Company upgrades your device without warning to a new and relatively untested filesystem.
Yeah, this is just one of the reasons that I don't touch Apple.
Apple today released iOS 10.3, watchOS 3.2 and tvOS 10.2 (14W265), the first two of all of which bring some pleasing extra functionality to iThings, But the main attraction in the new release is Apple File System, because it adds comprehensive encryption to the iPhone and Apple Watch. Apple's been very shy about the Apple file …
Yes, but all worthwhile file system upgrades have to start out as relatively untested. Don't get me wrong, I'm always corporate-skeptical but a significantly better encrypted filesystem and metadata is a good thing in the long term, even if there is the odd (hopefully) minor bug here and there along the way.
@Lee D
You don't actually follow IT much, 'cept p'raps to troll articles about Apple, do you?
Without warning - this has been in the pipe (as far as the public is concerned) since last year. Obviously, within Apple, it's been in the pipe for much longer than that. How much warning do you want? Is a year not enough for you?
Relatively untested? Well, there's been a public beta out for it for at least six months - and you can be damn certain that the t's have been crossed and the i's dotted (j's too) on this one.
Now I fully understand that you don't want to buy Apple. More power to you - more power to all of us. Competition is good, and a better file system for iOS means that Google will have to up its game on Android so you'll win as well from this. The converse applies too of course. This is good. But if you're going to choose a platform other than Apple fer chrissakes have a good reason for doing so - and kneejerk populist bloodymindedness doesn't count I'm afraid. And it especially doesn't count when there are plenty of good reasons that you could have chosen instead.
Muppet.
Ask people who have iPhones and iPads if they know about it.
Ask them if they know it's an in-place upgrade of the thing that holds all their photos
Ask them if they've read the release notes (which the article says don't mention it)
Ask them if they are on the public beta (the existence of which doesn't mean anything).
I don't buy Apple - My reason for not doing so is very simple - I manage several hundred Apple devices.
And an in-place filesystem upgrade is NOT something that you push with an update without even mentioning it in the release notes or giving people a chance to opt-in rather than just doing it.
> Ask them if they know it's an in-place upgrade of the thing that holds all their photos
If they keep all their photos solely on one portable, easy to lose, breakable device then they evidently don't care about their photos.
> Ask them if they've read the release notes (which the article says don't mention it)
What would they do with that information?
> Ask them if they are on the public beta (the existence of which doesn't mean anything).
Ditto.
I have a Nexus phone, so I just tend to go with the updates - so yeah, I'm trusting Google not to bork it (though my photos and contacts are backed up whenever I have WiFi, which is in most pubs these days). When I was on Sony, I'd wait a few weeks and see how other users online had fared with an update (just to make sure that the update didn't dent the excellent battery life or have some other undesirable effect)
"Ask people who have iPhones and iPads if they know about it."
First spend half an hour explaining to them what a file-system is, before realising that most people only have a pretty sketchy idea of what a file is and then realise that as long as their phone/tablet still works roughly the same, that most people couldn't give a flying fsck.
Sure, filesystem upgrades on a live device are...tricky, but if apple reckon they can manage, I'm inclined to believe them.
They have mentioned that they're going to do this (you can tell because we're commenting on an article about it), and they have offered an opt-in test to people who understand the issues.
(I once owned an iPod, that's the only Apple device I've ever bought. I'm not a fanboi is what I'm saying)
" a better file system for iOS means that Google will have to up its game on Android "
How so? The Muppet seems to be you, by assuming that Google is somehow lagging Apple in the file-system arena. (Android uses EXT4 and F2FS mostly these days, both of which are already modern, and support modern features)
@AC
I think, in your blind desperation to prove the superiority of the OS to which you have pledged allegiance, you might have missed the point.
The point is that, overall, neither Android nor iOS can claim superiority - both are excellent, and both are superior, in their own ways. Where one pulls ahead, the other will surely catch up and possibly overtake soon after. And so the dance continues, and both OSes are driven ever greater. For your purposes, Android is better. For mine, iOS is. Ain't that great?
With regard to the superiority of the file system, I dispute that EXT4 is better than APFS. Actually, in some areas HFS+ has the edge on EXT4 (although in aggregate compared with HFS+, and technically, EXT4 is the better overall filesystem). F2FS, so far, is a bit of a sideshow.
> IRT as "a relatively infested OS". Which seems entirely appropriate.
Relative to what, exactly? iOS is a heck of a lot more secure than Android, that's for sure. It's also a hell of a lot more privacy-aware too.
Now, if you want to complain that Apple are a bunch of condescending assholes that are doing their level best to take control of their hardware away from the user, that's an entirely legitimate argument.
If you want to argue that Apple gimps iOS to run slower on older units to encourage upgrades, I would suggest that newer OSes will always be slower on older hardware just cause they do more, but I won't discount the possibility that they may do something silly.
