Fusion is a type 2 hypervisor that runs on top of Mac OS X, so it is most likely a mac version of VMWare workstation, while ESX Server is a type 1 Hypervisor (also called a bare-metal hypervisor).

I agree with you, Apple has never been really serious about entreprise servers (it's mostly an on and off relationship) although Mac OS X Server is quite a capable server OS, and very easy to use and administer.

Regards,

Provisioning, management, monitoring, service, hell, it's not just about what application is running on the box for it to have a place in the data center.

Wasn't the apple AIX called A/UX? Remember once when selling computers back in the early 90's in Dallas we had some Apple corp. people come in for a customer event, still remember the name and title "Software Swami" of one of these guys.

I don't get the bit about Rosetta.

What would that have to do with running Unix applications on OS X? You can do that right now - all that's necessary, AFAIK, is that they be compiled for Darwin (as opposed to, say, FreeBSD) and run in X11. And X11 is already there on all Macs in /Applications/Utilities. Now X11 is not as usable or as aesthetically pleasing as Aqua, but, hey, these are servers.

http://www.macports.org/

As Wikipedia comments, "[MacPorts] allows the installation of a number of packages by simply entering the command 'port install packagename' in the Terminal, which will then download, compile and install the requested software, while also installing any required dependencies automatically."

Or there's Fink:

http://www.finkproject.org/

As for the overall strategy, I take it there are other areas in which Apple thinks it's easier to make money (at least right now). They may be wrong, but then again perhaps not ...

From where I'm sitting, the market for proprietary Unix servers is in free-fall - people aren't even willing to pay cash for Sun these days, when they can stick Linux on disposable Dell kit.

That doesn't strike me as the kind of market Apple are going to be particularly interested in, and it probably makes more financial sense to make OS X (client) integrate well with Linux based server networks, while perhaps providing some nice native client tools for remote database admin, etc.

Remember Apple isnt a "Computer" company anymore, they're a "Consumer Electronics" company.

Fusion is not ESX Server and of course your imaginary server is going to be more secure than Server 2008..

1/ Apple don't have the depth of engineering staff or experience to build a complete range of servers. To do so would require taking on more engineers and that type of hiring activity would have been spotted. Apple could wait a while and pick up the Sun x86 team cheap, but given their success or lack of it that might not be the smartest of moves.

2/ Mention of Sun brings us to another big hole in your argument - if vendors with experience and real enterprise market presence are seeing a downturn in sales, why do you think Apple could do any better? OS X is not a good enough differentiator, especially compared to cheaper enterprise Linux, and has a much lower application count.

3/ Which brings us back to the biggest hole which you mentioned but skated round in your article - nobody develops enterprise apps for Mac OS. And simply wishing them into existance doesn't work (simply ask Sun about Slowaris x86, and then consider Sun started wishing with a weight of the installed SPARC Slowaris base to beat the developers with). Apple has nothing to tempt the developers with and only the pain of porting apps to a non-existant product.

4/ Of course, Apple could pay companies to port their apps, but then I'm not sure Apple has the cash to spare. Whilst you diss Itanium, it had two big things going for it Apple doesn't - a market-leading chipmaker and a real enterprise UNIX vendor, and both Intel and HP were willing to invest big money and resource into porting programs, money and resource Apple probably doesn't have without stunting development of other product lines.

5/ And Itanium also shows another big hole in your daydream - Itanium can not only outperform the competition, but by being an easy porting platform it also offers multiple OS options at a cheaper price than enterprise competitiors like SPARC. Apple using Intel offers little differentiation from any other x86 vendor already offering Wintel/Linel kit, and would probably be more expensive. If they go back to Power then IBM will make sure their Power kit is cheaper.

6/ Which brings me to the final point - businesses make decisions on real business factors and not "coolness". The cool factor has allowed Apple to overprice their products for ages, but if they jump into the server market then they have to compete on price, something they don't like doing. Price is one key reason the Wintel PC won the desktop wars and relegated the Mac to the marketting graphics niche. I don't think even Jobs has the ego to want to try that kind of fight again, especially when he could be up against real interest in "free" Linux SMB servers.

