iCloud vs. AWS: Apple Has Considered Competing With Amazon in Cloud

[Powerbooky]

On Topic:
Apple probably has the money for it. The question is where to build them and will it fit with the "green" targets?
But I'm guessing that they also consider the risk of being bullied by the EU/EC with new demands: side-doors on their datacenter for so-called competition. 🤣

iCloud is awful for storage. I constantly have to fight with Files to have things remain downloaded and sharing via a simple link is near impossible. Especially since the cost for iCloud storage right now is substantially higher than competitors, why aren’t they passing on those cost savings now?

You can pin certain folders or files to remain downloaded since macOS Sequoia. Funnily enough I had the reverse problem in earlier macOS versions: iCloud drive stubbornly downloading everything, filling up the local drive with stuff I didn't need. I use iCloud drive for AfterEffects and others, these files can get very big.
Fortunately there was this little tool iCloud Control that helped a lot in removing downloads. Today I don't need it anymore... iCloud drive works fine now.

Sharing stuff also works fine... I can easily create a direct download link to a file. Whenever I need to send multiple files for a project I create one ZIP file rather than sharing a folder - usually this is the fastest way to transfer.

 

[jonnysods]

iCloud Drive just isn't ready for a business environment, I've considered trying to keep my files there but sharing, offline/online file management, reliability etc just seems to not quite be there. I use Workspace right now for files and I'd love to pivot from them.

 

[TriBruin]

Exactly. Anyone who thinks "Apple has M* processors they could compete with AWS" doesn't understand what AWS provides. They wouldn't even compete with Oracle cloud, instead Apple would be competing with players like Digital Ocean.

The better path for Apple if they wanted to do this would be to make a deal with Amazon to provide Apple Silicon as an option. Unlikely to happen though. Plus, ARM processors are already very common and popular with the hyperscalers. I moved most of my AWS workload to ARM years ago for the savings.

Maybe I am cofused by what you are asking here, but AWS already offers Apple Silicon as an option for Cloud Computing, but it is limited to macOS.

Amazon EC2 Mac Instances

Amazon EC2 Mac instances are built on Mac mini computers and feature the macOS operating system. Mac instances offer customers an efficient development environment by providing macOS with the benefits of AWS.

aws.amazon.com

Now, it is very limited. You are just renting a Mac mini in an AWS data center. It is not nearly what you can do with other AWS virtual machines.

 

[neuralengine]

But would they be able to bring themselves to let it have buttons?

AWS has a lot of buttons.

You don’t need that many buttons, but not sure if this is the best fit for the most button-averse company in the industry.

 

[Ghengis LeMond]

Building a business like AWS with zero experience in the space (iCloud does not count) is an order of magnitude in complexity and difficulty greater than building a car. They should stick to phones and movies.

 

[ksec]

Another one of those Apple PR articles for Apple to negotiate with Amazon on AWS pricing.

 

[tipoo]

I've always wondered about this. They make world leading perf/watt silicon, they're making their own servers with private cloud compute, they're working on the ACDC AI training chips akin to H200 etc. They seem to have a lot of the parts in place for competing in the server where electricity cost becomes a dominant factor and they lead on perf/watt. I'm just not sure about their experience and willingness to be a server shop for everyone else.

But hey make me an Apple Silicon Xserve again with 4 Apple Silicon sockets per 1U server, I'm in!

 

[nt5672]

The problem with this is that Apple's operating system software server components are terribly buggy.

Remote access locks up daily, is not available to reconnect for some time and can often require a reboot.

File sharing just stops working from time to time.

Time Machine is just a complete failure in a server environment because it cannot back up to Network Shares reliably.

SSH seems to work fine.

I think if Apple tried this they would have to significantly increase their support staff.

 

[TriBruin]

Personally, and I say this as someone who manages thousands of macOS computers, Apple should stick the client side in the Enterprise. It is already a struggle each year to get management controls for the new consumer faced features that Apple introduces each year. Or getting them to work with Enterprise systems.

There is still a lot Apple could do with Consumer level cloud computing. I am a shocked that Apple has not introduced a cloud based Time Machine solution. With dedicated Time Capsules discontinued years ago and network backups iffy at best, it would be a great opportunity and a way for Apple to sell more overpriced iCloud storage. Honestly, I would consider cancelling my Backblaze subscriptions for an Apple solution.

 

[QuantumPulse]

Wouldn't mind lower prices, iCloud is one of my most used services

 

[jonblatho]

Doesn't amazon already have graviton systems for exactly this reason? Apple isn't the only tech giant utilizing ARM in server environments. Meta does this too.

Yep. If my numbers are right, the majority of our CPU spend at my job is now AWS Graviton chips, and its share is growing every month (about 48% to 56% this year so far). We're certainly not spinning up anything new on x86 where we can help it. It's nice being able to develop on our Arm-based Macs and deploy to Arm-based servers.