But complaining that iOS is 'infested'? That's horseshit. iOS is pretty solid as an OS, and to their credit, Apple does respond reasonably quickly to security issues that are discovered. And unlike the majority of Android manufacturers out there, your device will stop getting updates only when the next version of the OS no longer supports it. So far that's been averaging the 4-5 year mark.
you don't have to upgrade your Apple iDevice on day 1 of the release you know.
Most really savvy tech people let others do the final testing for them and you know wait a week or even a month before applying the update.
Also, unlike a certain PC Operating system, while Apple might nag you a bit (well sometimes a lot) they don't force the update on you. You can ignore it if you so choose.
As they say, the jury is out on the new Filesystem. I'm sure this site will {gleefully} report any issues people experience with it. Then we can all make up our own minds as to if it is a POS or not.
It doesn't automatically update the OS.
If you do it checks / prompts to see if you have a backup of your device.
If you're worried that a software update might break something then don't update your device straight away, if there are issues they'll soon make the news, if nothing happens then it's safe to update your device, but make sure you have a backup first.
:)
Company upgrades your device without warning to a new and relatively untested filesystem.
.. which went through no less than 6 beta releases before it went public.
I'm an end-user beta tester (not a developer, so I only see stuff when it's pushed out for large volume beta) and I have seen 6 or 7 updates to the 10.3 beta (I lost count - too busy with other stuff) which is unusually high. 10.2 had about 4, I think.
This is not Microsoft, and they don't have to push updates because they need new revenue (which, with the subscription service, has now ended as an update motive for MS too, so maybe, just maybe, they'll now cut better code).
How do you know it was relatively untested? It's been in testing for years before it became public knowledge and has had several beta's for the public to test and is why we haven't hear many complaints! Hopefully the macOS conversion will be as smooth! Your loss for being closed minded and not wanting anything Apple!
Trust a single company I've paid money to do a job and who have a financial interest in maintaining a good reputation for not abusing that trust? Or trust a whole bunch of hardware and software vendors who occasionally pull in vaguely the same direction, with a result that is far too big for me personally to audit even if I had the skills to do so?
The point is, you have to trust someone. If for you that's the open-source community, then good on you.
"Yes sir, it's been encrypted. And the key seems to be based on some predicted future score at Headingly..."
England: 0
India[1]: 5
That one?
[1] My original though was "Australia: 5" and then I realised that that would be stretching the grounds of credulity *too* far..
I had the beta-test version, and it worked surprisingly well by the seventh iteration or so. Still foolable, though — like most aspects of most voice recognition systems. Remains pants on rugby, however — I asked about the 6 Nations, and was informed that no films having that name could be found.
HFS+ in iOS did effectively provide full disk encryption, as explained at Protected Until First User Authentication in the mind-numbingly long iOS Security document (old version, as the current one has already been updated). But that's not to deny that AFS' encryption facilities are better than HFS+'s in many ways.
Apple* have a habit of slowing devices with successive version of IOS. Any idea if the encryption in the OS slows the phones and tablets? A small hit would be worth it for the increased security, but I don't want my devices effectively turned into a treacle.
* Google too, eg the original Nexus 7
Much like Google, they have been doing hardware backed full disk encryption for many years. However they have recently moved away from it, due to the limitations it has, you can get better performance AND features from hardware backed file based encryption.
https://source.android.com/security/encryption/file-based.html
No data checksumming, unlike ZFS.
I've no idea about whether this is true or not, but could it be that the flash controller includes bad block detection and recovery (where possible; returning an I/O error otherwise) at a lower level? If so, perhaps there's no need for duplicated functionality. Besides, I think that read errors are much less of a problem with flash: it's write endurance that's the main problem.
Yes data SHOULD be backed up - however what proportion of users actually back up their data - probably no more than 1% of the users that have not already been bitten by lost data. If Apple (and others) wanted to add a useful new feature to phones (and PCs) - make them do an automatic backup whenever they are connected to the internet by a fast WiFi connection.
> If Apple (and others) wanted to add a useful new feature to phones (and PCs) - make them do an automatic backup whenever they are connected to the internet by a fast WiFi connection.
Indeed. I've written about an (in reality probably unworkable) idea that all laptops must be sold with an external harddisk or NAS by default (unless the buyer signs a 'I really know what I'm doing!' form).
@Duncan MacDonald
Perhaps I've misunderstood your point, or perhaps you jest. This functionality has been part of macOS and iOS for donkeys now. On macOS it's called time machine, on iOS it's just called backup. On macOS you have the option of backing up to a locally connected disk drive, or to a location on the network. On iOS you can backup to a connected Mac or to iCloud.
In either case, you can choose to locate your documents folder and desktop (and more besides) on iCloud - so that everything is kept backed up and synchronised at all times.
Do you want more than this? It's pretty seamless, straightforward and easy to set up…
Bit-rot takes place when the data stored on a disk becomes corrupted (e.g. disk fault or rogue write). This corruption will not (generally) be detected if the "backup" simply copies all the files to the backup media. Even if a rotation is used, you can get to the point where all of the backups contain the corrupted file(s) and the problem may go unnoticed for many, many years (e.g. those wedding photos that you want to show the kids).
