Is there a reason why Apple servers did not catch on?

Question

Is there a reason why Apple server technologies like Apple Workgroup Server and the Apple Network Server never gained widespread use and popularity?

Answer 1

I can tell you why it never caught on in my lab. We bought a Mac "Blue and White" to try out OSX server to add to our collection of Suns and (rather old) VAXes. While the hardware was nice, it was spendy. The OS was eccentric and had no obvious advantage over more mainstream *ix OSes. Linux was a rising star, and soon we could also run Solaris on cheap commodity hardware, so they were more attractive.

After a CPU upgrade, the Blue and White proved to be an excellent number cruncher, and with Linux as its OS had a long, productive life.

Answer 2

In a word: Linux

Why would anyone pay a premium for Apple’s proprietary hardware, when it is far cheaper — and more flexible — to get an Intel-based server and install some flavor of Linux on it?

For historical context, in the beginning there was the Apple Workgroup Server (1993 to 2003) that addressed the idiosyncratic needs of Mac OS at the time. Apple Workgroup Servers ran A/UX as well as classic Mac OS. Linux was only a glimmer in the OS world’s eye in 1993; the Linux 1.0 kernel was only finalized in June 1994.

But as time passed, the Apple Workgroup Server made way for the Xserve in 2003; Xserve was introduced in 2002 and the Apple Workgroup Server was sunsetted/obsoleted as a result.

Xservers were very sleek and cool machines, but they only lasted until 2011. From launch to 2006 they were PowerPC (G4/G5) based but in 2006 they were Intel (Xeon) based.

But by 2011, these Intel based were simply not selling. According to Steve Jobs himself:

“Hardly anyone was buying them.”

Now why they weren’t buying them? I have no solid citations but by 2011 Linux was 16+ years old, had significantly matured and — in my humble opinion — overshadow Mac OS X in terms of features, flexibility and ease of use as far as server software went.

In short, Linux maturing represented a sea change that redefined what the server market was and could be. And, as a result, swept Xservers out of the marketplace. It was just not economically feasible to justify the purchase a pricey Xserver when any old server in the world could run Linux cleanly and efficiently. More opinionated details follow.

---

I managed a few Xservers when they were effectively abandoned and EOL (End of Life). What pain in the a** they all were!

That said they were a two sided coin:

• Cool looking hardware.
• Abysmal OS software.

And there is a third side to that coin:

• Both locked into Apple’s consumer OS mindset.

Near the end the only real reason we had them was for use as a print server. And we tossed one of them and replaced it with a Mac mini (to act as a server) and the worst aspect of that was how to best rack mount them. We had some weird rack mounted thing that could hold the Mac mini; but it still sucked because it was basically a cake box on a rack with a Mac mini inside of it.

Software-wise, I heard my former co-workers managed to get rid of it and set up an Ubuntu Linux print server on some old hardware. An old rack mounted server that could die any day. But that was fine to last until a new budget could happen and a newer server could replace it.

And that new server? Cheaper, faster and more resilient than the Xserve and Mac mini could ever be. Old Windows server about to kick the dust? Just install it on Linux and give it a new life!

Or how about setting up a web server? Xservers could run Apache and PHP, but always with utterly ancient and out of date versions. Why deal with the headache of maintaining that junk? Just go with Linux and maybe setup Virtualmin and Webmin on it to more easily manage it.

At the end of the day, Apple — at it’s best — makes highly refined consumer devices.

And servers are not highly refined consumer devices. They are boring machines that run practical system software in a very practical way. macOS is not server software; Linux is much better suited for servers.

macOS is cool but not really as up-to-date as far as server software goes. Why hobble yourself with macOS when it is always trailing behind what the pros use? I mean look at how lazy Apple is with macOS nowadays; imagine if that mentality were applied to server software? I mean you don’t have to imagine; Apple was quite abysmal dealing the headache they self-created by insisting all Mac OS X/macOS be closed-source Darwin/BSD-based.

In contrast, Linux is open source and very server oriented. Yeah, you can have GUIs on Linux, but ultimately Linux is all about the command line and server software. Linux wins the server software war!

Again, the Xserve servers physically looked cool, but software-wise they really stunk. Good riddance!

Answer 3

Apple servers were found where there was a lot of Apple clients. For example, a university IT department that hosted the Mac OS (faculty or labs) would often have Apple servers. But I think you're asking more why didn't Apple servers make themselves more common in mixed computing environments.

There was a brief period in the early 2000's when the Xserve was making some inroads partially due to the vector crunching abilities of the G4 and G5 PowerPC processors. This wasn't so much for being a network server than it was more for getting these powerful CPUs available as server machines. I was at a company around that time and we were looking at buying into that ecosystem for some heavy number crunching duties, but Apple representatives shied the decision makers away from it, presumably because of the impending Intel transition. Once the Xserve became just another Intel x86 box, there was no advantage to having the Apple product and they fizzled out, from what I could see.

Answer 4

At least one Apple Xserve tester basically said that for them FiberChannel was required if a box would be deployed in the DC.

They tried a Xserve, was impressed by the speed but the machine was a non-runner in a center with hundred of machines.

Answer 5

I considered buying a few for software build-and-test (I produce libraries for quite a few platforms), but there was no real advantage over using Mac Minis.