Apple’s Xserves were awesome. Not cheap. Only marginally competitive. But a joy to work with because of all the nice little details like a FireWire port on the front so you could boot it off of an iPod.
Yeah I remember Xserve! But if anything it seems like it was evidence that Apple wasn't really able to "get" the server market, where hardware is a commodity and price and compatibility are basically the only concern. Not to say that couldn't change, especially if their own processors had a big advantage in price/performance, but there hasn't been a whole lot of evidence that they have what it takes to succeed in this space.
While most of your points are true, I will quibble with this:
there hasn't been a whole lot of evidence that they have what it takes to succeed in this space.
Does it matter? In 2006 and 2007 there were legions of pundits, columnists, and "experts" who said the same thing about Apple making a telephone.
I wonder if there's a market for a server optimized for iOS builds. Cram as many Mx chips and possible into a rackable case with abundant cooling. Certainly it would be useful to Apple internally. But I wonder if there are enough iOS developers with the budget to make this workable.
I think the pundits in 2007 were obviously wrong to anyone who followed Apple. I mean, perhaps you wouldn't have predicted the level of success, but it was clear that the iPod was all the experience they needed to pull off the iPhone, and it was also clear to anyone who used phones that if they could get the internet to work well there was a big market.
In contrast, I would say there's very little evidence Apple can pull off a majorly successful server play. It's not impossible, and I think the trends are potentially in their favor, but it definitely doesn't play to their strengths. Apple's biggest strengths are very consumer oriented, and the trade-offs they typically make are not great for an enterprise crowd (expensive hardware, willingness to abandon legacy compatibility, etc.).
There's probably a smallish market for very specialized servers that Apple could fill well, and maybe they will. Long term maybe that could translate to a wider market if they really make an effort. But I think it's at the very least a significantly longer time horizon than the iPhone.
Sure as I said, anything is possible, and that can change, but to be fair the iPhone is a consumer device, which was firmly within Apple's wheelhouse already when they were developing the first model.
Do you really think there's a market for dedicated hardware for iOS build servers? I mean there is some market for this since CI vendors already do iOS builds, but it seems like a bit of a stretch to imagine there being enough demand for Apple to actually put the R&D and product development resources into making this a reality.
We know they have tens of thousands of linux boxes running iCloud and the App Store and whatever else... if they can use their own ARM servers and reduce electricity use by 50%[1], you've gotta think they're looking serious at that, purely for themselves.
Not to mention buying their own hardware will obviously save them.
Purely for themselves, or also for entering the cloud infrastructure market. In this day and age if you are big enough (Apple surely is) and happen to have superior server hardware it wouldn't really make sense to sell that hardware, you'd rather keep it and rent it out.
An Apple alternative to AWS would be interesting, but it would probably be heavily geared toward iOS and Mac app backends, with little flexibility outside those use cases. The Xcode integration would probably be amazing, though.
It's hard to imagine Apple really getting the cloud market, or their product offering would look like. Cloud by necessity has to be pretty agnostic and flexible to meet the diverse needs of customers. Apple hasn't really demonstrated an ability to succeed with this type of product, and leans heavily toward very tailored, limited, "apple knows best" types of products.
The cloud is changing. At my gig we've embraced GCP's managed cloud services, and I do not want to go back to 2015. The operating cost savings (no devops people!) and the low cost makes engineering systems for cloud-agnosticism very expensive. If Apple provided a range of managed services that met clients' needs, they could find a lot of success. Would they beat AWS? Highly doubtful, but Apple already provides a lot of cloud data services to Apple developers and consumers as part of iCloud.
I'm hopeful with respect to Apple dogfooding its silicon in the data center because it may feed back into robust ongoing support for Unix/Posix on their publicly-facing platforms.
> I'm hopeful with respect to Apple dogfooding its silicon in the data center because it may feed back into robust ongoing support for Unix/Posix on their publicly-facing platforms.
I mean that's one possibility, but what seems much more "Apple-ey" (aka likely) is that they would release a very much locked-down platform of tightly controlled, managed services which would perform well and would have a beautiful dashboard design, but would offer a small subset of the functionality which is available from AWS or GCP, they would make some weird choices - like only allowing Swift for configuration files - and all of this would be very proprietary in nature and distant from their consumer products.
That's not what I'm getting at. Regardless of the details of the cloud offerings, migrating from Linux to Apple Silicon would help the MacOS and iOS developer experience, because it'll all (MacOS, iOS, Apple's rack OS) be using their BSD-derived Unix OS under the covers.
Right now Apple's cloud engineers are to some extent helping Linux be a better OS by filing bug reports, creating PRs, etc. That activity, directed toward an Apple Silicon rack OS, would redound to those of us sitting behind MacBooks, typing into the terminal, treating it as a Unix box with nice driver support.
The question is whether you can make a competitive cloud infrastructure service without x86. Imagine that would be a nonstarter for a lot of the market, at least now.
arm in the cloud is already a thing. A lot of companies have been moving towards server-less infrastructure in the past few years, and I can't imagine it makes any difference at all for a large set of use-cases whether you're executing your 50-line go lambdas on x86 or arm.
The vast majority of the potential market is running somewhere between some and all of their stuff on x86 stuff in the cloud. Of course someone could launch an ARM-only cloud. They'd either be severely limiting their potential market, or they'd have to convince prospects to migrate to ARM. Certainly not impossible, but is it competitive?
What migration? Most of the code I have run in the cloud in the past 5 years has been in the form of managed services. I don't really know or care if AWS is running my code on ARM or x86.
Moving their own servers to Apple Silicon would save a lot of money, buth when buying as well as operating the machines. But once they do, resurrecting the Xserve should be a further source of income and raise the chip volume (which drives down costs). Imagine the margin on an Xserve powered by Apple Silicon vs. having to buy Xeon chips.
> Not to mention buying their own hardware will obviously save them.
Maybe, maybe not. Apple doesn't currently produce servers, so it would probably take a lot of time and money for them to build out that capability to the point where it's break-even or better compared to just buying commodity hardware.
Or heck even just the general purpose chip space in general.
Apple's current ARM offerings are strictly locked to their own hardware, and there's been no hints that they may move into the ARM-PC space selling general purpose chips and letting other vendors produce board. I'd expect to see that first before an Apple silicon powered server.
Yeah Apple's never been a component company, they're all about vertical integration. Still it's hard to imagine if they actually do end up producing the best price/performance processors that these would be relegated to consumer products given the massive opportunity it would represent in the server space for cloud-based companies to bring down costs.