First, the usual unfounded Observation that an Apple server OS would be more secure than W2K08. Care to justify that a little Tim, it's pretty sweeping statement given that there have been no major security issues with W2K08.

Then there's the whole issue of building an enterprise hardware and software support organization to go along with the servers. Just having the boxes isn't enough, you need a full enterprise support organization or folks will just laugh at you.

We all know that if Apple did this, they would bring out an under powered server at twice the cost that has 0.2% of the applications and force you to install iTunes and Quicktime on it.

Seriously, the server tech guys will have spent a fortune learning to work a server on either the MS platform, Unix platform, Linux platform..... how many of them are qualified to run iServer? None? The cost not only in underpowered shiney hardware, but also in re-training their tech staff to run an enterprise server along with all the required software needed for a server, would price all but the most ardent mactards out of the running.

The reason Apple doesn't go into the enterprise market, they would fall flat on their face, it's not like a phone (which isn't a good business phone) or a music player (which isn't the best music player) its mostly mission critical applications which need to work, and they will not be able to beat MS/Linux in that area, both MS and Linux excel in the enterprise market.

Apple could buy SUN Microsystem and get the server knowledge it needs ... merge the good of solaris and OSX and be a good candidate for future server market especially in SMB and visualization markets

As Christophe points out, Fusion is most definitely not ESX. It's closer to Ye Olde Virtual PC.

As for Apple servers, I doubt there's enough mark up on the hardware to justify doing it off their own hook. Remember, IT buyers are more tech savvy and won't pay a premium for relatively cheap intel or AMD tin (particularly at the workgroup end) just because it has an Apple logo. Perhaps a licensing deal with one of the big players would work - they do the tin, Apple bring in OSX.

At the high end, I can't see they'd make enough Market penetration to cover the development of high end tin.

Additionally, as for OSX as a server OS. Sure it'll be secure etc, but what justifies it's premium over Linux? Also, there aren't any proven security benefits over Windows 2008 (contrary to the authors bias view). Maybe a small business server with an OSX flavoured mail/Sql/proxy/web server bundle might work, but Linux and Windows are established here also.

Apple should consider allowing OSX to be licensed seperately and virtualized. This would allow Virtual desktops as well as servers. How cool would that be?

Sorry no really why should apple bothe making servers for the enterprise. In fact why should apple bother with enterprise level apps. If apple want to make a server product then they should make a home server which any idiot can use and have your own mail, web server, file server, wiki , media server whatever. In fact OS X server already has most of it built in so already on the way.

Leave enterprise servers to hp and dell and vmware, theres no market for apple in replacing entire enterprise systems.

I have had the unfortunate job years ago to work in a Mac server room! We had to get it all ported to Windows, Exchange etc (yes the mail server ran on a Mac too!). As I had limited experience on Macs I had to replicate their environment back in our office to plan the port of their systems & mail...

Never seen a place like that since and hope I never do!

For all the Linux/MySQL fanboys, once you get into a corporate environment the support contracts, training etc required to provide proper support & maintenance of these 'free' products is shocking! The same can apply to freeware/shareware - tucked in their license agreement is a 'corporates not allowed to use' clause... big fees! Even Adobe Reader has this tucked away, corporates roll without an agreement at your peril.

The closest Apple gets to being business-class is media editing platforms for audio and video production, beyond that, There's not much more. Apple hasn't really penetrated business because as it seems, most of Apple's future vision are based mainly on rumors that have "leaked". You can't market rumors to a business-type, there has to be a viable road map, especially for backend \ systems where the price tag has to be outweighed by the power of the system.

Apple should stay away from big business and stick with personal machines and media editing systems when it comes to computing.

