https://www.zdnet.com/article/why-apple-had-to-kill-the-xser...
Unfortunately, Apple makes it somewhat uneconomical to use Macs in small bursts, as their license requires that Macs are leased for at least 24 hours.
The reason Apple doesn't do well B2B is nothing about their technology or marketing, it is likely all about their business operations. Apple is notorious for driving their own product-development lifecycle with heavy priority on their own design language. This is completely the opposite of how B2B sales work, where it is not uncommon for some development cycles to be dictated directly by a few (or even one) customer. I can't see Apple playing well with these expectations.
Development and sales (and support). Look at how many SKUs competing laptop manufacturers have because enterprises will specify some combination of ports and memory and CPU, and if the manufacturer doesn't have that particular combination, there will be no sale.
(Having said that: it's not obvious to me whether this is still as much of an issue as it used to be.)
Apple toyed with becoming more enterprisey at various times but their ultimate success came mostly from BYOD and enterprises proving Mac options because that's what many of their employees wanted. They do have business sales but it's still more of a pull than a push thing.
Now, Xserve didn’t work for a lot of reasons, but Apple exited the server market, and even stopped making the Server variant of macOS (even though it was just a handful of utilities at the end), largely because I think it realized two things:
1. It could never truly compete with Linux (and to a lesser degree, IIS) without making significant tradeoffs that are anathema to Apple (lowering prices significantly, giving up usage control, and decoupling vertically integrated hardware/software).
2. The cost of truly making a B2B play would require resource investments and commitments into an area Apple doesn’t need to care about and that would come at the expense of areas that are both more profitable and less of a PITA.
Apple is the richest company in the world and amongst the most profitable tech companies, why would they bother with enterprise when they can allow partners to deal with the integration side and focus on doing their stuff their way.
(There is one notable instance where Apple embraced enterprise, and that was when Apple added Exchange support in iPhone OS and admittedly, this was a huge deal and was ultimately a fatal blow to Blackberry. But it’s the exception that almost proves the rule. Apple added support for Exchange but left it up to others to figure out the broader MDM situations.)
Having said all that, I have always been surprised Apple hasn’t offered some sort of build/test service themselves and integrated it into TestFlight. It could be a driver for the all-important services revenue (in my imagination, the annual developer fee would cover a certain number of build minutes a month and then additional minutes would cost money). I have to think that such a solution is either in the works or that Apple looked at the support challenges and just decided to let other people do it at scale for them.
The thing is that much of the cloud is very much a commodity and Apple has successfully made its products and services anything but a commodity.
Yep. There have been a couple of specific things. Exchange and now MDM profiles. Probably a few others. But minor tweaks to remove major enterprise blockers. Small investments to remove complete showstoppers is just common sense.
>Apple hasn’t offered some sort of build/test service themselves and integrated it into TestFlight
Seems pretty logical. Who knows? It's certainly the trend with container platforms generally.
AWS was nothing if not a "B2B play" from the beginning. Having developed their own expertise with data centers and distributed computing, Amazon proceeded to sell that to other businesses.
In fact, as far as AWS is concerned, Amazon is just another big customer — and not even the biggest.
I may not be the typical end user, but I’m also not a business.
I predict they will build customer-facing 'cloud-powered features' where 'third party app code' can run on the apple cloud infrastructure 'on behalf' of a particular apple customer.
They will present a utility computing model to customers -- where each customer sees 'cloud-computing usage charges' displayed by app.
Ultimately, the cloud-compute/cloud-storage charges will tie back to the customer associated usage costs for storage or some other abstract usage based model of for cloud resource costs.
"Apple Cloud: build features for apple customers using usage-metered cloud resources -- with the customer paying for the usage charges they incur."
- They would be vertically integrated with Apple Silicon
- Apple themselves are increasingly depending on cloud services
- Async/Await in Swift will likely land next year, making Server-Side Swift much more appealing
- Apple is greatly increasing cloud / Kubernetes hires
- Could share a single Apple Silicon ARM architecture from client(ios) to dev(mac) to cloudApple, as a company, does not "do" long-term support; they're the anti-Microsoft when it comes to backwards compatibility. As soon as you start being an infrastructure service provider that means you HAVE to offer an LTSB branch of your OS, and Apple does not want to do that (and macOS has stagnated a lot since Apple devoted most of their OS engineering efforts to iOS). Apple doesn't offer a "real" server OS SKU at all (the "macOS Server" app package is for SOHO management, not for use as a high-availability application server OS).
...which means that Apple would have to provide an offering using Linux or BSD - such as a "naked" Darwin distro - but again, Apple does not want to have to support that, and I'm sure devs don't want to see the IaaS OS scene fragment further, and Darwin is far, far removed from being yet another Linux distro.
Amazon offers ARM cloud systems - why should Apple risk their profitability by competing in an arena that doesn't pose any risk to their bottom-line? And what do they have to gain?