That's were file systems like ZFS come in - each block on the disk is check-summed and verified each time it is read (including when it is read during a backup). This means bit-rot can at least be detected (so it doesn't make it to the backups) and can also be corrected if the file system uses redundancy. ZFS systems can also be periodically "scrubbed" to make sure that bit-rot is detected and, where possible, automatically repaired.
Macs will already auto backup if connected to a network, either wired or wireless so long as you have time machine setup. Every hour a backup will take place. Only configuration is pick the destination. Would be great if they extended this to idevices, so long as I could target local backup rather than cloud.
Macs will already auto backup if connected to a network, either wired or wireless so long as you have time machine setup
Indeed. Just don't bother trying to restore a specific file from 6 months ago unless you have a small geological era in which to do it..
(Apparently, Time Machine is going to go over to APFS at some point which, hopefully, will result in the death of the current cobbled-together kludge).
"Yes data SHOULD be backed up - however what proportion of users actually back up their data - probably no more than 1% of the users that have not already been bitten by lost data. "
Backup of an iPhone to iCloud is nothing more than changing a single switch in "Settings". Unless you have tons of data, then you need to pay for more cloud storage.
> "Yes data SHOULD be backed up - however what proportion of users actually back up their data - probably no more than 1% of the users that have not already been bitten by lost data. If Apple (and others) wanted to add a useful new feature to phones (and PCs) - make them do an automatic backup whenever they are connected to the internet by a fast WiFi connection."
To be fair, Apple includes an option in iOS device settings to automatically sync them with iTunes (yes, I know, ugh) whenever your iOS device connects to the same WiFi network, and, in turn, Time Machine makes it really simple to then back up your Mac. I am sure this ease of use means that rather more iOS/Mac users possibly do (organised) backups than do many Windows users. And, personally, I'd rather not backup my devices to "the fog", where who knows who is raking through your files..
Bit-rot is generally a concern for large disk estates, and fundamentally happens all the time. Generally you don't notice it, because the checksum process in the device controller corrects it before sending the data on to the OS. Each block or sector stored on a disk has a significant amount of error-correction added to it, because magnetic media is far perfect.
Unfortunately, the checksum process is not fool-proof, and multi-bit corruptions that pass the checksum calculations are possible. The more bit-flips and disk read operations that happen, the more likely an undetected read failure is to make it past the controller and up to the OS.
As the number of read operations goes up, both because the speed and size of storage estates is increasing, so does the chance of undetected corrupt reads, until eventually it becomes a statistical certainty. We are easily past that point with the largest storage systems around (think how big S3 must be).
Because magnetic devices (particularly) can have magnetic domains (bits) that become marginal and actually flip state both while the device is used, but also when it is idle, due to environmental issues, it is normal for many of the more sophisticated disk controllers to reduce this chance by periodically reading and writing back all data on the disk so that any bits that have been flipped will be written back correctly with new checksum information. This will provide higher confidence that the data read is correct by keeping the number of flipped bits down.
Bit rot in Flash devices is countered by similar processes, but its more common that once flash cells are damaged, the whole block will probably have to be replaced from the spare list, and this can make flash storage devices apparently completely fail suddenly when sufficient failures have happened.
I'm by no means a File System guru (in fact I'm not even an IT professional), so this may be a stupid question.
I gather ZFS seems to have very many desirable characteristics that we're now getting with AFPS (bit rot detection being one of them). Would it not have been easier for Apple to simply implement ZFS rather than writing a whole new FS instead?
ZFS is released on a CDDL licence.
What exactly could Oracle get sniffy about?
"Derived from the Mozilla Public License 1.1,[5] the CDDL tries to address some of the problems of the MPL.[6] Like the MPL, the CDDL is a weak copyleft license in-between GPL license and BSD/MIT permissive licenses, requiring only source code files under CDDL to remain under CDDL. Unlike strong copyleft licenses like the GPL, mixing of CDDL licensed source code files with source code files under other licenses is permitted without relicensing. The resulting compiled software product ("binary") can be licensed and sold under a different license, as long as the source code is still available under CDDL, which should enable more commercial business cases, according to Sun.[6][7][8] Like the MPL the CDDL includes a patent grant to the licensee from all contributors ("patent peace")."
The flash in iOS devices already had full disk encryption. The reason you're able to erase your iPhone so quickly is because it doesn't erase anything except the encryption key it randomly chose for the filesystem, choosing a new one and marking all blocks as being available for reuse. If it actually erased all the blocks it would take more time and a 256GB phone would probably get a bit warm in your hand with all that flash block erase activity (that's the most power intensive thing NAND flash does, by far) If you've ever done a secure erase on your SSD you'll know it is something that takes at least a minute, maybe longer on a large one.
So I'm not sure what APFS is adding to iOS in terms of security. The big changes might be more for macOS.