I thought Apple were waiting for some third party to come up with a killer workgroup app for Xserves. They would then sell the solution AS/400-style. Then competition would create more third-party expertise who would sell at the department level to corporates.

It would have been OK if someone else had done the work, but no-one did.

Apple doesn't care about that market, firstly because they can't supply it well. Secondly, if Apple wants to continue making the world a better place (I think that is their motivation), wasting time trying to get on the good side of a load of enterprise IT departments is a waste of their expertise.

Apple currently does sell some business servers, so that's one reason to take it seriously. If you are going to seek customers for any product your company makes - you should take that product seriously to maintain your company's reputation for quality.

As well, selling business servers based on the Macintosh architecture, and doing it well, could presumably help Apple get experience which it can apply to building better Macintoshes for the ordinary desktop.

The fact that there's money to be made there, on the other hand, is no reason by itself; after all, Apple could sell cars, flowers, or groceries and make money.

But if Apple has strengths that make it easy to make and sell servers, then that's a reason to take advantage of an opportunity to make money. Of course, since it is a computer company (among other things), that may have been taken for granted.

Fusion is not ESX. Nor is Rosetta going to be a good path for running linux apps. Rosetta is about architecture emulation. Linux binaries are typically (or at least available) as x86 - yes you'll need a compatible library layer, a sort of Linux "WINE" (Take a look at the definition of the WINE acronym). As OS X has a BSD userland I'm not sure why you wouldn't just compile all those OSS apps anyway.

That aside Apple seems to have serious security problem and bug acknowledgment issues that would not allow a lot of organisations to take them seriously.

apple hasn't been able to sell into business in the past, and are unlikely to be taken any more seriously now. they'd also have to create an entirely new support service in order to cope with enterprise customers.

what's the point? given Sun's decline there's no obvious benefit to me to apple distracting itself with enterprise systems.

What was Apple's profit margin again? 30%?

What's the profit margin in the server market? 3%?

Now think again whether Apple will ever enter the server market.

Bearing in mind that OS-X is basically a bastardised BSD, why would you want to spend the extra money on an Apple server when you can source perfectly serviceable and cheaper kit from countless other manufacturers and stick a proper BSD (or even Linux if you really want to) on it?

These are servers! So all the (questionably) lovely Apple GUI stuff is irrelevant (and once you take that away, all you have left is the bare OS - ie - BSD). All you need is a bare-bones OS that can run your apps. Any intel box running BSD/Linux will do that, from the cheapest of the cheap up to proper server kit. And you won't be shelling out unnecessary extra cash to "upgrade" your OS (licence) so that it can accept more than 10 incoming network connections (a bazaar restriction in this day and age).

I can see the writer of this article *really* likes their products, but this is just silly.

So you are saying that Apple should launch a new server line, in times of economic difficulty, for the small startup company that doesn't want to use Windows or Linux on it's servers has has no existing IT staff. That's a tiny market at the moment, even if the basic idea was good.

Any buisness with existing server infrastructure isn't going to suddenly convert to Apple for the server 'user experience', when the only people that ever directly log on to it are the IT staff. That's a non-starter. And yes, user experience is really what sells Apple products. Noone buys them for their openess, compatibility, performance, stability or security (unless they are spectacularly ill-informed, or just terrible at maintaining other OSes)

I'm not a big Windows supporter, but I don't buy that an Apple server would be automatically more secure. Apple have an atrocious record of managing vulerabilities (even MS score better here), and have had a lot less time to get their server platform to a point where the apps, rather than the OS are more likely attack vectors.

So a company that prides itself on NOT being used for "boring work stuff" (rather they prefer to be used to pose with) should break into the dullest of the enterprise markets, which has bugger all profit margin compared to what they're used to?

They also pride themselves on having the cuddly-uddliest of GUIs. Which isn't that much use on a server.