I would love to see Apple have an LTS branch of macOS, but I just don't see this happening. Too much would need to change in the company for that to happen.
From a macOS developer's standpoint, it's really a huge PITA there is no LTS. But, we adjust. Because we have a low-level product (compiler), we sometimes get into a situation where we need to sunset a product before we would normally do that, just because the older product doesn't run on the supported macOS versions. Doesn't happen often, but it does happen.
Obviously would work as FaaS or PaaS.
I think they don't want to cannibalize Mac sales. If you want to make an iPhone app, you have to buy a Mac to do it. They don't want to make it easy for you to rent one just to compile the app.
Some companies only charge for commercial use of their products. So the “it’s free” argument doesn’t really hold water.
Companies have tried to “rent DVDs” digitally by playing the DVD on their hardware and streaming it. It got shot down by the courts.
> issue opened October 2013
> __ is broken on MacOS
> __ should fix it but I don't have a Mac to test.
> I do have a Mac but ___ can’t be replicated on my specific dev-tainted environment, plus I’m using macOS Y and you use macOS Z which is a pain for me to set up a VM for, if at all possible.
The raise of a few free macOS CI mitigated that a bit but the offer is honestly poor (except at CircleCI, props to them), and seems like Apple could care less so much they’re now actively making that harder than ever.
If Apple should be sued for anti-consumer behavior this is why... It costs time and money to fix software with every MacOS release and I'm getting close to suggesting we add a 50% Apple tax on products to pay for the additional support it entails.
As a matter of fact I went through the trouble to set up 10.9 through 10.14 in VirtualBox, driven by gitlab-runner via the virtualbox executor[0] (which is pure genius: it uses linked VM clones from a reference VM to spawn parallel runners, as well as using snapshots once SSH is ready after first boot for <2s spawn + full state reset).
Given the painful macOS on VirtualBox situation, at the time I wanted to move to VMware Fusion[1] but lacked time to implement an executor. It seems someone took that in their own hands and implemented a specific runner (in Swift!).
[0]: https://docs.gitlab.com/runner/executors/virtualbox.html
[1]: https://gitlab.com/gitlab-org/gitlab-runner/-/issues/1679
[2]: https://gitlab.com/gitlab-org/gitlab-runner/-/issues/1679#no...
> A "lease period must be for a minimum period of twenty-four (24) consecutive hours."
[0]: https://news.ycombinator.com/item?id=25562532
So you won't be able to rent out one device to dozens of tenants per day who each use it for ten minutes or so while ensuring inter-tenant isolation.
Last, those Apple rules might as well be non-enforceable legally, but Apple might still retaliate by e.g. removing apps from their store that were uploaded by tenants of those EU companies.
I'm not sure if it's accurate, but I and many others see Apple as a fairly vindictive company, from stories about how Jobs operated to how they treat developers that speak out. My impression is that it's part of their DNA now.
Besides that, I forgot where it came out about the amount of money Amazon retail makes on it’s dedicated Apple store. Also, Apple and Amazon Video have special deals where you can do in app purchases of video within the app on iOS devices and the purchases are charged directly to your Prime account. No one else can do that for physical goods.
Apple and Amazon have been chummy since Jobs licensed the “one click patent” when the iTunes Music Store opened in 2003.
Yes, no, and yes.
Apple can't practically enforce that against you, an individual using a mac in your home. But then again you're probably not reselling your mac's functionality directly for profit to others. They for sure can seek to enforce it with AWS or others who sell it commercially. It's the reselling of their software and functionality for profit that is covered by the terms of the agreement.
I did find it interesting that Apple took the entire last run of that hardware for themselves, so that they could run their own internal VMware clusters on real Apple hardware for their developers.
There’s no way that Apple could build an AMI to run on hardware designed and built by someone else. Apple has a small number of people who really know what they’re doing, and they get shifted around to whatever project Apple thinks is most important — whether that’s hardware or software.
Apple needs to own the whole enchilada here, for multiple reasons. So, they can sell Apple hardware to AWS for inclusion inside of AWS data centers. But it has to be real-deal Apple hardware, running real-deal Apple OS, even if it is under a hyper visor like VMWare.
Server OS seemed to be targeted at small businesses to run on a Mac Pro or Mac Mini connected to NAS and hosting internal applications. But it wasn't very different from regular OSX.
If I remember correctly they previously bundled icloud services where each user got a slice of storage on the cloud, so I could see them extending that model to a time slice of compute as well, hosted on some form of K8s. I think the previous model was too limited for most use cases, but it’s not fundamentally broken.
I don’t think the economics of the M1 with 32+ cores make sense unless they go down the route of either build a cloud platform and amortise the cost of the chip design over a larger number of devices (or a chip let approach, but I’m not sure this matches their general design approach).
