Bring on the goats! Apple's cloud failure demands further sacrifice In four decades, Apple has done more than most companies to seamlessly blend software and hardware, creating beautifully functional products. However, software (and hardware) no longer live exclusively on our devices, and the one thing that continues to frustrate Apple is the cloud. Way back in 2011, Apple chief executive Tim …
And whilst I wish Efrati and Nellis (whoever they may be) the best of luck securing a highly paid consultancy gig at Apple, I expect that even with the backing of El Reg, they won't succeed, and the clever folk at Apple already know all about clouds, and are doing something about it right now.
I wouldn't say they are bad at cloud services (they were once bad at cloud services), They now run some excellent cloud services that, for me at least, have been very reliable. The problem they have (that I will get too), is something different.
First what is good,
I've had Apple TV since the the second gen streaming device hit the shelves. It has always worked as near as dammit flawlessly and the model of rent and stream or buy it once, stream it as many times as you like, that Apple pushed is simple for the end user. You tend to just rely on it and forget how reliable it is, which is testament to its success.
iCloud file management was for years vey much simplified, but worked well as well. They struggled for a bit with drop-box style general file management supporting folders and files (folders and files were for a long time a bugbear if Steve Jobs because the user interaction model is so needlessly complex and through the history of computing has been the cause of much wasted time and data loss). After a shaky start (let's not mention yesteryear file management services they tried, which were dreadful). Now for file sync it works very well, and the only feature it wants for is folder sharing.
Maps famously is a similar story, getting off to an even worse start. But now becoming something far more useful. It wasn't reliability of serving up the service, but the bad quality of the data that was the problem. It's only recently I have been found I have no need to reach for Google maps though and that isn't because they have completely reached parity. They haven't. It's because Apple has now crossed the "good enough that I can't be arsed to reach for Google maps" line.
App based sync for notes, Apple email, calendars, reminders and photos have worked for me near flawlessly. Though I have heard lots of complaints from some people about calendars. Also for me, for many years inviting non Apple users in heterogenous calendaring environments (which is basically outside an individual company true for everyone) was for years pretty bad, especially for including/inviting people who might have got one Apple device, but don't use it as their main device but by a year or so,p ago it seems they had solved these problems and it has been pretty good.
Apple Music got off to a feature rich but very shaky start and whilst not terminally unreliable, was annoyingly glitchy. How Music Match service tracks got overwritten (or appeared to get overwritten) by DRM Apple Music streaming based tracks WAS a huge failure and caused some users a lot of pain. They have fixed that now though and it now works very well indeed. The UI categories remain a dogs dinner though. But once you dive into them and avail yourself of the human curated suggestion lists, the service shows some real strengths over rivals,
iCloud backup has always worked flawlessly for me doing exactly what it was intended to do.
The new photos service is one of those in the best category - it works so well you don't notice it. And the new photos app and iOS device storage space optimisation strategy works so well few understand or need to understand how awesome it is (e.g, the strategy of storing low res versions on iOS devices and retrieving high res only when needed and zoomed in has been implemented seamlessly and the non destructive editing in a simple but powerful photo management app is just great).
Apple's Office apps have matured into excellent very easy to use office alternatives to MS Office that do exactly what you would expect such an alternative to do. Really the only things they miss are 1) Scripting language support to the level of VB. Needed by enterprise (though less and less in a cloud based world). 2) Numbers lacks Excel style Pivot Tables - these are a huge feature for those that use them and are missed sorely. This is more a comment on the excellence of this feature MS implemented in Excel than criticism of Numbers as a spreadsheet. Numbers, Pages and Keynote are a definite usability upgrade over Office, though of course in most places a feature downgrade (only Keynote, the least important of the trio bests its MS rival all round). The cloud sharing and browser based versions of these apps have been a quiet success which beats Google equivalents by a long way in the usability / end user experience department.
IMessage also has worked reliably as a service for some time (the biggest issue being a while back now and that was the one that affected users who switched to Android, where their Apple account and ID would remain as an active entity getting in the way of messages Apple user friends were sending them).
So overall their cloud services, with some notable past problem areas, work very well across the board. So what are they doing wrong?