I guess they COULD create a beginner-server or something like that. Maybe a home server for emails and the like. And the fanbois would love it and the mass media would hail it as revolutionary. The Linux communities would decry it as they've been doing it for years, even if it did take 6 months to learn how to hack through the config files to manage it, Microsoft would point out that it's had their Server Editions for years- and even webservery stuff included with all versions of XP Pro since it came out- and then someone would notice that Facebook or Myspace lets them show their black and white pictures of them looking moody just as well.

And every time their "I'm a Mac" advert was shown, Lester would be there to tell everyone that they're not all graphics editing and looking good, they've also got a server business- and how conventional is that?

Anyway, they'd only go and create the "Ballamer" server size rather than the standard "U"s. It'd be thin enough that you could fit 5 into a 1U box.

Exactly. Apple practice aggressive market segmentation (as whined about at length by the people who want a cheap Mac minitower), and the falling, low margin server market isn't likely somewhere they'd go, no matter how hard some people wish for it.

And having run OSX Server, it's OSX with some GUIs for the included open source daemons. "any x86 Linux/BSD" box would work just as well. It's clear from using the thing that little expense has gone into it and it's not a priority.

I worked for a company that ran an Apple A/UX box, which ran on PowerPC hardware and had a nice, MacOS front end. We ran an Oracle DB on it as well as file sharing for our all Mac clients. It just worked. Always. I shut it down once after a couple of years and vacuumed the 3/4" of lint out of it (wool carpets), and fired it back up. We replaced it with one of those Apple-branded IBM AIX boxes, which was more complicated to administer, but it, too, just worked. For years.

I recognize that things are more complicated now, but remember, some people don't want to have to putz around with the server all the time - or any of the time unless necessary. Was the extra cost worth the virtually no maintenance? Absolutely. Whether it makes business sense today for Apple to make server hardware, I don't know. But I know they brought server capabilities to the unwashed masses who don't want to have to work hard just to get and keep a server running.

HP, Dell, IBM have the standards-based server business locked up, with also rans Fujitsu, Sun, Supermicro.

Lenovo is attempting to get into this space and Cisco is rumoured to be launching blade servers.

Is there really room for another? What does Apple bring to the table that's unique, other than higher pricing, a true white box (!) and compulsory automated iTunes and Quicktime updates?

So you were employed to work on Macs, but as you didn't know anything about the hardware or operating system you replaced it all with Windows. I very much doubt you "ported" anythnig, as that requires an understanding of both environments. So, your employer was stupid to employ you as you had the wrong skill set, and your opinions of Apple hardware and software are pointless as by your own admission you don't know how to use them.

Now for the punchline. I do know MacOS and Apple hardware well, but I still prefer a mixed Solaris and Linux environment.

Personally the single biggest issue with Apple in the server space is proper support. Apple have zero experience in this. They provide no support for their server products, essentially outsourcing what little they do to companies that send around a guy with a replacement part. If you buy into the server market it isn't about shifting the boxes. It is about a long term provision of support for those boxes. That is why IBM, HP, SUN, even SGI (actually especially SGI), can charge the premium. If you have a problem there is real knowlegable support, local parts, small number of hours turnaround. And not just for hardware. But the whole integrated system. Also, certified known good configurations. Complex multivendor solutions that "just work". And an eco-system that you can grow into. Apple have no relationship with large scale storage, backup, archive, data repository management etc. Sun bought STK, IBM roll their own, and very nice it is too. And so on.

Being able to provide this level of assured business is not something that can be created quickly. Compaq had to buy it (in the form of DEC.) It is arguable that HP's ability in this area is still the DEC heritage. It takes decades to build up the expertese, the people, and eventually the trust. Even a small SME, with not a lot of money to throw about, should be able to see past the raw price of the boxes, and any glitz they come with, and see the critical value in risk management that comes with getting into bed with a long term real enterprise player.

What Apple do do, and do very well, is create the same well integrated, "just works" system for the desktop. They do it better than anyone else. It took them a long time to learn how to do it, they forgot a couple of times and had to relearn. MS have never learnt, Linux probably never will, and the proprietory Unixes will never need to. But the task needed to play in the server business needs the same level of investment.