It's tough to imagine what "killer feature" Apple would be able to provide in this market, and they'd absolutely need (at least) one at the outset in order to be successful. At first I thought maybe cheaper macOS VMs, but the market for that is mainly iOS (and other Apple platforms) developers, and they already sell Xcode Cloud for that.

More likely, it would cost more for less control and they would somehow present it as a "feature." No, thank you.

 

[vgmrm]

This sounds like the XServe to me: equipment that could have been, but never was.
While they were great machines for their time, they had serious shortcomings and were far from cheap.
The current rack-mountable MacPros are just an attempt to turn a farm tractor into a rack-mounted unit. They are an extremely expensive piece of junk with no incentive to be productive.
The only niche market I can currently see for them is Apple system virtualisation, primarily for licensing reasons rather than performance.
For any other task, you can get Linux equipment offering equivalent power at a fraction of the cost.
Then there's the problem of integrated memory and storage: no one in their right mind would invest in equipment that can't be upgraded.

 

[vantelimus]

The only customers who will care about Apple's specifics are those that specifically want something from the M-series, but that's a niche of a niche in the market. You'd need to develop that market for decades to get it to a point that it would be worth taking seriously.

Sure, if you are only skating to where the puck is now. While it is wise to be skeptical that Apple has figured out an enterprise strategy, predicting it would take "decades" before Apple would have a Cloud offering worth taking seriously discounts 1) the niche they might capture or create; and 2) the speed at which technology advances now days -- a decade can be a lifetime.

 

[indychris]

Dang! Only a few years to have smart glasses like Meta and now thinking about cloud services like Amazon? Tim ‘The innovator’ Cook just can’t be stopped!

 

[mdatwood]

Maybe I am cofused by what you are asking here, but AWS already offers Apple Silicon as an option for Cloud Computing, but it is limited to macOS.

Amazon EC2 Mac Instances

Amazon EC2 Mac instances are built on Mac mini computers and feature the macOS operating system. Mac instances offer customers an efficient development environment by providing macOS with the benefits of AWS.

aws.amazon.com

Now, it is very limited. You are just renting a Mac mini in an AWS data center. It is not nearly what you can do with other AWS virtual machines.

I didn't even realize they started doing that. It goes to show how large AWS is. Unfortunately, it's useless to me as all my servers run Linux though maybe I could move to BSD. I'd be curious of their usage beyond 'iPhone app test runner'.

 

[obezcinnet]

In my experience, iCloud works fine for backing up files when I’m upgrading or restoring my phone, but beyond that, I’m not interested. I use other more reliable file storage options, and iCloud ranks dead last in that regard.

What is reported on this article is not a service for the end users like yourself. so you can't use this service even if it existed.

 

[Aichon]

Sure, if you are only skating to where the puck is now. While it is wise to be skeptical that Apple has figured out an enterprise strategy, predicting it would take "decades" before Apple would have a Cloud offering worth taking seriously discounts 1) the niche they might capture or create; and 2) the speed at which technology advances now days -- a decade can be a lifetime.

I wasn't saying it would take decades to make an iCloud competitor to AWS. I was saying it would take decades if they tried to use the M-series as the primary differentiator. Doing it that way would take decades of investment that they've demonstrated neither a willingness nor desire to make. And given the costs with transitioning to the M-series (sure, it may cost less to operate, but the up-front cost for retooling is massive in terms of time, money, and risk), you'd need to far exceed competitors like Snapdragon, which come with far fewer constraints and far more ability to work with the tools that their potential clients are already using (i.e. 90% of the ongoing cost savings at a fraction of the migration cost).

I also think you're underestimating how slowly things move in this part of the tech stack. Swift is 11 years old and only got around to supporting Android—the most-installed OS in the world—2 weeks ago. Apple Silicon is 15 years old, but Linux kernel support (i.e. to run natively on the metal) only arrived with a stable build in late 2023, is still in active development, and—tellingly—hasn't been receiving support from Apple.

Meanwhile, AWS already offers M-series chips running macOS to anyone interested, thus undercutting any offering Apple might provide (also of note: it didn't exactly make a big splash). Azure is 16 years old and laser focused on enterprise. Again, these services are commodities, and Apple has never demonstrated an ability to work well in commodity markets.

It's also worth considering customer satisfaction with what's already out there. I hated my cell phone prior to the iPhone. I have a VERY distinct recollection of thinking on my first day with my first cell, "This is terrible. I bet Apple could do this better". A few years later, I was proven right.

In contrast, we frequently use AWS and Azure for the work we do (I'm in software consulting), and these platforms are rock solid. They work great. Clients are happy. Engineers are happy. The only complaints I hear with any regularity—depending on which stakeholder you talk to—are (1) cost, (2) difficulty navigating the interface, (3) confusion over which specific services to use, and (4) lack of support for X technology. Apple is well-suited to differentiate on #2 or #3 (and poorly-suited to differentiate on #1 and #4), but #2 and #3 are the two least important axes on which to differentiate, because those are up-front issues that stop mattering once you set things up or once you develop any expertise in the space (which nearly everyone in the space has, since these are aimed at enterprise, not laypeople).