Also I’m pretty sure they think they have a competitive advantage over traditional x86 and ARM vendors so they’ll likely want to strike while that still holds.
The average Apple consumer customer can make plenty of use of cloud storage. What use does the average user have for cloud compute resources?
Unless Apple starts offloading some processing to the cloud, and gives them some free, I'm not sure how regular users would use this, and offering it for free assuming the majority of people won't use it is just asking for some popular app to come along that takes advantage of that and screws up the economics. Not to mention people probably won't be happy to have a limit and pay for overage use for stuff that is likely provided by free on other platforms (if Google Photos prettifies your photos for free and Apple does it through metered compute with some given free, that's bad optics).
My understanding was they ditched Azure.
Netflix doesn’t use AWS for video, Facebook doesn’t use AWS for most stuff, Google and Microsoft obviously don’t now because of competition but they didn’t before GCP and Azure were things.
They're Apple after all, so they can build macOS to boot from a standard environment - UEFI, kernel that doesn't check the copy prevention string on SMBIOS, etc.
Us, mere mortals, are more restricted, as we don't have the keys of the kingdom.
I've read before on HN about people using Mac hardware in production, when they needed access to macOS graphics API, but since this only concerns a software implementation, and hardware was standard until M1 processors, it's reasonable to assume that Apple could port that part of OS to Linux, if they needed to use it from servers.
Why?
For one thing, the tcp stack is ancient, and easy to syn flood. (It was forked from FreeBSD in 2001ish, before FreeBSD added syncookies and syncache). Synflooding os x makes the whole network stack unresponsive, including other interfaces like localhost.
https://en.wikipedia.org/wiki/Xserve
With their new CPUs maybe it is time to bring them back?
I mean surely they are not going to license Dell to put a M1 chip in a rackspace computer.
Ridiculous :-))
I do wish they have less restriction on macOS usage timing. And a services for Web Developers to test their site on Safari. Which is currently impossible to do with buying a Mac.
I wonder how "usage" is defined here: for instance, if I run a managed service on a fleet of Macs in the cloud, which serves many end customers from each machine, is this allowed under the terms?
IANAL (and haven't read the full EULA) but probably. The 24 hour restriction seems to apply to leasing. I would assume that running a multi-tenant managed service (or, for that matter, serving web pages) from a Mac would not be seen as leasing.
Not necessarily. A Mac Pro would work well for single-tenant usage where you want to run a series of applications that can each burst to the whole machine but not take long to finish.
That being said, I don't think Apple has enough people who would be interested in hosting on MacOS to justify an Apple cloud offering. I think they know that.
Is it really? Lots of devs use Macs, yes. But compared to the sheer number of regular, non-dev users, that has to be tiiiiny.
I found lots of stats on the internet about what percentage of developers use Macs, but none about how just many Mac users are developers, so it'd be fascinating to see that number of anyone can find it.
Developers fuel their platform. Without developers MacOS and iOS are nothing. Apple knows this.
That's the real target.
And you can see it by how they treat the development tools (stuff like this: https://news.ycombinator.com/item?id=25563276 or the many deprecations, for example: https://news.ycombinator.com/item?id=25562604). Or CI/CD setups using their tools.
People like iOS because of the software, both the OS, and the software which runs atop it. This is why WWDC is Apple's single biggest event of the year.
Likewise, people don't buy Macs because they are metal and have a shiny fruit on the backside. Developers were among the first to migrate to Apple back in the 00s when they discovered its Unix underpinnings made it great for doing web development.
This idea that Apple doesn't care about developers doesn't hold water. If they didn't care, they wouldn't have bothered to add Hypervisor support into the M series CPUs. This is a feature which has almost no purpose outside of software development or as a server platform and Apple doesn't make dedicated server hardware so where does that leave us?
However, I don't quite believe that there really are 20m active developers on the Apple platform. Evans Data estimates that there is a total of 27m developers worldwide [3]. It's not plausible that 74% of them should be writing software for Apple platforms. On the other hand, not all developers using a Mac are developing for the App Store. A lot of them will be Web developers.
I wouldn't be surprised if developers' share of Mac sales is 30%. Maybe it's just 20%. I don't know, but it's definitely not tiny.
[1] https://techcrunch.com/2018/06/04/app-store-hits-20m-registe...
[2] https://www.macrumors.com/2020/01/13/apple-mac-shipments-dow...
[3] https://www.future-processing.com/blog/how-many-developers-a...
Are you serious?
Apple's single biggest event of the year is a developer conference. They aren't putting on week-long events for video editors.
They also provided a bunch of Mac minis to Mac Stadium for OSS developers to build on.
They added virtualization to the M1 (it isn't on the A series) and have HyperKit built into the OS is all specifically for developers.
Well, Apple isn't giving them away, but they do indeed sell them to Mac Stadium.
They could probably just kill that part and make enough money on iOS anyway.