Only four OS's are Unix 03 certified, and on X86, only two: Solaris 10 and Mac OS 10.5. That presumably carries some weight in this discussion.

http://www.opengroup.org/openbrand/register/

It's a tough lot really... They could build a low end presence with X64/X86, but they lose out on the highly profitable high-end of the market. Also, to be a real server vendor you need to scale. That's why Dell will never do anything more than differentiate on price, which is not where Apple has traditionally wanted to be. X86/X64 does not have the reliability features required to compete with the high end CPU's (Power, SPARC, or even PA-RISC which is now EOL'd).

I don't really see Apple doing Sparc since they are already an Intel OEM, though I think it wouldn't be a bad idea. They've already thrown out PowerPC, so I doubt they would go Power... That really only leaves Itanium. Itanium is dangerous as every major vendor except HP has thrown it out. Intel has had the worst record of any chip manufacture with Itanium. They've been late with just about every release. Even when they came close to releasing on time they've had to reduce the promised features. Even Fujitsu is rethinking their Itanium decision and going back to Sparc. Intel seems to have no interest in Itanium any more as HP owns more than 95% of the market there. Sun sells 4x the number of Sparc servers than HP sells Itanium servers... That is not the scale that Intel envisioned, so Intel will require that HP keep throwing money at them to keep Itanium going (what was the last bribe to keep developing Itanium? Was it 3 Billion?). How long will it take for HP to get tired of being the only one putting money into Itanium? It's clear, to me at least, that HP will have to move everything to X64 and just abandon Itanium. Apple certainly does not want to have to go that route, though they've shown no problems with changing CPU's in the past...

So, what should Apple do? Just partner with a real Server vendor for scale and then keep differentiating at the low end with OS/X. I think Sun would be an interesting merger/purchase for Apple though. When Sun was the purchaser I did not think it would be a good idea. With Apple as the purchaser, it might just work...

And there he comes out of the woodworm - mr PHUX fanboi.

Let's face it - nobody in the real world uses or likes PHUX except you and your little buddies.

Now just climb back under the rock and the rest of the world will move on

Disclaimer - I actually have had a fair bit of experience of supplying server based systems to Small and Medium Businesses (SMBs). I have installed Linux, Windows/Dell/IBM/HP and Apple based solutions

The cost of the kit is the least of an SMBs problems. Many SMBs are held hostage by their (outsourced) IT after the kit has been installed. A quick look on Apple's site shows that a realistic Xserve small rack mounted server with 2x 1TB SATA hard drives, a Quad Core Xeon, 64MB Radeon graphics, a single power supply and a superdrive with unlimited clients costs ~2,400 quid - AppleCare Premium Service and Support costs an additional 635 quid. A Mac Pro (Tower) with similar specs and 256MB Radeon costs less than 1,850 quid - Now add Mac OS X Server (Unlimited-Client) at 616 quid and an AppleCare Protection Plan for Mac Pro at less than 200 quid... The 19 inch rack solution comes in at a 3,035 as opposed to 2,666 for the tower. For a 50 user system they are both a bargain.

With Linux you could probably do something similar for less than half of that - On the other hand, if you used a Microsoft Small Business Server (SBS) based solution, the software and client access licences (CALs) ALONE would cost you 4000 quid for 50 users; so add say another 1300 quid for the hardware. Oh! you wanted a SQL server as well sir! So that would be another box and licence and CALs - or you could spend about the same and use the MS SBS Premium product.

Compared with the pratting-about your Linux consultant will probably charge you for, and Microsoft's CAL Tax, an Apple Server solution may be a bargain.

In an ideal world Apple would make lots more use of it's OS X Server platform.

I really wish they would come out with some low cost XServes. My small business would eat them up in a second, and we could do away with outsourced IT support for good - we don't really need it now, except for to add new users to the domain (ffs!)