The Bad
In a word connectivity. They haven't provided the API glue that allows their cloud services to be anything much more than silos. Great for personal inter device transfer/sync, but not great participants in programmable data flows that are increasingly becoming available for coordination by other services, such as IFTTT. This failure to participate will bite them hard because it means their services will become marginalised. They still, just about, have time to address this deficiency, but the problem is they are showing no signs of the desire to do so. The only significant move they have made in this regard is two years ago providing a competent but still somewhat restricted developer web API access to a users iCloud account.
Do they actually need to build out the cloud infrastructure?
What about putting an abstraction layer on other cloud services so they can use whoever is cheaper and/or actually working at any given time? After all, the key selling point is supposed to be "computing/storage" as a commodity, just like power or the ISP networking, and its the data that is precious and needs protection (encryption + backing up) and management?
Surely you encrypt before storing it remotely?
Certainly things like reliability and backups are dependent on the service they make/buy, but again, if possible it would be better to duplicate on two providers so if one goes TITSUP and/or hikes the price too much, you keep the other and migrate to a new "2nd copy" for the next contract negotiation round.
Well if you really want continuity and resiliency, you're going to have multiple encrypted copies and multiple providers with a wide geographic separation and distribution in case shit happens as a matter of course, not just as leverage with your vendors come negotiation time. Disaster happens a lot more often than people seemingly would like to think.
Apple has no experience in this space. 2 main reasons:
1. They didn't build a reasonably large amount of their OS underpinnings. My understanding is that the BSD or similar licenses are responsible for this. Polite correction please.
This may also be why running OSX on non-Apple HW is possible and so simple. You didn't write this, you don't understand this and you can't control this.
2. Apple is a consumer electronics company, not a computer company. It's like asking Hermes to build a power station.
It would be a power station with some nice (shit, overpriced) watch straps though...
1. Google and Amazon don't build their OS. Apple built Mac OS. Pre-Apple teams still in Apple built NextStep OS. Apple adopted the BSD Core architecture from NextStep and have built on their own filing system that is so venerable it now needs a total overhaul. So a long and deep history of OS Work.
2. Apple was the first successful personal computer company, before IBM, before any other computer company around today actually. There is no company still around with a deeper and longer reaching claim to the personal computer than Apple. Name one who cover hardware and software with the depth of understanding of Apple. The only answer is Microsoft and for them they have only done the hardware side in more recent years.
Asking you to understand this is like asking an 8 year old how you can listen to music on the original Walkman.
> They didn't build a reasonably large amount of their OS underpinnings
??
Apple has been building its own OSes since day one. I never used the Apple II but I used the original Macintosh and Mac SE (68000-based) and it was absolutely their own OS. The developer documentation was superb - I used to own five volumes of "Inside Macintosh".
Unfortunately, by the time they got to System 7, Mac OS had become as bloated and unreliable as Windows of the day. So I was very happy when OSX moved to a solid Unix design as its underpinnings.
You can't blame Apple for this. Look at Microsoft and authentication services. They got it badly wrong several times. In the end they gave up on trying to reinvent the wheel, and just used Kerberos and LDAP instead.
You're calling System 7 bloated compared to Windows 95?
It had other problems.
Such as no multitask, no protected memory, no networking, no security, You name it, it doesn't do it. no... no... no...
Not bloated. A little disorganised and buggy, perhaps, but there's no way anybody but a 1970s IBM computer scientist would call 3.5Mb operating system RAM consumption excessive.
Problem with olde worlde mac OS was... it was so basic it wasn't even an OS. More like a scientific calculator, a few levels up from feeding lists of instructions onto a sheet of bare metal.
2. Apple is a consumer electronics company, not a computer company. It's like asking Hermes to build a power station.
^ This would only be a statement that somebody like a stock market trader might suggest, or somebody who doesn't know the history.
Avie Tevanien, or Betrand Serlet, these are proper systems designers. When Jobs died, Serlet walked almost instantly, his quote was something like "I do computer science not product marketing".
As for whether Apple or NeXT knew anything about web services... there's WebObjects! As for whether Apple knew about the web in general... that's where WebKit came from.
As far as heavily multitasked "distributed" computer systems are concerned... well... Tevanian worked on the original Mach.
