that much anymore. What that would indicate might shake confidence in the company, but that’s another story. Apple is transforming itself into a CE company. If you fit their definition of “consumer” then they’ll probably have a new pet rock for you every couple of years, while getting further locked into their personal computing model might not be so painful.

The big quandary is for those current Apple customers who don’t fit into Apple’s consumer model. You’ve liked Apple Mac stuff because you think it’s easier to use or maybe just because it’s familiar or maybe because you like the industrial design. Whatever. But you have a problem when it comes time to replace equipment because the puck that Apple is skating to is no longer in your skating rink.

The Xserve betrayal could start a seismic shift in Apple brand perception, making it uncool with the hipsters who still need to pay their bills. Or, it could be a small puff of smoke as if it were dragged off the dock inadvertently. But you won’t measure this on Apple’s bottom line and trace it back.

Chandra Coomar

Why do you think that Apple plans to abandon servers and OSX Server at such a time in its history of failure in the enterprise market? Apple has never had a higher profile in, higher interest levels and a stronger foothold in the enterprise as it has now. The iPhone and iPad have brought significant and rapidly growing awareness of Apple as an innovative and highly capable player in IT.
OSX Server may thrive in future as a hybrid OSXiOS offering that moves enterprise computing into the realm of ease of use, administration, maintenance and TOC. Offer the ITMs the same familiar creature of old, with an easy path to transition to even better solutions. Incorporating aspects of iOS into OSX/Server compromises nothing. The iOS mode can be trialled, used or ignored according to circumstances or preferences.
Apple walks away from outdated old ways to promising new ones.
I think this augurs very well for Apple in the server space, both in hardware and OS/Pro software futures.

Log in to comment (TMO, Twitter or Facebook) or Register for a TMO account