It should be noted that apple's version of the XNU kernel is *not* a BSD license. Mostly Apple Public Source License.
Saying "Apple has not succeeded in cloud" is pretty much nonsensical as far as I can see.
For an Apple user, "the cloud" is a place where they can store personal content, enabling them to access it from any Apple device where it has internet access. All fine and dandy and very convenient for millions of users. But that's all it is.
Compare this to the other "cloud" offerings. Here we are talking about hosting virtual servers, under the complete control of those who purchase them. I know MS has pushed heavily into this space with Azure, to the point of offering decent Linux VMs, but why would Apple try and muscle in on it? Surely they've no desire to start hosting other OSes, and I can't really see a significant desire for virtual OSX boxes as I can't imagine what they would really offer, OSX being such a personal computing OS.
I see the person who commented that Apple is more of a consumer electronics company above got downvoted, but I can sort of see their point. Of course this misses the point that Apple stuff is professional grade kit that is used massively in all corners of the media, but surely that's the essence of the appeal of their non-gadget computers? Pro standard media focused computers at a consumer level price.
I'm not surprised Apple haven't felt the need to try and emulate Google or AWS like Microsoft did, they seem to be doing quite well without that, and I don't see them becoming a loss making company any time soon.
That place where people store their content is exactly what the article is referring to. That requires massive amounts of servers and storage to handle everything. Obviously at even a smaller scale it is cheaper to "do it yourself" than to use a public cloud provider. Apple would save a lot by doing more themselves.
Just like Dropbox recently announced they had(or were going to) pull out of public cloud. Drop box I read at one point had more than 10,000 servers in data centers but still used public cloud never understood why, but whatever they wokeup eventually and saw the value. I suspect dropbox management started off exclusively in a cloud provider, performance pains hit them and they moved their performance sensitive stuff onto their own gear, and were too lazy to do the rest, until the costs got too high.
Apple has no need(at this time) to get into the public cloud space to offer VMs and other things to customers. Their cloud is for their devices.
Though in several years time it would not surprise me if they did get into it, after they have built up enough, it is a natural progression. It's what spawned amazon and google and microsoft's clouds after all. You build up enough internal resources and talent(because you must have it to run your own stuff anyway) and it doesn't take much more effort to offer it to the public.
The article I read on el reg I believe stated Apple will be cutting $600 million checks to google (per year) for use of it's cloud. You can do a lot with that kind of cash. Back when apple was swimming in massive growth that kind of money didn't matter, now the bean counters are counting beans more carefully and it matters more now and even more so in the future.
In my mind at least the biggest challenge to doing it yourself and not going to a cloud provider for services is getting the right talent to do it. People like me are rare, and most companies aren't willing to recognize the value provided by doing stuff on your own. The non tech types that have not experienced cloud usage don't know any different. They can't understand why for example investing $5M in doing it yourself can save you $30M from using a public cloud. The math is often so simple they can't compute it, and don't believe it. Which is unfortunate.
The coming tech implosion will be interesting to see how it affects cloud adoption.
The cloud is a systems-administration model rather than a specific technology.
Apple haven't announced anything new since the lock-in of iOS, though.
The big problem is that apple were turned from an artisan's studio computer into a consumer company, totally forgot about supplying tools for enterprise.
Apple could have cleaned up on small & medium sized businesses, but that'd have required a lot more software to be written, and a lot of moody customers.
Of course they took the jewellery option, it's easy money.
Apple currently has at least six distinct cloud infrastructure projects underway....
Well, there's the next fuckup, right there. You see, this cloud thing? There really should be only one of it. Six distinct projects will result in six things with very obvious joins between them.......
Unfortunately Apple have been backed into a corner as a maker of shiny toys.
I can tell you what put the dent in apple's professional & enterprise business:
Migrating from PowerPC to Intel when IBM decided they'd make Xbox with the new Power chip, and IBM refused to sell apple any more chips.
Up until that point, there was a lot of stealthy apple stuff to do with making the OSX environment truly distributed & internet-capable. After that... it's chase the iOS sandbox dollar & shiny shiny marketing.
It should never have been forgotten that Apple develop operating systems and that's the core product. That's been forgotten for over a decade now.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
